EDGE Can Out-Perform 3G; Here's Why

goombah99 writes "Blackfriars's communications has an interesting discourse on why the practical difference between 3G and EDGE cellphone data networks is less than it appears to be based on a naive bandwidth metric. Their argument is that the user experience of TCP/HTML is much more impacted by latency, error rates, and processor speed than by bandwidth — and Edge had the edge on all three. Additionally, EDGE may consume considerably less power."

Comments

Is this article sponsored by Apple?

[Corporate Troll]

Or some other iPhone lover? ;-)

Re:Is this article sponsored by Apple?

[Calinous]

The author owns Apple stock - so not only an iPhone lover.
      But what he says sounds true (not sure about his Nokia phone being slowed down by too fast a transmission speed)

Re:Is this article sponsored by Apple?

[clarkkent09]

I don't know but I looked up the author's (Carl Howe) writings on that site http://blackfriarsinc.com/blog/index.html and just about all the entries are positive comments about Apple and iPhone

Re:Is this article sponsored by Apple?

[jrumney]

• iPhone: 667MHz Samsung S3C6400
• Nokia E61i: 220MHz TI OMAP1710

Yep, must be the network causing all that slowdown.

Re:Is this article sponsored by Apple?

[Anonymous Coward]

Yep, clockspeeds tell the whole story and there are absolutely no other metrics that need to be accounted for to determine wireless data bandwidth. No sir-ree.

Re:Is this article sponsored by Apple?

[stewbacca]

You can buy his "iPhone Summer Heat" guide from his site. Of course his information is horribly biased and unreliable, especially since he is obviously trying to make a quick buck off the iPhone hype. Any aspiring grad student can smell the lack of validity a mile away.

Re:Is this article sponsored by Apple?

[MrMr]

I suspect this may shed some light on the motivation of any brand-XoverY fanboy:

http://en.wikipedia.org/wiki/Motivation#Cognitive_dissonance_theory

Re:Is this article sponsored by Apple?

[A Commentor]

And the submitter must also be an iPhone lover... The three things the submitter claims that 'edge' has the advantage is BS.

First 'processor speed' - that's not determined by the network, but the mobile vendor. Faster processors are not reserved for Edge. The article stated one example of a slower UMTS phone, and the submitter generalized it to all Edge phones have faster processors than UMTS phones.

Second 'latency' - the article does not state that Edge has an advantage here, it just states that wireless latency sucks. So the article's argument was that since wireless latencies suck, bandwidth is not a factor in loading web pages so...

The result: loading Web pages on a 3G phone may actually take about the same amount of time as a phone loading those pages over an EDGE network because all the network time is spent setting up and tearing down connections, not actually sending big amounts of data. And so far, most carriers have preferred to optimize bandwidth at the expense of latency. Why? Because it's more marketable (see erroneous analyst quote above).

The problem here is that the HSDPA addition to UMTS, has drastically cut latencies. He also does not state the typical round-trip delays for each of the technologies, they may be similar, but most likely not.

Third, error rates - the article makes the claim that 'higher speeds' translate to 'higher error rates'. He provides no evidence that is in fact the case. It would be like a company that just put out a 2400 baud modem stating that the other 9600 baud modems out there is not going to improve your download speeds because there will just be more errors on the line. In fact, Edge and UMTS transmissions are very different with UMTS using significantly more spectrum than Edge. So, Edge may have higher error rates, but without some actual data, none of us can make either claim.

Overall, a really bad article submission to accept...

Re:Is this article sponsored by Apple?

[Corporate Troll]

Well, the moderators must be iPhone lovers too. My post is currently moderated: 30% Funny, 40% Overrated, 10% Insightful. Not that I'm complaining, I knew it would turn out that way. Originally I was just trying to be Funny....

Re:Is this article sponsored by Apple?

[theelectron]

No, but the GP has a point. Processor speed will have a large impact on the appearant bandwidth, which is all consumers would care about for mobile browsing.

Re:Is this article sponsored by Apple?

[encoderer]

You're right. That's why my 3ghZ P4 totally outperforms my 1.6GhZ Core2Duo. ....oh, wait....

Re:Is this article sponsored by Apple?

[jrumney]

Are you saying that an ARM9 is equivalent to a Core2Duo, while an ARM11 with 2D graphics accelerator is equivalent to a P4?

Re:Is this article sponsored by Apple?

[kkonrad]

Im in Italy and im using umts with my laptop and edge with my hacked iphone... edge sucks

Re:Is this article sponsored by Apple?

[shinmai]

First 'processor speed' - that's not determined by the network, but the mobile vendor. Faster processors are not reserved for Edge. True, but EDGE communications may require less processor time than UMTS.
Other than that, I wholeheartedly agree with your comment. The article is unprofessional and very transparent in it's marketing-BS. Apple should just admit their mistakes and push out an improved iPhone, instead of sugar-coating all the what's wrong with their phone.

I've been very unimpressed by the iPhone from the very beginning, and I didn't even have anything against Apple before they launched the damn thing. All the features advertised in the ads I've seen (granted I haven't seen more than a few. The ones I saw didn't spark interest, so I dismissed most of the hype after that) have been pretty much standard on any decent smartphone for quite some time, and the only "innovation" here is the touchscreen, which I've heard is a bit of a pain to use. Top that with sub-par features like EDGE instead of 3G and I'd much rather save my money for a more "traditional" brand of mobile phone.

Re:Is this article sponsored by Apple?

[devjj]

So anyone with a history of liking Apple and/or the iPhone must be incapable of offering an unbiased opinion?

Re:Is this article sponsored by Apple?

[Anonymous Coward]

Yes

Re:Is this article sponsored by Apple?

[Anonymous Coward]

Some phones have a separate processor for baseband functions. This was commonly the case
until Symbian updated their real-time capabilites to allow all processing to be done
on the same CPU, saving manufacturers million on BOM costs.

Re:Is this article sponsored by Apple?

[Score Whore]

Honestly I'm not into following cell standards and technologies, but I can tell you this. I'm currently attached via a samsung A900 phone via sprint's evdo network tethered with a USB cable.

A speed test via dslreports tells me I have 167 ms latency to Speak Easy, I have 979 Kbps incoming and 97 Kbps outgoing. I'd not want to play Quake at 167 ms latency, but it's perfectly usable for a remote shell, IM and web browsing. I dunno if EDGE has latencies better or worse. It'd be interesting to know.

Re:Is this article sponsored by Apple?

[dgatwood]

They're probably right about most of the time on some web sites being spent waiting for creating and tearing down connections. For a lightweight web page (CNN), the vast majority of my time using an iPhone on AT&T EDGE is waiting for the initial DNS lookup.

That argument breaks down very quickly, however, when you have a web page with large graphics (Flickr), streaming content (YouTube), or large pages of text (any FAQ from back in USENET days). At that point, I'd be amazed if 3G weren't faster, as EDGE is downright painful. I just don't go to those sorts of sites unless I'm on Wi-Fi.

What could really help performance a ton in phones would be:

• Take advantage of HTTP pipelining. It basically makes the latency a non-issue for requests from a single server, DNS latency notwithstanding. Unfortunately, a lot of web servers are broken and don't support it correctly....
• Heavy local DNS caching to the maximum degree allowable by the TTL.
• Use TCP instead of UDP for DNS requests.
• Use multithreaded DNS resolvers.
• Don't block displaying the page on any content ever, even if that could potentially cause a relayout. If a site doesn't resolve quickly, load the content when you can.
• Use a transparent proxy server. By doing so, you can pipeline all HTTP requests through a single connection without caring whether the final servers support pipelining or not. More importantly, you can pipeline all of the requests through a single connection without doing any DNS lookups over the high latency connection. If you design the proxy correctly, you might even be able to send all requests on the same connection without caring about what host will eventually receive it. I'm not sure if that's possible with the current proxy protocols or not.

Re:Is this article sponsored by Apple?

[jrumney]

If the article had been comparing a low-cost feature phone where the manufacturer cut corners in this way, it would be relevant, but both the iPhone and the Nokia E61i have separate baseband processors, so it doesn't really come into it. What is important is how quickly the browser renders, or appears to render, the received data to the screen. In this respect, the iPhone clearly wins, due to I suspect a combination of raw app processor power, and the fact that the iPhone was designed around the mobile browsing experience, rather than a phone with the rest tacked on as an afterthought but only if it meets the cost profile that the network specifies.

I don't doubt that the iPhone has set new standards when it comes to mobile web browsing experience, but it certainly isn't due to the EDGE network.

Re:Is this article sponsored by Apple?

[Lars T.]

• iPhone: 667MHz Samsung S3C6400
• Nokia E61i: 220MHz TI OMAP1710

Yep, must be the network causing all that slowdown. IOW people should buy the Nokia instead of the iPhone because it has the faster 3G instead of EDGE, the fact that it is too slow to use the speed of 3G, and the iPhone actually works faster non-withstanding.

On The Edge

[Anonymous Coward]

Edge (EDGE) has the edge on processor speed? You mean processing efficiency? Or is EDGE an embedded processor brand now as well? (Or is that the Edge the summary wants to talk about?) Or are we talking about The Edge the guitar player?

I find this edgy and I'm not a Ubuntu type.

Re:On The Edge

[hey!]

Like many engineering issues, this is one of balance.

It's not that Edge has any advantages when it comes to processing efficiency, it's that to take advantage of 3G's greater bandwidth you need more processor bandwidth than you can get today in a reasonable mobile package.

If you take a Dodge Neon on the autobahn, you don't enjoy the unlimited speed offered by the highway very much. If you are making frequent side trips, you might do better with a slower road with more frequent exits.

In the end, there is no single thing as "speed" when it comes to networking. There are several, such as bandwidth, response time, and latency. If I had my choice, 3G would be my choice for applications that have to deliver large volumes of bits at a consistent rate. That doesn't describe most web use by a long shot. It does describe streaming high quality video to a device that can display it, but has limited buffering capacity. If you think about that, in mobile applications that's a rather narrow niche in which to have a killer advantage.

It comes down to balance. Does 3G widen the narrowest bottleneck in my mobile network use? If not, then it's advantages don't mean much to me. It may be that other bottlenecks have to be widened before anybody needs 3G's peculiar advantages.

Coming back to TFA, it may be that the iPhone would be better suited to exploit 3G's advantages than other phones. But you can't get an iPhone on 3G, so it's an academic question. The practical question is whether the less powerful devices on a 3G network can exploit that network well enough to outweigh the iPhone's attractions?

Personally, it doesn't matter to me one way or the other, because nobody has good enough coverage to render that issue irrelevant. I don't care how "fast" a network is unless I can reach it every place I have to go. If I lived and worked in Manhattan, this might not be an issue, but then I'd have better things to do with my time than watch videos on my phone.

Re:On The Edge

[Anonymous Coward]

If the long term average bit rate is same due to HTML/TCP connneciton issues then the lower average bit rate transmission mode can be done be more effieinntly on a slow processor compared to a fast processor trying to handle bursts but idle most of the time.

Re:On The Edge

[Anonymous Coward]

>If you take a Dodge Neon on the autobahn, you don't enjoy the unlimited speed offered by the highway very much. Until you drop a nice turbo at 35psi onto an srt-4 motor, and goto 850hp.

Re:On The Edge

[amsr]

Actually 3G coverage is pretty good, and with a population of 1.5Million, many of who could probably afford and want a 3g iphone I'd say they would probably sell a lot. The "3g coverage" argument is a false one. 3G deployment is mainly in densely populated areas for a reason, thats were most customers are. Nothing about having a 3g phone means you can't use EDGE when you go out in the boonies... I think the power consumption argument is far more viable, but even then 3g blackberries do alright so who knows...

Re:On The Edge

[hey!]

3G deployment is mainly in densely populated areas for a reason, thats were most customers are

If I'm going to stay put where my bandwidth is, I'll stick with DSL.

But you underestimate my bias here. I'm talking about voice. The most important thing about a phone is that it carries phone calls. And yes, I do sometimes visit the boonies, and I want to be able to make calls from there.

Re:On The Edge

[amsr]

Edge vs. 3G support has nothing to do with your ability to place calls in the boonies. The iPhone works, even without EDGE coverage, in areas where it can get a GSM signal. The visual voicemail is dependent on the data network, but if you notice, if you don't have EDGE coverage on your iphone, it prompts you to "call" your voicemail.

The latency issue is for real

[allcar]

When performance testing web applications, I typically find that latency does indeed have a very significant impact. Obviously some types of application are more susceptible than others. Bandwidth is critical in data intensive applications. Latency is much more important in highly interactive applications. Rich Web 2 applications, making lots of (Ajax) calls to the server for small amounts of supplementary data are badly hit by latency problems.

Re:The latency issue is for real

[Saint Fnordius]

Translating into "layman's terms", EDGE is more responsive than UMTS or HDSPA, but the 3G protocols are better at shovelling huge files up and down the stream. That means that lots of small data like an IM client will feel faster on EDGE, but downloads and video will be faster on UMTS/HDSPA.

I can accept that argument. If this is true, then Meebo would be faster on EDGE; but YouTube faster on UMTS. Using my cell phone as a modem (no DSL in my neighbourhood), I can say that my experience has been pretty much like that, though I thought it was due to longer "handshaking" at the beginning of a UMTS connection...

Re:The latency issue is for real

[somersault]

I already thought it was in layman's terms all :P

Re:The latency issue is for real

[somersault]

That'll learn me to use smily faces that involve the less than sign and then a greater than sign later in the post..

I already thought it was in layman's terms? More layman than UMTS and HDSPA anyway :p This is assuming that your average layman has played Counter-Strike, and so knows that latency > all :P"

Re:The latency issue is for real

[allcar]

Thanks for all those "layman's terms", like UMTS and HDSPA.

Re:The latency issue is for real

[Anonymous Coward]

Translating into "layman's terms", EDGE is more responsive than UMTS or HDSPA, [...]

Wrong. Translating into "layman's terms", the author of the article claims that EDGE on one device (iPhone) is more responsive than UMTS on an older device (Nokia E61i).

He did not even mention HSDPA or HSPA, which would have blown EDGE away. And contrary to what the author of the article claims, real-world measurements of network latency (not even mentioning bandwidth) show that HSDPA is faster (more responsive) than EDGE.

The whole article is based on a flawed comparison between only two devices, and one of them (the Nokia phone) is known to be inefficient for 3G. If he would have used other phone models from Nokia or other manufacturers like SonyEricsson, he would have got different results. But then, the article would not have been so interesting for someone who owns Apple stock.

Full disclosure: I don't own Apple stock.

Re:The latency issue is for real

[arachnoprobe]

I would actually go further. In my experience with several to-end-user web applications, latency is the only thing that counts, as long as you are displaying some kind of useful progress-bar for any bandwidth-limited actions (e.g. uploads). The "normal" user knows about limited bandwidth, therefore he is used to that. Powerusers will switch to working in multiple tabs, so they don't have a problem with that either.

Re:The latency issue is for real

[Anonymous Coward]

That's just plain wrong.

HSDPA latency is significantly lower than for UMTS, thanks to a couple of enhancements (Lower TTI, HARQ, etc). There's been a major effort to reduce latency in the 3G/3.5G systems in order to make VoIP viable.

http://www.umts-forum.org/component/option,com_docman/task,doc_download/gid,1632/Itemid,12/

Re:The latency issue is for real

[CastrTroy]

So why not have 2 chips in the phone. Use one network when you want to stream huge video files to the phone, and the other for all the smaller downloads, which require less latency. Granted I think that on a mobile device, the need for low latency outweighs the need for higher throughput. If you want video on your phone, just have a SD card, and download the video from your home internet conneciton. I got a new iPod Nano Video, and I have no problem loading the videos and music on in the morning before I leave for work. I think that the networks need to make huge leaps before we can expect a large amount of people to be streaming video to their cell phones.

Re:The latency issue is for real

[Anonymous Coward]

Here's an even better Article about Latency in the wireless networks.

http://www.ericsson.com/hr/about/events/mipro_2007/mipro_1137.pdf

Re:The latency issue is for real

[duffbeer703]

I recently went on a NY->South Carolina road trip, and used my Treo 700w with Verizon on the way down for getting maps from Google and doing very casual web browsing. Alot of the I-95 corridor is still using 1xRTT, and I was expecting really abysmal performance.

My experience was similar to what is being discussed by others here -- overall response times, especially for applications like Google Maps that are doing alot of small image loads, seemed better than on 1xRTT than EVDO. I didn't notice any difference in battery life -- the Treo just sucks ass.

Re:The latency issue is for real

[Anonymous Coward]

Generally correct.

However, the resolution of cellphone screens removes the need for high bandwidth networks in the first place. Most videos will be QCIF/CIF, heavily compressed with H.261/263/264..hardly bandwidth intensive. The only time bandwidth is going to matter on a phone is when you do large file transfers.

Re:The latency issue is for real

[sigmoid_balance]

Simply because they cost money, they drain more battery power, the increase phone size, the make the phone hotter and require more cooling, thus increasing the phone size even more and so on. Did I mention battery?

Re:The latency issue is for real

[Saint Fnordius]

I should have been clearer. I meant that if I understood the argument correctly, then the argument was that EDGE had less turnaround between request and reply, and UMTS had the faster transfer of the individual files. Whether the argument is actually valid or not is another matter. I just was trying to ensure that I understood correctly how it could be.

My own anecdotal evidence comes from using a Motorola RAZR V3xx as my modem, and it often will start in 3G, bump down to EDGE, then GPRS, then 3G again without losing the connection. Until now I thought it was due to broadcast strength from the cell tower, but this article made me more uncertain.

Re:The latency issue is for real

[TechyImmigrant]

> seemed better than on 1xRTT than EVDO

That's because you are possibly the only person in the world left using 1XRTT. You had a lightly loaded service.

Re:The latency issue is for real

[homer_ca]

Is EDGE so different from UMTS that it needs different chips? My CDMA phone runs both EVDO and 1xRTT.

I know what you mean about streaming video. If you know what you want to watch, it's easier to load it on the flash card at home, but with Youtube, it's more about surfing around and finding funny clips you didn't know about before.

Re:The latency issue is for real

[drew]

So you're saying that once I upgrade to a 3G cellphone, my network performance should be high enough to support real time voice communication with other people? I can't wait!

Re:The latency issue is for real

[Mike Buddha]

Translating his article into layman's terms would be useful if his article wasn't pure crap. I have EDGE and 3G (HSDPA) phones that I use with the same account on AT&T's network. Hands down, HSDPA has lower latency. His 'Arguments' are pure crap. The battery issue is the only thing he mentions that holds any water, and it's really not that noticeable. If you charge your phone every night, does it matter if it has one bar left or two bars left?

I use a Cingular 3125(EDGE), a Samsung Sync(HSDPA), and a Samsung BlackJack(HSDPA) on AT&T's network. For about a week I had to use the 3125 as my sole internet connection. I went right out and picked up the Samsung Sync after that experience. It was like night and day. Playing WoW through the 3125 was like pulling teeth. I almost (God forbid!) stopped playing (well, at least until the end of the week). Doing the same with the Sync was a joy. So much so that I avoided installing broadband at my house until I got a roommate.

Re:The latency issue is for real

[chihowa]

EDGE and 3G can fit in a small phone. My Nokia 6120c has both GSM/EDGE and UMTS/HSDPA and switches between them easily. I'm not sure what the battery life issues with 3G are supposed to be, either. This phone with bluetooth always on, occasional calls, occasional data, and constant application use lasts easily as long as any of my other non-smartphones. It easily outlasts my laptop battery when used as a bluetooth tethered modem, also.

Re:The latency issue is for real

[Reaperducer]

I think it depends on a lot of factors like the chipset and how much you use it.

I did some battery life tests on my SE M600i, and generally speaking (because I don't have the numbers in front of me) turning off the 3G function and relying purely on GSM about doubled the battery life. In fact, one of the first things many Americans who use this phone do is go into the settings and turn off the option to search for 3G networks in order to immediately get an extra hour of talk time out of it.

The chip makers seem to be in a rush to produce new low-power 3G mobile chips. I think they assumed Apple would settle for what they already had and were surprised that Apple went Edge with the iPhone. Jobs has always said it's about battery life, and when I see how the chipmakers are responding, it seems to be true.

Re:The latency issue is for real

[kraka40]

Ok folks .. yes latency is real, but it impacts each network stream independently. The problem with EDGE networks is that the bandwidth is not available to adequately multi-task and fill the gaps (created by latency of other streams) . 3G offers a fatter pipe, so even if there is more latency (assuming you have the CPU power and programming in your mobile device) you now can take advantage and fill those "gaps" with requests for more rich content. It is not the initial GET that is of any bandwidth concern, it is the subsequent GETs (which should be executed in parallel as much as possible) to fill in the rich content necessary for the page.

Of course as mentioned earlier this is all moot until the mobile devices have the proper programming and hardware speed to take advantage of it.

Probably the best test of this would be to test this on a laptop with two separate mobile networking cards from each network. I'm pretty sure you'll see much better performance from a 3G network.

Diggdot?

[Anonymous Coward]

So one guy who owns an iPhone (and Apple stock) writes an argument, based on his own limited experiences with an iPhone and a Nokia, without any precise measurements, concluding that EDGE is better for mobile web browsing than 3G.

Submissions to "articles" like these are making Slashdot look more and more like Digg. I don't know about the rest of you but in my opinion, that's a Bad Thing.

Re:Diggdot?

[jellomizer]

Well for some odd reason everyone thinks Apple uses the latest and greatest technologies in their products... Apple doesn't they use tried and tested technology they are rarely on the latest and greatest Processors and Video cards and displays... (I am sure the iPod touch/iPhone touch display technology has been out for years) Apple most likely went with the EDGE for 2 reasons. 1. Less Power Consumption, if you I phone couldn't keep a day charge then people won't like it... 2. Availability if the iPhone only works on the internet in selected spots ignoring small towns then people won't like it. If the 3G was more widely deployed and their technology used less power the next version of the iPhone may support it. But I agree his comment was a lot like Steve Jobs rational that the G4 Chip out preformed the Intel Chips of the Day. Focusing on one benchmark not overall benchmarks. Granted latency is what really effects people from staying the internet is slow or fast. But if it happends in a split second verse a second people don't care much.

Re:Diggdot?

[Luscious868]

So one guy who owns an iPhone (and Apple stock) writes an argument, based on his own limited experiences with an iPhone and a Nokia, without any precise measurements, concluding that EDGE is better for mobile web browsing than 3G.

Clearly you didn't bother to RTFA. The author makes the argument that while 3G networks have more bandwidth than EDGE (they can transfer data at a faster rate), that higher bandwidth comes at a cost of higher latency (the time it takes for the transfer to begin) and more power consumption. He, correctly, points out that the higher latency of a 3G network can make your browsing experience slower than EDGE depending on the type of content you're browsing. If you're dealing with large files (think audio and video intensive sites) then having the higher bandwidth of 3G is an advantage but if your dealing with a typical website (made up of many small files) having the lower latency of EDGE is an advantage.

What's the point of having a connection that is twice as fast when transferring small files if the time it takes for that connection to begin is twice as long? The more small files being transfered, the "slower" things will seem.

My boss learned the difference in the late 90's the hard way. We were stuck sharing a dial-up connection at the office and my boss wanted a higher speed connection so he tried a satellite. I tried to explain the difference between bandwidth and latency and how they can affect your experience. He didn't get the difference and after the satellite was installed he complained like crazy when browsing the web because the connection didn't seem any faster than when using our old 56K modem. It was only when downloading large files from the web that you really felt the effects of the increased bandwidth.

Re:Diggdot?

[zafo]

Based on my own unscientific comparison, my personal Helio Ocean with 3G seems to be faster at bringing up web pages than my business Blackberry with EDGE. I haven't done a side-by-side comparison, though (accessing the same web page at the same time).

Re:Diggdot?

[Anonymous Coward]

Clearly you didn't bother to RTFA. Actually I did RTFA. I never said that what he claims cannot be true.

I'm saying or implying that:
1.) This guy has an obvious bias
2.) His method (or lack thereof) of proving his argument is worthless.
3.) Fanboy blog articles of sub par quality should not be on the frontpage of Slashdot.

When I read Slashdot and I see a headline saying "EDGE Can Out-Perform 3G", I expect a serious analysis or experiment (no, I'm not new here)

Re:Diggdot?

[johnkzin]

So, then, in order to publish an article doing a comparison, it has to come from an outfit big enough to have a formal lab? Isn't this a version of the "corporate journalist vs blogger", or "corporate programmers vs open source contributor", arguments?

So, he's one individual who brought up a topic. He doesn't have a huge formal testing environment... and while you may find him biased, he did at least have the integrity to STATE his bias (you wont find an unbiased source, those don't exist; but finding a source which is up front about its biases is rather valuable).

This is the bazaar, not the cathedral. If you don't like/believe his results, come up with a way of creating the hard number analysis you think is necessary, and publish it as a rebuttal (or in agreement, or whatever you find). Complaining that he's not a cardinal (person in the cathedral) is just silly.

And, frankly, hasn't /. _always_ been about the bazaar of journalism?

Though, maybe it would be a good idea to ask /. to implement a reputation system other than the default karma score ... something that gives a long term measure of the person's journalistic integrity (I don't get the feeling that that's what the karma score is really giving you). There are people who get complained about whenever they post some article ... if there was a built in measure of that public reaction, then it would be easier to sort the wheat from the chaff. And newbies would be somewhere in the middle.

Re:Diggdot?

[oliderid]

Before:
"A mouse with one button is better than a mouse with two buttons."

Now:
"Edge is better than 3G"

Mac fans look sometimes a bit errr...fanatic :-)

Re:Diggdot?

[stewbacca]

Submissions to "articles" like these are making Slashdot look more and more like Digg. I don't know about the rest of you but in my opinion, that's a Bad Thing.

thumbs down...

Re:Diggdot?

[stewbacca]

As a long time Mac user, I take exception to your jab, sire.

Me Before (about 10 years ago):
"Why are my fellow Mac users too stupid to use 2 buttons?"

Me Now:
"Why do PC zealots keep bringing up the one-button mouse argument when it hasn't been valid for 10 years now?"

Re:Diggdot?

[stewbacca]

Why did you qualify your statements with Apple "most likely" went with EDGE for 2 reasons? Jobs has been crystal clear on why they chose EDGE. You summed it up nicely. Everything in life is about tradeoffs. In a rare mistep, Apple chose the wrong tradeoff with the iPhone (IMHO). Does anybody really want longer battery time at the expense of slower data transfer? I know I don't (but bought the phone anyway).

Re:Diggdot?

[Dionysus]

I seem to remember the Mac being one of the first machines that had Firewire and USB

Re:Diggdot?

[Anonymous Coward]

Completely agree. Worthless article, worthless point, iFanboys in full effect. Stories like this are why I took Digg off my RSS reader. It used to be good, but then the interesting articles were replaced by stuff prising Apple, Linux or Nintendo, or crapping on Sony and MS. Then came Ron Paul and the onslaught of fanatical political stuff.

Anyway, iFanboys, the rest of the world still thinks your little toy phone is funny and outdated. It's slow and locked down, you're a sucker, deal with it.

Re:Diggdot?

[Anonymous Coward]

A mouse with one button IS clearly better. If you need two buttons, then the women who buy them would have to take off the oven mitts just to see the rest of the recipe.

Re:Diggdot?

[jellomizer]

"rarely on the latest..." not "never on the latest..."

Sure every once in a while Apple comes up with some new technology. But normally they will take a step back and work with what works.

Phrases like rarely, usually, show a trend toward, most likely, for the most part.... tend to qualify that there exists exceptions to the statement but those exceptions are more of a minor occurrence then the general statement.

Men tend to be taller then Woman doesn't mean all Men are taller then all Women, Just the average height of men is greater then the average height of women.

Re:Diggdot?

[LarsG]

Oh man, don't they teach source criticism any more?

The public explanation from Jobs is "better battery life". Now, that might be true, and that might be the only reason they went EDGE only. But it might also be salesman spin ("we know our valued customers want good battery life, so we limited the iPhone to EDGE for your benefit"), and the reasons for EDGE only might be completely unrelated to what Jobs is saying publicly. The point is - unless you are His Steveness using a pseudonym you can't categorically state that battery life is the one and only reason for his choice.

One reason can be the limited roll-out of UMTS/HSDPA in the US.

An other could be cost, they wanted a higher margin.

A third that they might want to go least-common-denominator (GSM/EDGE) so that the rev1 iPhone would work, and hence possible to sell, in pretty much all markets; that is, they wouldn't have to do a rev2 before they entered the eu/asia market. If rev1 was intended for US only, it might have been better to go CDMA. Especially considering that a CDMA phone is easier to lock to a carrier. But that would require Apple to do a rev2 before they could enter the other markets.

A fourth that it would be unwise to include everything in rev1, since they need to put something better in rev2 so that you have a reason to buy the next one.

Re:Diggdot?

[Bodrius]

No one is complaining that the blogger created the article.
What they're complaining is about the relevance and editing of the item - is this front-page material or not? Is the headline accurate or not?

Slashdot was supposed to be "news for nerds, stuff that matters", and it is questionable whether every random blogger argument matters enough to be in the front page, with a headline that seems a categorical statement of fact.

If you want to use the open-source metaphor, that works too: not every random open source project appears in the front page, nor should they. Most projects (like most anything) do not matter that much; not until they acquire some traction and someone has put a good amount of effort and good code.

I see this as yet another 'editor complaint' - perhaps the article should not have gone to the front-page. Perhaps it should have been combined with other similar/related submissions to estimulate a discussion.

I personally would have been happy if Slashdot had not used a headline that, unlike the original blog's:
- Claims X is better than Y as a fact
- Gives an expectation of evidence

I don't care much about the subject either way (I'd like more phone and less net on my phone, thanks).

But I agree with the parent that the only difference between Slashdot and Digg is that there is an editorial staff which is supposed to filter 'what really matters', and provide context / research to the submissions where necessary.
If there is no filtering and no editing, Slashdot might as well go with the Digg model and remove the 'editorial' process -let the masses and moderation systems deal with that.

Re:Diggdot?

[nxtw]

The author makes the argument that while 3G networks have more bandwidth than EDGE (they can transfer data at a faster rate), that higher bandwidth comes at a cost of higher latency (the time it takes for the transfer to begin) and more power consumption. He, correctly, points out that the higher latency of a 3G network can make your browsing experience slower than EDGE depending on the type of content you're browsing.

The author is right about 3G networks having more bandwidth and higher power consumption, but 3G has *much less latency.

The article is full of bad assumptions and his reasoning for 3G's higher power consumption is faulty.

Re:Diggdot?

[Wdomburg]

The author makes the argument that while 3G networks have more bandwidth than EDGE (they can transfer data at a faster rate), that higher bandwidth comes at a cost of higher latency (the time it takes for the transfer to begin) and more power consumption.

He claims it comes at the cost of higher latency, but fails to present a real argument for it. Benchmarks I've seen peg latency on EDGE four times higher than UTMS, which kind of kills his argument.

Re:Diggdot?

[Anonymous Coward]

And a fifth could be that Apple engineers were not capable of making 3G work.

After all, these are guys that have to be sent to 'Mordor' (i.e. Shenzhen) to sit at the feet of crappy Foxconn engineers. (I never got to Shenzhen, Shanghai's much better).;

Re:Diggdot?

[elrous0]

Thank God is was for quality reasons, and not because AT&T gave them a shitload of money.

Re:Diggdot?

[Score Whore]

Unless latencies are stunningly bad on the 3G networks, I'm not sure that the argument that low latency/low bandwidth beats higher latency/high bandwidth is true.

Consider this. Let's say that EDGE has 50 ms RTT and 200 Kbps bandwidth. And let's say the 3G are running 200 ms RTT and 1000 Kbps bandwidth. Now you hit cnn.com and load the front page. Currently that is 133 KB of just the html. So you get your connection setup in three packets and then send your request and wait to receive your reply.

So for EDGE, your three-way handshake takes 150 ms, then send your request and wait another 50 ms before data starts arriving. Then the time for just the text to load is a little over five seconds. Then, depending on if your browser supports keep-alives, etc. you tear down the connection or not. Anyway the total time from start of the request to receiving the last of the data is about 5.2 seconds.

For 3G, the handshake is about 600 ms, send your request and wait 200 ms before data starts arriving. Now the 133 KB page takes about 1.3 seconds on your 3G network. Total time, about 2 seconds.

If your average webpage is built up of lots of itty bitty images served from lots of different servers, then it's possible that EDGE might be able to game some advantage, but you'd have to have all your objects being no more than 15 or 20 KB before the bandwidth penalty kicks in and the 3G networks stomp all over you.

But if the images are being streamed from a single server, then it's a very different situation since you don't have to keep repeating the handshake. Turn on keep-alives and woot, your handshakes go down by a factor of ten or more. Turn on pipelining and you don't even have the lock step request/response cycle to contend with. In that situation bandwidth is king.

----

Oh hell. I just hit Google to see if the numbers that I pulled out of my ass above are at all attached to reality. And I'm afraid to point out that they're not. EDGE latency is around 500 ms to BBR's test page. My EVDO phone comes in at way less than that. And bandwidth is much higher.

I think the story is pure BS.

Re:Diggdot?

[Reaperducer]

The public explanation from Jobs is "better battery life". Now, that might be true, and that might be the only reason they went EDGE only. But it might also be salesman spin ("we know our valued customers want good battery life, so we limited the iPhone to EDGE for your benefit"), and the reasons for EDGE only might be completely unrelated to what Jobs is saying publicly. The point is - unless you are His Steveness using a pseudonym you can't categorically state that battery life is the one and only reason for his choice.

I didn't entirely believe the battery life argument at first, but more and more it's starting to ring true. In the last few months the chip makers have been falling all over themselves announcing how their latest and greatest and next generation chips will drastically cut power consumption and extend battery life compared to current chips. If 3G chips were OK already they wouldn't need to do that. I also think they're all vying to hit that sweet spot in power and price that will get Apple's attention and make the company buy their chips.

Re:Diggdot?

[Anonymous Coward]

I totally agree... this article is beneath slashdot! I want quality news, not non-sense ramblings by someone who has no idea what they are talking about!

Re:Diggdot?

[LarsG]

Power consumption is pretty certain to have been one of the reasons, I was just pointing out to stewbacca that Jobs making a big deal about it doesn't mean that it categorically was the only reason. Jobs is a great salesman, so it is only to be expected that he will promote the positive sides of technical choices Apple makes in their products. And Jobs does have a point. If I turn on everything on my N95 it will suck the battery empty in no time flat.

As for 3G, yeah the current crop draws more power than 2G. And as always the mfgs are pushing for cheaper, less power drain and smaller (more integration in a single-chip/few-chip design). The mfgs would obviously be delighted if they scored a design win with Apple for iPhone2, but one also has to remember that the iPhone currently is a niche product; a design win with SE, LG, Moto, Nokia or Samsung would pop as many if not more champagne bottles in their boardroom.

Re:Diggdot?

[Anonymous Coward]

Submissions to "articles" like these are making Slashdot look more and more like Digg. I don't know about the rest of you but in my opinion, that's a Bad Thing.

But posts like yours (and other posts which have yet again educated me on something I didn't really understand that well) keep making Slashdot look like the Slashdot I've loved for 10 years for just that -- the knowledged and critical community laying the smack on the topic.

If you feed lions, maybe they like a light snack too every now and then. The difference between Slashdot and Digg is thus a zoological one IMHO ;-)

This may be true...

[reidconti]

... but my iPhone is still slow as hell doing anything on EDGE.

It was well worth the (lower) price, as 400 is what most of my phones have cost, and they last me a long time, but I get the feeling I won't have this one for very long if the 3G version comes out soon :)

Re:This may be true...

[ozmanjusri]

my iPhone is still slow as hell doing anything on EDGE

Just as a quick test, Redhat takes 48 seconds to load fully on my 3G M600i, and 11 seconds on my ADSL2 line at home.

How long does your iPhone take to load the same site?

Re:This may be true...

[pgillan]

It took me 40 seconds to load redhat.com on my iPhone.

Re:This may be true...

[Curmudgeonlyoldbloke]

Some more numbers...

90 seconds on an Xda Orbit (Windows Mobile 6, GPRS only SIM card)
20 seconds on a Blackberry 8800 (GPRS only SIM card)
(and about 13 for me with Firefox and Noscript via ADSL)

What's going on here isn't that RIM have some magic beads that make GPRS 4 times faster - different pages are getting served to each device. Redhat serves something pretty close to the "full" page (the same as the PC browser gets), whereas the Blackberry doesn't get sent the graphical tabs arrangement at all (although it does get sent the graphical adverts).

So comparing "address X on device Y" as a test of speed needs to take into account what's actually sent, and how usable it is on the device when it arrives. Redhat.com's actually an example where the Blackberry page is better suited to the device than the Windows Mobile one (it has proper links on rather than a mangled tabset), but this probably isn't typical.

Re:This may be true...

[Tony Hoyle]

I'm currently up to 2.5 minutes and I'm still counting.

iPhone is pretty useless for browsing websites outside wifi zones.

btw. WTF have redhat done to their page that takes so long?

Re:This may be true...

[nxtw]

It took me 21 seconds on my Samsung SGH-i607 Blackjack, using HSDPA & the Opera browser set to desktop rendering mode (so it shows the same page as desktop Opera would)

It didn't start downloading until 5 seconds into that for some reason, and it was browseable after 16 seconds or so.

(The page took about 4 seconds to load on my ADSL line)

Re:This may be true...

[Anonymous Coward]

It just took 12.5 on my N95 over wifi from home.

Re:This may be true...

[BorgDrone]

Took about 17 seconds on a Nokia 6120 over 192Kbps UMTS (I have HSDPA disabled, enabling it would not increase the transfer speed but it would decrease the latency, also, it cuts battery life in half)

Re:This may be true...

[timster]

34 seconds on my iPhone (on EDGE). Maybe you live somewhere with poor AT&T service?

Re:This may be true...

[gEvil (beta)]

Maybe you live somewhere with poor AT&T service?

Yes, we call that "America". : p

Re:This may be true...

[Tack]

11 seconds on my BB 8300. RIM's magic beads involves recompressing images in transit. But your speculation that BB doesn't get served up the graphical tabs isn't quite the right perspective. What's really happening is that the tabs are rendered using certain CSS properties that BB doesn't fully support. So BB receives the same page and css, it just processes it differently due to its incomplete CSS support, so it doesn't end up requesting the background images used to build the tabs.

Re:This may be true...

[kiso]

I have two ADSL lines from two different ISPs at the office and one beats another on various pages. Why? Completely different routing, bandwidth sharing policy, squid caching policy, packet loss, etc. There are so many variables, that simple loading a page through different ISPs and counting seconds doesn't make a benchmark of the channel between you and the carrier. Try loading something from the carrier's internal network, both through EDGE and 3G. That'll make it somewhat closer to reality.

Re:This may be true...

[steveo777]

Samsung Blackjack with 3G... too bad they haven't implemented 3G in the Twin Cities. So I'm on EDGE.

The first two times I tried loading the page I got to 34KB and my phone stopped loading, so I reset it and now it hangs at 33KB. I have had nothing but problems with this POS and I'm very close to just out right paying AT&T the 200 to get off my contract and go elsewhere. Although I have no qualms about blaming it on AT&T's shoddy network.

Tried loading the page after killing all apps, gets to 152KB after loading for 25 seconds and hangs (gave up after 5 minutes)

Re:This may be true...

[Anonymous Coward]

So....different pages are being served to different devices?

Regardless of whether they can process them correctly, they're still getting different pages.

Re:This may be true...

[stewbacca]

It took me about 2 seconds to navigate to a new browser page. Then about 2 seconds to type "redhat". The page loaded in under 4 seconds. But then again, I'm using 10mbs wireless from my apartment. Which is the whole point of the iPhone offering WiFi.... Edge is a (painfully slow) alternative for those times there is no WiFi signal. Jobs has made it pretty clear that EDGE is not intended to the primary data network.

Re:This may be true...

[Tack]

Not really. In some cases I'm sure that's true, but in the redhat.com case, it seems as if all devices are being served the same HTML and CSS content. The visual differences are a matter of rendering.

Re:This may be true...

[WhiteWolf666]

Not an iPhone, but Redhat.com takes 8 seconds to load on my 3G EVDO Motorola Q, as long as I'm using the Opera browser.

That's definitely not cached, as I've never visited Redhat.com, and I'm using the "Desktop" view, so all images are being loaded.

Re:This may be true...

[meepzorb]

redhat.com took 19 seconds on an HTC Mogul (Sprint/EVDO, Windows Mobile 6, using the vanilla IE Mobile browser).

Re:This may be true...

[Loki_1929]

Doign anything at all is slow as Hell on EDGE. I've used several networks since they became available. EDGE is just plain slower than anything else I've ever tried. Verizon's 1xRTT is better than dial-up, but tends to drop a lot of packets when your signal isn't so hot. As such, the overal bandwidth drops as the signal degrades. Sprint 1xRTT does incredibly well with low signal. It's slightly slower than Verizon's 1xRTT under ideal conditions, but doesn't have any of the same issues when the signal degrades. I don't know how they do that. Verizon's and Sprint's EVDO services are both pretty fast and reliable. Each are solid DSL speed, though Sprint can be quicker (especially if you're in a RevA area).

All in all, I've found that Verizon's 1xRTT on virtually no signal still works faster on web pages and whatever else than EDGE. EDGE is just plain slow; there's no way around it. We can talk theory all day long, but fact remains that from New York to Miami to Las Vegas and at least a dozen cities in between, EDGE is a joke compared with anything else.

Re:This may be true...

[Rycross]

My iPhone took 55 seconds. EDGE is good enough on sites that don't have a lot of big images... it really seems to choke on those. Sites like Kotaku bring my iPhone to its knees. But I mostly use the EDGE to do web-based MSN chat and email, and EDGE works well for those applications.

Re:This may be true...

[godefroi]

31 seconds on my HTC Apache (with a WM6 ROM and IE) on Verizon's EV-DO.

Re:This may be true...

[godefroi]

... and 19 seconds using the offices' 10mbit connection over wi-fi (HTC Apache)

Re:This may be true...

[Curmudgeonlyoldbloke]

So other GPRS / EDGE devices are loading the same page in 11-90 seconds (from the numbers above) and the fact that it's slow on your iPhone is Redhat's fault?

Misleading title?

[MoonFog]

To me, it seems the author is concerned about one thing: Web browsing. Problem is, off course, that most of the advantage from 3G comes from other services such as wathing video or video chat on the phone. In Norway, we can watch football (soccer) games over the phone, something the 3G phones handle a lot better than Edge ones. 3G network is put out there to give us these kinds of services. This article, on the other hand, only talks about TCP connections and HTTP. This isn't a case of Edge outperforming 3G in all aspects, just that it appears (he doesn't mention how he measures this) faster browsing regular HTML web pages.

Re:Misleading title?

[Calinous]

He compared it with his other phone supposedly, an Nokia E61i, and he feels the iPhone faster.
      Not much base for a scientific comparison, but I would tend to accept its conclusion (pending my own testing of the devices :D). As for high bandwidth, sequential transfers... I don't know, but you seem to have at least as good an opinion on those as the article writer has on that web speed.

Re:Misleading title?

[Professor_UNIX]

Problem is, off course, that most of the advantage from 3G comes from other services such as wathing video or video chat on the phone.

Video chat on a cell phone!? What's next, you're going to want to use it as a camcorder and send movies via SMS? I suppose you've got an instant messaging application, 3G, user-customizable ringtones, Bluetooth data transfer, and replaceable battery on your fancy-shmancy European chic phone too right? These things are superfluous on a revolutionary communications device in America.

Re:Misleading title?

[zoney_ie]

3G is rubbish here in Ireland, where people desperate for broadband have bought 3G data modems for internet access. The problem is that the system is not very scalable, and it is too expensive and slow for the operator to upgrade capacity to provide more service. The bandwidth they advertise for example (as "up to 3 Mbs") is shared for each cell - so even just two people using it solidly means half the bandwidth - but in city areas it means that the conditions can be worse than fixed-line dial-up.

From what I gather, EDGE is nice and cheap and can be more easily scaled. I believe O2 are now planning to roll it out in Ireland despite having a 3G network already.

Re:Misleading title?

[Anonymous Coward]

From what I gather, EDGE is nice and cheap and can be more easily scaled. From what *I* can gather, that's because it can be done as a fairly minor upgrade to the existing 2G infrastructure, which is well-established and (*now*) has pretty good coverage. Whether it's actually inherently better is another issue altogether- I don't know if that's what you had in mind.

I believe O2 are now planning to roll it out in Ireland despite having a 3G network already. Yes, I was quite surprised to see that (despite having invested heavily in 3G), O2 in the UK are doing the same thing, and have decided to increase EDGE support after all. This might be because of the iPhone or because of existing Blackberry users, who are apparently pushing existing EDGE installations to the limit.

Clearly then, they can get more mileage out of the existing 2G network than they are doing at present, but I'd assume there are limits beyond which they'd have to build new infrastucture and costs would make going further less attractive. And although it looks like the networks have done a volte face, I doubt they're going to abandon their long-term commitment to 3G.

Re:Misleading title?

[LarsG]

Not to mention that he only provided handwaving to back up his assertions.

He starts out with the observation that the iPhone on EDGE feels faster than an E61i on UMTS, and draws the incorrect conclusion that EDGE must be better than UMTS for web browsing. Then he goes and makes up a lot of mumbo jumbo "explanations" for why that is; I could go on a point for point debunk, but this article has already wasted enough of my limited time on this planet.

In fact, UMTS (and especially HSDPA) not only has better bandwidth than EDGE/EGPRS but also has lower latency. He didn't even bother to look up the specs.

Insert obligatory rant on the declining quality of articles posted to /. here.

Re:Misleading title?

[bernywork]

All correct, the EDGE upgrade is a software upgrade of the 2G GSM cells in most situations, but because of the size and complexity of the networks, this can be enough of a challenge for operators not to do it as for most of the time, they don't get any money out of it.

From their point of view: "Faster access for the customer, more network utilisation for us, no more money for us, costs us money as we have to run a large project to do the upgrade.... Hmmm, how about we go do something that makes more money?"

With the iPhone release and everything else, O2 basically got pushed into doing the roll out. They figured that now they have the financial incentive to do so.

EDGED out?

[LaTechTech]

So according to this; 3G is getting EDGED out?

Re:EDGED out?

[asliarun]

So according to this; 3G is getting EDGED out? Sir, unfortunately that was lame and you're losing your edge.

Apologize for being blunt.

Skip 3G for 3.5G

[jettoblack]

My personal experience in Japan over the past 5 years has shown that 3G does little or nothing to address latency issues, but 3.5G (aka HSDPA/HSUPA or together just HSPA) has made a huge breakthrough in cellular latency.

I have used data services via 2G (9600bps), PHS (32-128kbps), 3G (384kbps), and now 3.5G (3.6-12mbps). While the bandwidth has gone way up and monthly charges have gone way down, everything before 3.5G had horrible latency (400-900ms), not to mention ridiculous fees (think $20/MB or more).

Now I use a 3.5G (HSDPA) cellular data service called eMobile which sprung up just over the past few months. I get about 300KB/s (bytes not bits) down and 100ms latency, unlimited use for about $50/month. Not quite as fast as the gigabit fiber I have at home for $40/month, but it certainly works well enough for a snappy browsing experience, and WoW and FPS games are perfectly playable.

Re:Skip 3G for 3.5G

[Tony Hoyle]

Isn't most '3G' 3.5G anyway? Certainly is here... It's just marketed as the same. Certainly my N95 comes up as 3.5G wherever it is, even when I was stuck in the middle of nowhere camping..

Re:Skip 3G for 3.5G

[jrumney]

Isn't most '3G' 3.5G anyway? Certainly is here...

Maybe if here is the US, where conversion to 3G has lagged. In Europe and Japan the networks are still upgrading their old UMTS equipment, and a lot of 3G handsets were sold that are not HSDPA capable, or have that capability disabled due to inability of the network to test it properly at time of release.

Re:Skip 3G for 3.5G

[fons]

I totally agree.

Re:Skip 3G for 3.5G

[Gadzinka]

aka HSDPA/HSUPA or together just HSPA

Ah, I was eagerly waiting for the merger of those two technologies under the name HSDUPA, and the bastards decided to go with the shorter name HSPA. They've lost such a chance of the century on the name of technology that actually sounds like what it does ;)

Robert

Re:Skip 3G for 3.5G

[PopeRatzo]

the gigabit fiber I have at home for $40/month

I just bit through my coffee mug.

Re:Skip 3G for 3.5G

[mnmn]

You're only posting to incite hatred and anger in everyone here. "unlimited use for about $50/month" is like racing your Lamborghini down the street on the first gear at 1am.

Re:Skip 3G for 3.5G

[Sketch]

Sounds right to me. I use ssh on my Zaurus via bluetooth to my phone on AT&T while on the train. On most of the trip, I'm on HSPDA and everything is great. But there are a few spots where it drops down to UMTS or EDGE, and then latency is awful...makes me wish for my old Sprint PCS TDMA dialup connection. ;)

Useless article.

[NeuralAbyss]

EDGE smedge. Try using HSDPA.. it royally kicks the arse out of EDGE or UMTS. Downloading at 3mbps whilst in a train at 80km/h.. I don't think that comes anywhere close to the "user experience" of EDGE.

The article doesn't cite /any/ real-world statistics to back up the arguments.. sounds like a load of iHype to me.

Re:Useless article.

[Anonymous Coward]

Both while and whilst are ancient, though while is older. Theres no difference in meaning between them

Re:Useless article.

[Anonymous Coward]

Please Stop using this word instead. It makes us non-native English speakers without a similar word in our native language sound rude in comparison.

Tired, hungry, and thirsty non-native passenger on an airplane: "Could I have a beer?"

Stewardess: "Only if you say the magic word!"

Now also confused passenger: "WTF? Abracadabra?"

Re:Useless article.

[NeuralAbyss]

Australia.. whilst is certainly a common word over here. Provided you avoid the under-25 crowd.

And to keep it on topic.. 3G is certainly commonplace in urban settings here, as is HSDPA. If you're willing to pay an exorbitant amount, you can get HSDPA in the bush via Telstra's UMTS850 network.. but it's not cheap by any measure.

Re:Useless article.

[cthulhu11]

Try using HSDPA

With what NIC, and where? Yeah, that's what I thought.

Re:Useless article.

[NeuralAbyss]

Anywhere that's not the USA, pretty much. (read: GSM-using countries)

And as for the NIC..

USB: http://www.huawei.com/mobileweb/en/products/view.do?id=282
Expresscard: http://www.huawei.com/mobileweb/en/products/view.do?id=780
PCMCIA: http://www.huawei.com/mobileweb/en/products/view.do?id=145

And that's just one manufacturer. I'm sure you can find miniPCI versions around. UMTS is bloody good.. but expensive.

Bullshit

[RzUpAnmsCwrds]

UMTS/HSDPA can easily hit 700kbps, as can CDMA2000 1x EV-DO. EDGE hits 180kbps on a good day. On a REALLY good day.

The "error" argument is bullshit. All digital cellular technologies have extensive error correction.

Streaming media (Verizon/Sprint/AT&T all have services), downloads, and pretty much everything else benefits from more bandwidth. There is absolutely ZERO way that your browser is going to get slower because you have a faster network link, unless your browser is a piece of crap. Your browser may not get much faster if it's CPU constrained (pages don't load any faster on my 770 using the 15Mbps campus network instead of 1.5Mbps DSL), but it's certainly not going to trip the browser up or any garbage like that.

As for battery life, yes, UMTS/HSDPA takes more power. You also spend less time downloading, because it's faster.

T-Mobile doesn't have UMTS/HSDPA in the US right now, so I use EDGE every day - on my phone or on my laptop. EDGE is slow and has horrible latency. There's simply no other way to slice it.

Re:Bullshit

[clonmult]

Sure, it takes more power, but not a significant amount.

I've gone through this on two distinct 3G phones - Nokia N73 and an SE K800i. Switching 3G off on either doesn't make a significant improvement on battery life.

And the browser on the 770 is painfully slow at times! Still love it as a handy little toy though :)

Re:Bullshit

[minimum]

UMTS/HSDPA can easily hit 700kbps, as can CDMA2000 1x EV-DO. EDGE hits 180kbps on a good day. On a REALLY good day.

It all depends, how much PDH/SDH/SONET channels are given to the RBS. HSDPA can support 14Mbit as well, just be the man and transport 8xE1 to the base station controller; for EDGE, usually 3(6) 64Kbit channels are given per cell.

Re:Bullshit

[Kineticabstract]

I concur with your bullshit assessment. I've used three different phones in the past three years: the Cingular 8125 (EDGE only), the 8525 (3G when available, EDGE otherwise), and the iPhone (EDGE only, WTF?). I can tell you that, from my perspective, the 3G network is FAR more responsive, regardless of the application. It made a huge difference in my web browsing, and I miss it greatly. I don't know what this guy's on (nor do I care), but his basic point is complete bullshit, in my opinion.

Re:Bullshit

[stewbacca]

As for battery life, yes, UMTS/HSDPA takes more power. You also spend less time downloading, because it's faster.

An interesting test would be to have a list of 1,000 websites to visit, then embark with two different phones with similar battery lifespans. Record how many websites on the list you get through before the battery runs completely dead. Even though 3G takes more power, the time you save awaiting downloads probably makes up for it in "real world" usability. But then again, maybe it isn't THAT much faster that it would recoup the lost battery power. In any case, as a user, I'd pick faster network/shorter battery life every day of the week.

Re:Bullshit

[Anonymous Coward]

http://www.phonescoop.com/news/item.php?n=2462

T-Mobile's First 3G Phone On Sale ...took 'em long enough.

Re:Bullshit

[clonmult]

Unless you find that the battery doesn't even last a full day .... but otherwise I'd tend to agree with you.

Re:Bullshit

[swordfishBob]

Yup, mostly.
UMTS is arguably more power efficient, and actively manages the spreading codes and error correction ratios as required. Acknowledgements are retransmissions are incredibly fast - effectively "immediate", so packet loss is just not an argument.
HSDPA uses more power because the radio must be continually active and listening - it can receive notice of a packet at any time. When HSDPA is not active, the radio and its decoder can shut down most of the time, knowing there are specific times to listen.

Latency is also improved by structural changes in the back end of provision of UMTS and HSDPA c.f. GSM.

Give me edge any day!

[Verte]

I'm not sure if it is the same in the US, but I really miss the CDMA network [I think CDMA is EDGE, or similar]. I never had any problems with CDMA, but one thing has driven me up the wall ever since going to 3G- that I have to update the time myself. On CDMA, the clock would automatically sync. No need to worry about time zones or daylight savings, or the clock being a couple of minutes off. Always perfect time. I haven't seen any benefit to 3G that could outweigh the lack of time sync. Bandwidth is not so interesting when you can only get decent bandwidth in certain areas, and I'm pretty sure for most people here, that means in the office, where you've got a computer and don't need to use the phone for media. /rant

On topic, the point they make about phone hardware not keeping up with bandwidth is a good point too. My 680i seems to respond to key presses slower than a vt100 [it drops fast hits], so I can imagine that is where the bottleneck for html rendering is, not the bandwidth.

Re:Give me edge any day!

[Anonymous Coward]

I have a new Treo which is dual mode: it uses both the Edge and 3G networks depending on which is available / has more signal strength. In my usage experience, the 3G network definitely runs faster both in terms of latency and bandwidth.

Re:Give me edge any day!

[Mr2001]

I'm not sure if it is the same in the US, but I really miss the CDMA network [I think CDMA is EDGE, or similar]. Nope... EDGE is a GSM technology. CDMA's 3G network is EVDO, which is comparable to UMTS/HSDPA in terms of bandwidth, but AFAIK it doesn't have the latency problems that the submitter complains about with HSDPA. It's a shame that people like the submitter assume there's only one kind of 3G network.

Re:Give me edge any day!

[Verte]

Ah, I see. How confusing. Thank you.

Re:Give me edge any day!

[Andy Dodd]

Yeah, and lack of time sync sounds more like a phone issue than a network issue.

If we really wanted an apples-to-apples comparison of EDGE vs. 3G performance, we'd need a comparison done on the same phone with the same browser, for example maybe an HTC TyTn II with one of the "bandswitch" apps that allows UMTS to be forced on and off.

Re:Give me edge any day!

[rwwyatt]

The operator can send the MM_INFORMATION message on 3G/3.5G Networks, but not all operators choose to do so.

Arguing backwards

[don.g]

The iPhone is superior. The iPhone uses EDGE. Therefore, EDGE is superior.

Which is a load of crap. UMTS does need more power than 2G GSM (don't know about EDGE), and latency isn't wonderful -- but no worse than EDGE.

Radio protocols designed to run IP (even WiFi) have forward error correction (i.e. ability to cope with noise) to reduce dropped packets and thus keep TCP happy.

Why are blog posts of people who don't know what they're talking about ending up on the slashdot front page?

Re:Arguing backwards

[cropus]

Even the basic 3G _has_ forward error correction - even two types of it. I believe all modern radio systems use it.

BTW, TCP retransmissions are poison to radio networks. That's why HARQ is used in HSDPA and upcoming standards to prevent TCP retransmissions i.e. if the packet didn't arrive correctly, the PHY layer handles the retransmission (not even the complete TCP packet but just the invalid part to make it simple).

I think bad network coverage is not a reason to rant a radio system. Complain to the network operator . Please understad that it takes time and money to build the network. It may even be that the network gear manufacturers are not able to deliver the needed equipment.

Check out 3GLTE for good Interweb experience, it is coming out in couple of years (www.3gpp.org).

Re:Arguing backwards

[Anonymous Coward]

I have to post anonymously because I am working for one of the companies producing network equipment and mobile phones. But I can state two things:

• Contrary to what the author of the article claims or assumes, the latency in many UMTS networks is significantly lower than in EDGE. One of the goals of 3G design was to reduce the latency, which had been identified several years ago as one of the bottlenecks for a good web experience (because of TCP's three-way handshake and slow-start). The author is just speculating about the latency but does not give any real statistics. My experience is exactly the opposite: UTMS round-trip time can be more than 40ms faster than GPRS/EDGE.
• The comparison with the Nokia E61i only shows one thing: this Nokia phone is not very good at 3G, both for speed (bandwidth + round-trip time) and battery lifetime. But it would be interesting to run the same comparison against better phones from SonyEricsson, Samsung or other manufacturers. Or even other models from Nokia.

Re:Arguing backwards

[intheshelter]

"Why are blog posts of people who don't know what they're talking about ending up on the slashdot front page?"

Because if this weren't allowed on Slashdot then NO ONE could post anything.

Re:Arguing backwards

[AceJohnny]

Why are blog posts of people who don't know what they're talking about ending up on the slashdot front page?

To provoke discussion within the (sometimes) better informed slashdot crowd.

With these kind of articles, I regularly read up the comments before the article itself, and most of the time I get a better picture of reality through a couple of highly-moderated comments.

Though for every good +5 comment, I find two crap +5 comments.

So which one's yours? ;)

Re:Arguing backwards

[stewbacca]

why are blog posts of people who don't know what they're talking about ending up on the slashdot front page?

a) because anything about Apple, AT&T and iAnything is guaranteed to create lots of traffic for slashdot
b) "IF" the story claims are true (doubtful), the findings go against common knowledge, which makes for good conversation
c) slow news day?

Can't we use both?

[Procasinator]

Don't 3G phones fall back to 2G (GPRS) when they can't get a 3G signal? Would it not be possible offer EDGE (sometimes refered to as EGPRS) and 3G, and let the user decide on a case-by-case basis (if they want to; don't force them to) which one they want to use.

I mean, if I want to view a simple webpage, I could use EDGE. If I want to download a song or a video, then 3G would be the better option.

Re:Can't we use both?

[notgm]

someone would have to build and sell a phone that could use GSM and CDMA, and I don't think anyone in the US is willing to sell a product like that.

bear in mind that an EDGE network in New York is not an EDGE network in rural Missouri, and a 3G network in LA is not a 3G network in Idaho - not every carrier on every site has been built out to full capacity. for the network that will work best *WHERE YOU LIVE*, you need to do some real world testing. bring a laptop with you to the store, ask for a demo. you'd be surprised how many stores don't have optimal coverage.

i've been on an edge network in arkansas that rocked a 3g network in illinois, and i've been on a 3g network in north carolina that rocked an edge network in texas.

edge != edge != edge.

just like realty, location, location, location.

Re:Can't we use both?

[lucas teh geek]

someone would have to build and sell a phone that could use GSM and CDMA, and I don't think anyone in the US is willing to sell a product like that.

where does CDMA come into it? CDMA sure as hell isn't a 3G technology

Re:Can't we use both?

[Procasinator]

Valid point. I'm talking from a European stance myself (personally in Ireland). We are lucky enough to have a fairly decent 3G and EDGE networks throughout alot of the continent. Why should we have to suffer EDGE on the iPhone when we have infrastructure for more? I think it is a bad decision to release the iPhone in the UK without support for both. The US infrastructure is silly. That's why Nokia has little market dominance over there - their too far ahead.

Re:Can't we use both?

[wikinerd]

just like realty, location, location, location.

Correct: Im in EU and I have connections with all 3G (1.8-3.6 mbps) operators active in my country, because not a single of them is 100% satisfactory in coverage, and there are many places where no one of them covers at all. As an independent professional, without 3G coverage my performance may suffer, so I have to use multiple carriers to do my work.

Re:Can't we use both?

[Tony Hoyle]

In the UK only one provider (O2) has Edge *at all* and they installed it so they could sell the iphone. And that covers by their own estimates only about 20% of the country. I'm not sure how that translates to 'decent coverage'.

The iphone is effectively limited to GPRS acrosss most of the country even on O2 which sucks the big one, and could mean its ultimate failure here (since wifi is as rare as hens teeth too).

Re:Can't we use both?

[Anonymous Coward]

You should look at, say, an AT&T 8525 (mentioned because I have one). It will switch between GSM and WCDMA (UMTS/HSDPA) with a little 3rd party software app. I leave it on EDGE 90% of the time, since for email push it's much more battery efficient, but I switch to 3G (HSDPA) whenever I need a lot of data, and my ping times drop from 250ms average on EDGE to 90ish on HSDPA, not to mention the drastically increased bandwidth.

Re:Can't we use both?

[notgm]

while this is true, you'd be hard pressed to find a (US) nationwide carrier serving up a 3g network on anything other than CDMA.

Re:Can't we use both?

[Procasinator]

In the UK only one provider (O2) has Edge *at all*

I was under the impression Orange did too? (though probably very little coverage)

Either way, a lot of Europe has EDGE too (well, at least reportedly), and they should have 3G with fairly seamless degradation to EDGE, and when necessary, a further degradation to standard run-of-the-mill GPRS.

Re:Can't we use both?

[Procasinator]

I've just realised I've made myself look dumb by linking to an Orange page about using EDGE abroad.

Nevertheless, they still seem to have EDGE support.
http://www.3g.co.uk/PR/Feb2006/2548.htm

Re:Can't we use both?

[Andy Dodd]

Sure it is. CDMA2000 1xEV-DO is 3G.

That said, there's now a lot of confusion between CDMA the modulation scheme and cdmaOne/CDMA2000 (Qualcomm's protocol suites based on CDMA modulation). If you are using a 3G network, you are using CDMA modulation, regardless of whether you are using UMTS (the protocol suite for 3G GSM, which moved from the GMSK modulation used for 2G/2.5G GSM networks to a CDMA based modulation scheme), or CDMA2000 1xEV-DO (Qualcomm's 3G protocol suite.)

Re:Can't we use both?

[clonmult]

Not sure if's that easy to be selective, but it was pretty cool to find my venerable (!) Nokia N73 dropping to EDGE when there was no 3G signal (which wasn't that often).

That would have been an alternative for apple - drop in a chipset that supports EDGE and 3G, and possibly give the user the option of which one they prefered (kinda like a battery life/net speed slider).

Re:Can't we use both?

[Mwongozi]

Orange UK has supported EDGE for years. They also support HSCSD. (Remember that?)

Orange used to be quite the innovative network, back in the day. They jumped the shark when they killed off Wildfire.

Re:Can't we use both?

[tengwar]

[I work for one of the major mobile telcos in a group that deals with mobile data clients]

Yes, it's perfectly possible to get the device to use EDGE as well as 3G and many devices do that. However very few networks upgraded their GSM equipment to use EDGE as they went straight to 3G so in practice I've only ever seen this when I visited the USA and roamed on to the Cingular network.

While I don't have the figures, as far as I know GPRS/EGPRS (i.e. EDGE GPRS) has much higher latency than UMTS, HSDPA or HSUPA. The reason is that it uses time slicing to get multiple user on to a single frequency, whereas WCDMA (the radio technology that UMTS et al. are based on) allows simultaneous transmission. The max throughput is hugely different: for instance I have a pretty common handset that theoretically does 3.6Mb/s, and I get something close to that in practice. 7.2Mb/s is coming soon, if not already out (I'm not up to date with the network plans). The only area that the older technologies are better on is power consumption: there's a big difference on this, but if you need to eke out your battery, most handsets will allow you to select GSM/GPRS only.

Latency my arse

[neokushan]

The only real comparison this author makes between EDGE and 3G is the latency. Everything else is just speculation (for example, he claims that the phone will be slower on 3G because it has to process more....I want whatever it is he's smoking), but right now I'm on a connection that has an excellent latency (think 50ms for most European severs) yet when I try to play Halo 3 on live, I get so much lag that entire rounds pass without me even knowing, all because I don't have enough bandwidth.

I wish someone would sponsor me to write some FUD, it seems like a cushy job.

Has the author of TFA even used 3G?

[MikeyVB]

Because here in the Netherlands I recently got a 3G phone (Sony Ericsson W880i) and included with my subscription is ~9 channels of televsion. My TV is streamed over the 3G connection, and only buffers for about 3 seconds when I switch channels, with stereo sound too. No artifacts or funny business even with low signal strength, nor switching between cell towers (I only use the TV when travelling to and from work on the train)

Also in my subscription is a couple of free songs that I can download using the 3G. I have any downloaded song within a minute. Web browsing (on Opera Mini, with HTML and NOT mobile pages) feels nearly as fast as my computer at home. Can EDGE, at only 0.2 Mb do that????

Of course, maybe it does, as I have never used EDGE, but at least would try BOTH technologies before I claim one is better than the other.

Re:Has the author of TFA even used 3G?

[d3vi1]

The author also seems to forget that 3G is not only Data calls. 3G is video Calls, high bandwidth for both upload and download, data sessions NOT interrupted by voice calls, IPv6, Mobile Radio and TV (as mentioned above).
From my point of view, the fact that the Apple iPhone uses the inefficient EDGE network, while the world (except the US obviously) uses 3G, and more recently WiMAX. In Romania, while Orange and Vodafone said they will introduce the WiMAX networks later this year, using a WiMAX enabled device you can already see their networks in the 3.5/3.75 GHz spectrum.
I actually had high hopes from the iPhone, and after it came out it simply disappointed me. My next device will be the Nokia Nx00 with WiMAX, that I hope is coming out Q108.

Re:Has the author of TFA even used 3G?

[fsmunoz]

The article is a post-factum justification for the iPhone, nothing more. If the iPhone uses EDGE, then OMG I'll find some way to make EDGE better! You want TV channels in your phone, as by now common in Europe? I'm sure the guy will say that since the iPhone doesn't have it seeing TV on a phone is actually a Bad Thing that nobody should want, it's actually a "feature": "iPhone: now with No Eye Stress advanced technology, protects the eyes by deflecting rapid changing images in stream form!".

It's pathetic really: if people like the iPhone, by all means, buy it, but making the obvious shortcomings it has and spin it around to make them virtues is an act of desperation. This "EDGE is better" is just one more of those; thankfully the iPhone doesn't use GSM alone, or we would be hearing how less power demanding it is!

Real world conditions.

[jsiren]

What I want to know if $MOBILE_DATA_PROTOCOL is still usable once the train is doing 200 km/h in the middle of nowhere.

(The good thing about 200 km/h is that the tunnels around here don't last long enough for connections to time out...)

Re:Real world conditions.

[Anonymous Coward]

With increasing frequency, tunnels have perforated coax installed for cellular reception.

Re:Real world conditions.

[itsdapead]

What I want to know if $MOBILE_DATA_PROTOCOL is still usable once the train is doing 200 km/h in the middle of nowhere.

I'm from the UK you insensitive clod!

"Real world conditions" is that the train is stopped in the middle of nowhere because the rail system is being run at 150% capacity and if one train has to slow down (because, e.g. some slippery leaves have fallen on the track; its a bit windy; its a bit sunny; we've had the "wrong type of snow" or the embankment has collapsed because the cut down all the trees to stop the leaves falling on the line...) the entire rail system gridlocks (usually shortly followed by the mobile phone system collapsing as everybody tries to phone home).

OK, that's slightly cynical, but...

reminds me of DSL vs Cable broadband claims

[Anonymous Coward]

On paper DSL should be faster too.
In reality, it's not in virtually all markets.

Rubbish!

[pandrijeczko]

There is no way that U2's little guitarist bloke can outperform the likes of Joe Satriani, John Petrucci & Paul Gilbert!

Re:Rubbish!

[stewbacca]

Ooh, nice analogy! But of course, you have to include a car reference since this is slashdot. The Ford EDGE outperforms all the other Crossover vehicles in it's segment, even though it has no alliance to AT&T or Apple.

Yearly 'better' standard

[nurb432]

Well, i guess it keeps people buying new stuff each year to avoid the fear of having an obsolete brick. Since we all know they cant get us wth genuinely better products.

If they could build in a 'obsolescence' function where consumer electronics would just self destrcut after so many hours of use they would. The manufacturers are a victim of their own success in cost cutting and reliability.

Lovely

[ArAgost]

I like the "High bandwidth radio networks are more error-prone" part. It's so full of evidence, data, and accurate reasoning that you just can't help but agree with it.

hmmmm...has he used it

[munkey_bwy]

ive just switched over from a 3G based phone that i managed to get data connections all over the world to an EDGE based phone of a more prolific carrier. Guess what, i very rarely get a data connection on edge, even in my office which is in the UK's second largest city and if i do get a possible data connection it get refused based on quality. IS EDGE REALY THAT GOOD?

Whatever! EDGE Sucks, 3G Rocks

[Anonymous Coward]

Whatever.

I've been a long time user ... daily for both of these technologies. EDGE for over a year and 3G for 6 months before my circumstances changed and I didn't need mobility any longer.

EDGE was next to worthless for PC use.
3G was good enough to **replace** my high speed connection at home. Yes, it was that fast. I used VPN a few times a week for 10+ hours - no issues. The only reason I didn't drop my broadband services was because I run a tiny server out of my house connection and all email for my domain from there.

EDGE is ok for mobile-only sites with extremely low bandwidth needs, but even checking the local traffic maps is painful. ANY graphics are too much for that network. My Blackberry is still on EDGE - -I know what I'm talking about folks.

3G really is THAT good - at least in my experience traveling in major cities. Both have drop off points and your mileage may vary for your hardware, drivers and location. The closer your mobile devices are to a real computer, the more important 3G or better connectivity becomes, regardless of what apple may claim.

At least one glaring incorrent point

[marvinglenn]

FTFA: "Power consumption of any chip increases according to the frequency squared."

Wrong. The power consumption is proportional to the square of the voltage, not the frequency. If you double the frequency, you only double the current, not quadruple it.

Other points in TFA may be correct. I don't know.

IAAEE. (I am an Electrical Engineer.)

Re:At least one glaring incorrent point

[jm1234567890]

But if you double the current don't you quadruple the power? P = I^2 * R

Re:At least one glaring incorrent point

[Aaron Isotton]

I'm not an electrical engineer, but I have a little bit of EE background.

As the parent post says, the statement "Power consumption of any chip increases according to the frequency squared." is obviously wrong.

I believe that the author of the article misunderstood the "rule of thumb" for CMOS circuit power consumption, which states the following (from Wikipedia and other sources):

Power = Voltage^2 * Frequency

This obviously applies to the *same* chip only, i.e. if you run the *same* chip at half the clock rate it will use roughly half as much power. Concluding that a 10 MHz chip will consume 10 times more power than a different 1 MHz chip is obviously wrong.

I definitely recall that in some EE lectures (related to real-time scheduling in low-power systems) we were told that the power consumption of a CMOS circuit is proportional to the *square* of the frequency, which happens to be "more similar" to what the article claims. Can anyone comment? Is this related to the fact that you may decrease the voltage when you are decreasing the frequency?

Re:At least one glaring incorrent point

[Andy Dodd]

In addition, the limiting factor in terms of operational (i.e. transferring data) battery life of a phone is almost always not the CPU, but the radio. Regardless of whether the signal processing scheme uses any more power, CDMA-based modulation schemes (such as UMTS - yes, 3G/3.5G GSM uses a CDMA modulation scheme, even though it is not part of Qualcomm's CDMA2000/cdmaOne protocol suite.) require a linear (inefficient) transmit power amplifier, while the GMSK-based modulation scheme used for GSM/EDGE could use a nonlinear (efficient) power amplifier. Early UMTS handsets had AWFUL battery life due to this and handset manufacturer's complete lack of experience with CDMA modulation schemes. Since then, Qualcomm has gotten involved with UMTS and is now one of the most common suppliers of UMTS mobile chipsets (in addition to chipsets for their own protocol suite) which perform much better than earlier UMTS chipsets due to Qualcomm's years of experience with CDMA. That said, there seem to be some underlying flaws in UMTS (likely a higher peak-to-average ratio of the signal) that make it use more battery life than Qualcomm's own CDMA-based suites.

Re:At least one glaring incorrent point

[IkeTo]

> The power consumption is proportional to the square of the voltage, not the frequency.

I suppose the "frequency" here means the frequency of the clock of the chip, rather than the frequency of the channel (the two should be quite strongly related, but not necessarily proportional).

The above is not exactly correct. To understand this, one needs to know why the chip consumes (most of the) power. It consumes power when its transistors switch. When they don't switch, they are at a high impedance state, so nearly no current flow through (and thus consumes nearly no power) although there is a voltage. When they do switch, a current flows for a brief period of time, and of course this leads to energy consumption following V square over R. But since this happens only for a brief moment we need to know the total length of period of time this happens in order to know the total energy consumption. And it isn't hard to see that this is proportional to the frequency f. So the energy consumption is actually proportional to f times squared V.

But this is not the whole story. In order to support a higher frequency, the chip requires a higher voltage (in order to force the states of the transistors to change quickly enough); and the dependency is approximately proportional. So it means that if you are a designer of a chip/product and can adjust both f and V, or if you use a chip which can adjust the voltage dynamically depending on the frequency, the power consumption varies approximately as cube of f (or equivalently cube of V).

So running fast is actually very expensive!

IANAEE. Instead I'm a computer scientist doing research on energy efficiency.

Re:At least one glaring incorrent point

[amram9999]

This article makes a lot of claims and doesn't back them up.

3G phones have better spectral efficiency than 2G phones, so they transmit less power for a given amount of data. For a given web page, this means a 3G phone will consume less power if all other parts of the system are equivalent. However most 3G phones have faster processors which offset this to some extent. The peak currents in a 3G phone will be higher because it can transmit data faster, and this means you can drain your battery faster if you are using it extensively.

Another important fact is that actual bandwidths can be far less than advertised ones due to the error correction used in both EDGE and 3G networks. If you are far from the base station it will reduce your effective bandwidth until your bit error rate is acceptably low. Thus real world EDGE speeds are less than 200 kbps.

EDGE also supports multislot which uses several time slots to increase speeds. Some networks might support class 8 EDGE while other support class 12. For a better explanation look at http://www.gsmworld.com/technology/gprs/class.shtml. The article talks about GPRS but applies to EDGE as well.

IAAEE as well and I work on cell phones. Please don't just accept rubish claims without any factual evidence.

Re:At least one glaring incorrent point

[lexiconbt]

IAAEE (and working in power management), and I can't believe I'm agreeing with the CS guy. The original author's comments are valid. The efficiency of such com chips is getting quite better as far as efficiency goes over the years. For example, the energy per bit, on your RS232 serial port is quite a bit higher than energy/bit of a new Ethernet MAC/PHY. This is a (good) assumption, however. Anybody know the actual data on clock/power requirements of EDGE vs. 3G to put some real data into this discussion?

Re:At least one glaring incorrent point

[shabushabu]

No. Typically current increases *because* resistance decreases (Larger transistors, for instance). Usually it is the voltage that remains constant. Therefore P = V*I is the relevant formula here.

I have a 3G phone

[EmagGeek]

And it's ALWAYS faster on an EDGE network than on a "3G" network, unless I am tethering. But, on the phone, TFA has merit. It's unfortunate that the SYNC is such an idiot's phone, with no user capability to choose network, cell, or whatever. I get coverage from two cells at my house and the phone is always stuck on the weaker one. With my old MPX220, I could force-select the stronger one. Ugh...

Speed at a cost

[sfm]

Both EDGE and G3 are marketed based on speed, but this comes at a cost. I wonder if anyone has done the comparison on the basis of $/bit and included the lower speed "low cost" networks as well?

Re:Speed at a cost

[Tony Hoyle]

Cost is the same, you pay by the MB.. whereas you could argue that since Edge is about 10* slower than HSPDA it's 10* cheaper because you could only transfer 1/10th of the data.. that's be a silly argument.

Re:Speed at a cost

[Zelos]

I'd think that most people who are going to use mobile web will have an unlimited plan.

I'm not sure you can even buy an iPhone without one.

Re:Speed at a cost

[Andy Dodd]

The OP is probably from Europe, where unlimited data plans did not exist until the iPhone was rolled out in the UK. (On the other hand, their pay-as-you-go data was MUCH cheaper than what U.S. carriers charge for pay-as-you-go data.)

In general, it appears to me that Europe has more of a pay-as-you-go approach than the U.S. "flat rate monthly contract" approach, but European pay-as-you-go prices are MUCH more reasonable than U.S. pay-as-you-go prices for both voice and data.

Re:Speed at a cost

[marsu_k]

The OP is probably from Europe, where unlimited data plans did not exist until the iPhone was rolled out in the UK. Bullshit. I had unlimited GPRS data in Finland in 2002, it was about 20e/month. Expensive by todays standards, granted. Now you can get unlimited 3G data for 10e/month (max. ~380 kbps), or if you have HSDPA, 20e/month (~1 Mbps) or 30e/month (~2 Mbps).

(for some reason Slashdot seems to reject euro symbols?)

Re:Speed at a cost

[Andy Dodd]

Oops, I guess I was wrong. I recall that the addition of unlimited data plans for the iPhone in the UK seemed like really big news.

Re:Speed at a cost

[Zelos]

It's changed recently in the UK, about 12 months ago unlimited data plans were a little hard to find, now they're everywhere.

Edge is ok for simple things

[crunzh]

I have a phone that supports both EDGE and UTMS (Nokia N73) and edge is fine for checking text email and surfing mobile friendly websites but if I want to download stuff, surf large webpages or use the phone as a modem for my laptop edge makes it allmost unbearable but UTMS it is perfectly useable.

Some Practical Experience

[SwashbucklingCowboy]

Having used both EDGE and 3G (I'm posting this over an EDGE connection), I can say with great certainty that 3G beats the crap out of EDGE.

Re:Some Practical Experience

[vjouppi]

Same here. I use a lot of PuTTY for SSH'ing to my home box from my phone (Nokia E90) and I can safely say that the GPRS/EDGE latency is absolutely terrible when using a terminal application.

3G/HSDPA has noticably lower latency and editing text over SSH is nowhere near the frustration it was with GPRS/EDGE.

But yeah, I'm sure the article author's iPhone is great and does everything he wants.

ugh...

[Anonymous Coward]

I wish I could digg this story down. Typical Apple fanboy comments with no real evidence to back-up their claims. Latency on EDGE is almost a full second in a typical ping.

Re: Strange reasonning

[neutrino38]

Good point for the processor issue.

I don't get the argument that high speed networks are more error prone. Following this rationale, this would mean that our broadband or wifi network would be much less reliable than a good old modem.

Also, I fail to see why EDGE technlogy would be less prone to radio interference and reverberation than UMTS. COuld the author please provide pointers of more factual articles about this point?

What is certain is that 3G has an advantage over GPRS as the later uses time slots that are unsued by voice call to transmit packets. Therefore, this technology was more or less piggybacking on the GSM system.

I don't know if EDGE retains the same principle but for sure, in 3G, data is now a first class citizen. This means that you got a good chance to retain your data connection if the nework becomes overloaded.

I have a razr v3xx (3g phone)

[majortom1981]

I have a motorola razor v3xx wich is 3g (the 3.6mbps version) and I can select the networks manually. I get faster speeds on edge then I do on 3g. I don't know why but I do. I think its because att doesnt really care about 3g.

Re:I have a razr v3xx (3g phone)

[Seakip18]

Second that. I'm in an area where they are building a 3g network. It sucks since I'm constantly switching between EDGE and 3G. I can't stream anything since I lose the connection to my server as the networks change over. But in my own highly unscientific experiment, 3G beat the crap outta EDGE with Google Maps. Half the loading time at least.

Re:I have a razr v3xx (3g phone)

[HeavyDevelopment]

I think that is more of a function of the fact that ATT doesn't have great coverage with their 3G network yet. This is probably why Apple went with EDGE because the coverage ATT has with EDGE is really extensive. ATT invested a lot of money early with the 2G/EDGE network. I have used 3G with ATT (Blackberry 8100) and it was faster....often but not always. Then again I live in a major metro area. In a rural area I imagine it wouldn't be that great. I had a 3G wireless card with Verizon that was awesome in terms of speed. Verizon a later player in the game invested more in 3G and has excellent coverage even outside of major metro areas (I was able to get 3G wireless at Mammoth Mountain ski resort for example). No latency issues at all. I used it like a regular broadband connection and canceled my landline service. Well that was until Verizon branded me a pirate and terminated my account. Apparently in their contract you can not upload and download files (of which as a web developer I do all the time) nor can you stream anything (which I was doing with internet radio often during the day), or at least that is how the security and fraud "specialist" explained it to me. Things may have changed since then, but if you can't upload and download files and/or stream media what's the point to having broadband wireless?

3G is faster

[Anonymous Coward]

Edge is like a 24.4 dial-up in my experience. Sometimes 3G isn't a lot better. Truthfully, my AT&T card is laughable as a "broadband" device.

The latency argument is bogus

[Anonymous Coward]

The latency argument is bogus. UMTS and especially HSDPA have lower latencies than EDGE. EDGE is depending on the "old" 2G framework and will never have better latencies.
HSDPA latency is around 100ms. GPRS is on 500ms, and I expect EDGE (we don't have it in the Netherlands) to be slightly better than GPRS. With new network sw releases the latency of both technologies tend to go down every year (GPRS started on 1100 ms in 2001)

Also the availability is depending a lot on whether you live in the US or Europe. In Europe, HSDPA is very well established and as far as I know only Germany(T-mobile) has a real EDGE capable GSM network or will have it soon.

From personal experience I can say that HSDPA is great. Througputs upto 1Mbps are very common.
On a laptop it almost feels like a fixed line connection when browsing and emailing. On a phone it is less of an issue, unless you are downloading youtube movies.

I accept the battery and power argument though. This is a real issue, that keeps people from switching to UMTS, and in my opinion the real reason Apple chose EDGE.

Regards,

Bert

Streaming

[Doc Ruby]

What I want out of my mobile "phone" is streaming. Streaming audio/video, mainly audio: Internet radio. For low-latency, the basic voice protocol is what's important. Streaming doesn't need low latency, or any of those other EDGE features - it needs the higher bandwidth of 3G. And though it consumes power, that's consequent of any app that that delivers steady content to my terminal.

Re:Streaming

[WhiteWolf666]

Correction.

EDGE doesn't have low latency. EDGE is a high latency protocol. EDGE typically has latencies over 1000 ms when the connection is loaded.

There is no way in which 3G is inferior to EDGE. This is someone's attempt at implementing the proprietary Steve Jobs Reality Distortion Field.

3G > EDGE, for everything (Voice, Data, Streaming, Interactive, Non-interactive, Downloads, Porn, Whatever).

Re:Streaming

[CompMD]

Get an EVDO phone. I can listen to my local collegetown indy radio station anywhere in the US on my phone as a 256Kbps MP3 stream, and I never have any problems with it.

Re:Streaming

[Doc Ruby]

Do you get bursty bandwidth, where your 256Kbps rate is only an average over longer times, longer than a 30s (1MB) buffer? So your buffer runs out before it can get a blast of the latent packets?

When last I tried EVDO, on a Sprint Treo a couple of years ago, it could deliver only about 100Kbps average, but would alternate between 40Kbps and 160Kbps in about 45-60s peaks/troughs.

Re:Streaming

[CompMD]

Nope, its continuous. I never had the music stop to buffer as long as I had an EVDO signal. I've consistently been able to pull down close to 1Mbps on Sprint, and peak speeds of 2Mbps. I used the field debugger to verify this. I use an HTC Apache (PPC6700), its by far the best phone I've ever used. Tethering it to my laptop I was able to download files at ~100KB/sec (not kbps). According to the laptop I was connected at 921.6Kbps, but I don't know how much I trust that. The quality of the EVDO service definitely varies by location; I've had the best luck in Chicago and Kansas City (Sprint's home town). Even the 1xRTT speeds on my old PPC6600 in the KC area were respectable (~80-100kpbs).

Mind you, I'm not a Sprint fanboy; as far as businesses go, I hate their practices, but I cannot say anything bad about their technology.

I smell the comeback of an old Apple ad...

[lonesome_coder]

I was making a call on my 3G phone, and it was like bleep bleep bleep...and I was like...ugh.

Question: But, next-gen iPhone will have 3G?!

[intmanofmystery]

This guy probably owns an iPhone or works for Apple! Please prove me wrong.

It is interesting to note that Steve Jobs has announced that the next generation of the iPhone will come with 3Gtechnology. This article is real nonsense since we can now only assume that even Steve Jobs agrees that the iPhone should come with 3G.

Please, nothing to see here, move alone.

Re:Question: But, next-gen iPhone will have 3G?!

[stewbacca]

I own an iPhone (but don't work for Apple) and EDGE sucks. Does that make me wrong? hehe. But seriously, I've never used 3G, but it can't possibly be any worse than trying to surf the web via the EDGE network in downtown San Antonio (AT&T Corporate Headquarters for cryin' out loud). EDGE is what keeps the iPhone as 4/5 stars instead of 5/5. It "almost" pushes it 3/5, but the phone is so greate inspite of the AT&T limitations, that I can't bring myself to do that. But then again EDGE is THAT bad, so maybe 3/5 is more realistic.

3.5G is fine

[SirJorgelOfBorgel]

I have a HTC TyTN II which has 3.5G and it's wonderful. Ping is usually around 80ms, vs the 30ms I get on my landline, for servers on the AIX. Speed is incredible. No, I've never used EDGE (they pretty much went straight from GPRS to UMTS here, skipping EDGE). The point is, there is nothing wrong with 3.5G, the latency is more than good enough for usage in a phone. If you would use it as data connection for your PC, it might be less than perfect, but I doubt you could play an FPS better over EDGE with it's slow ass connection speed, even if the ping is a bit lower :) In other words, the dudes point are irrelevant, even if they were true :)

Why?

[Anonymous Coward]

Why is Kdawson such an apple fanoboy? Why?

What a load of baloney

[Anonymous Coward]

Currently I can get a 1.7Mbit service via HSDPA/UTMS service with network upgrades will soon provide 3.4MBit then 7.4Mbit.

Only in a country with crappy 3/3.5g infrastructure would be promoting EDGE as a superior wireless platform.

HSDPA is heaven

[wikinerd]

I have used both EGPRS (2.75G 236kbps) and HSDPA (3.5G 3.6mbps and 1.8mbps), as well as plain GPRS (2.5G 53kbps) and UMTS (3G 384kbps), and according to my personal subjective observations: GPRS sucks big time even for browsing, EGPRS is not very different than UMTS in terms of speed but appears to have lower latency, UMTS really sucks because of too much latency, and HSDPA is heaven, as it has much lower latency than UMTS and much higher bandwidth.

In plain user's terms, according to my experience: With GPRS I can read some pages specially made for mobile devices (eg WAP) and I actually do use it sometimes to quickly read some BBC or other news on my phone while I'm standing in a bus, etc. But when I get only GPRS signal on my laptop then I cannot really do anything except some SSH. I have used EGPRS only briefly, but I can say it's satisfactory both for browsing and for SSH, but not for downloading or uploading. UMTS is not very satisfactory for SSH (high latency), but downloading is so-so (uploading still not good), and Web browsing is usually ok. HSDPA is perfect, as it is very good at SSH (lower latency than UMTS) and Web browsing, and also very good at downloading and uploading as well: You can actually be in the middle of the sea on a ship and transfer all your server backups or download a GNU/Linux distro and burn it while you are on an island or a mountain - provided there is coverage and you have enough batteries with you in your backpack or trolley. You can even use a 3G router to connect your LAN to the mobile network as a backup in case your DSL fails.

I actually many times work out of my home office thanks to mobile networks. I pack a laptop and lots of batteries in a backpack or convertible trolley bag, get a ship, and go to explore various islands while working over the 3G connection. I have even mapped the most significant network blackspots in my usual destinations so that I can avoid them. This mobile lifestyle wouldn't be possible without 3G.

Re:HSDPA is heaven

[TooMuchToDo]

You can actually be in the middle of the sea on a ship and transfer all your server backups or download a GNU/Linux distro and burn it while you are on an island or a mountain - provided there is coverage and you have enough batteries with you in your backpack or trolley.

Correction: You can do all of the above, even at sea, as long as you're in range of a cell tower on an island. Also, your speed will be limited by the backhaul connection the tower has available to it.

Re:HSDPA is heaven

[wikinerd]

Correction: You can do all of the above, even at sea, as long as you're in range of a cell tower on an island. Also, your speed will be limited by the backhaul connection the tower has available to it.

Yes, of course. No tower, no signal. This is a problem for places where density is low and there are too large bodies of sea with no land in between... but here my local geography is quite the opposite: Generally satisfactory tower density with thousands of islands spread everywhere :)

Tuesday

[Zebra_X]

From the post...

But the question left unasked as been, "Does 3G really improve the user experience dramatically?" Most pundits would reply...

Well, given that the author seems to have little or no experience with day to day 3G use I'd say this might be a case of the pot calling the kettle black.

Yes, wireless is as much about latency as it is bandwidth. Broadband wireless latency is not as good as it should be.

Here are some general real world metrics for 3G: latency ~ 300ms Download: 800 - 900 Kb/s Upload: 200 - 400 Kb/s data transfer from 100 - 200 KB/s

Typcial advertised connect speed: 3.6 Mbps

Given these stats the following tasks are highly useable: Web, IM, Streaming music or video, Skype Video, check e-mail, download larger files quickly.

EDGE allows for a reasonable user experience for e-mail and IM but that is about it. File transfer speeds are abysmal and downloading web sites is also pretty poor. It seems most sites have decided that it's OK to have a 300k home page. Streaming music is far from reliable and video is out of the question (although you tube has been able to decimate their clips enough for iPhone use).

His point about phones not being equipped to handle the additional data and subsequenct processing is well made. This is probably true for the vast majority of phones. In the context of the iPhone however, I think there are major benefits to having 3G. Further, defending Apple's decision is probably not the right way to go on this. The iPhone is the only phone on the market truly capable of exploiting a 3G connection - but lesser phones such as the nokia mentioned provide a faster connection. In addition, that poorly performing nokia phone will do just fine as a mobile broadband bridge.

The bottom line is that especially for laptop users, and a lesser extent phone users, 3G is a huge step up. It's as fast as a shared public wi-fi connection, and easier to access. The major difference is that an access point is not needed. 3G is a very liberating experience lol

Re:Tuesday

[stewbacca]

video is out of the question (although you tube has been able to decimate their clips enough for iPhone use).

Well, which is it then? You can't say it isn't possible, then qualify that with saying it is possible as long as it is crappy. For the record, YouTube videos look better on the iPhone, because the limited resolution hides the massive amounts of artificats present on a 20" (and bigger) screens.

Re:Tuesday

[Zebra_X]

I didn't say that it wasn't possible. It is - but the end user experience is not that good. I was lumping streaming audio and video together.

YouTube is more of a file download than "streaming" in the truest sense of the word.

Re:Tuesday

[WhiteWolf666]

I think you may even be underselling 3G.

Here, with either my EVDO Rev0 phone, or EVDO RevA USB card, I get latencies between 50-150 ms. Downloads easily in the 1500 kbps, and uploads in the 1000 kbps range.

In downtown chicago, both go nearer to 600 kbps; but the latency is still excellent, and everything is vastly, vastly better than EDGE.

Re:Tuesday

[stewbacca]

As much as I hate EDGE for surfing, I have to admit that the YouTube videos play nearly as soon as you click it. Not sure how they do it, given how awful edge is for just displaying an html page.

Re:Tuesday

[Zebra_X]

To comprare Skype is able to use ~ 32 - 46 kbps for sending full motion video @ 15 fps. Using h.264 one should see similar capabilities. While I'm not 100% certain the, iPhone can use the highest compression methods for h.264 because it's video decompression is all hardware based. Given the screen sizes and fudging with the frame rates you can get a pretty good experience. 32 - 46 kpbs is well within reach for EDGE - it's just that you need dedicated hardware to achieve this to get decent battery life. It's also a lot easier to make this scenario work for on demand content - than live streaming video.

Re:Tuesday

[Zebra_X]

Better to understate than overstate :) Especially

Over all, it's pretty awesome to be anywhere and be able to download a 20 mb file in a reasonable amount of time.

It's pretty amazing how far we've come in 10 years.

He forgot "YMMV"

[Thumper_SVX]

The problem with cellular data technologies is that a lot depends on the implementation. There are so many factors involved in the reliability, latency and bandwidth available (not least of which is interference on particular wavelengths) that unless you're doing a controlled laboratory test with your own controlled equipment, then quite frankly the results you get are going to be incredibly varied.

My suspicion is that the author of the article tested in a location where interference on the EDGE "channels" is low, and 3G channels suffered a much higher error rate because of interference. I've seen the same thing myself... though I'm lucky to live in an area where the 3G coverage is excellent and error rates are relatively low.

While there is a latency difference even in the lab between the two, it's not nearly so great as to cause serious pain. I have used my laptop tethered to my 3G phone often, and have no problems using it... but then there are other times when it's almost painful. EDGE on the other hand is slightly less painful than 3G at its worst, but EDGE is at least CONSISTENTLY painful across the entire usage spectrum. If you like your pain to be predictable, use EDGE :D

Seriously, though; my experience with 3G is really good. Perhaps that's partly due to the TYPE of traffic I use; mostly IMAP for my mail, and some light web browsing. The only time I really browse the web on my 3G device is really when I'm bored, waiting for something and reading news articles. Even then, when I do so I'm typically on "device-optimized" sites rather than the "full blown page". Mostly my news comes from Google News and the BBC... both of which work fantastically on my 3G device.

Only when I'm using my laptop tethered to my device do I really notice that 3G is slow relative to a good broadband connection... but even then I rarely use it much because I've normally got access to a WiFi connection in most of the places I use my laptop. Even if I don't, in the US we do have the advantage that there's usually a coffee shop relatively close that has free WiFi for their customers. As a result, I don't really use the tethering much, but it's nice when I need it. And when I need it, 3G beats EDGE every time.

fanboy alert.

[b96miata]

Here's all I need to know:

I've used my GF's brand new WM-based EDGE-underpinned smartphone side by side with my 2-year old Samsung (regular)phone running J2ME on EVDO. My antiquated handset blew it out of the water.

This guy's on Frickin crack. He's a fanboy. That's all. The comparison might, *might* be valid in very specific areas, but wholly because of how fucked AT&T's 3g network is in the few places it exists. Get on a proper 3g net (like the ones in other countries, or the CDMA based ones in the US) and there's no comparison.

Re:fanboy alert.

[Whiteox]

Good 3G service here (Hutchinson-Qualcomm and Telstra NextG) work very very well in Capital cities. NextG works well in rural areas as the frequency is quite low (850mhz I believe) so it travels further.
No matter. Good coverage makes 3G work very well.

Personally....

[r_jensen11]

I refuse to buy a phone that's locked in to one carrier. If it doesn't support SIM cards and let me use it wherever I go, I'm not getting it. If a carrier doesn't support GSM, then I don't support them.

The title of this is wrong

[cdhowe]

I'm the author of the original article on Blackfriars Marketing, and the title of this thread is misrepresenting that article. The original title is "Why EDGE versus 3G matters less than you think.", not why EDGE is getter than 3G. I've posted a followup to the article today here. It's not nearly as inflammatory as implied here.

Re:The title of this is wrong

[Iloinen Lohikrme]

I think that one reason for much of the negative feedback here in Slashdot is that you make general judgment based on the American mobile infrastructure and market which both are less developed than in Europe and Japan.

Also in your follow up article you touch some subjects and say that what consumers want is low-latency mobile applications including plain phone calls, messaging, web etc.. Unfortunately you forget that 3G is absolutely must if you are mobile user in urban areas. The original reason for introduction of 3G wasn't to have a higher data access, but to enable plain phone calls in heavily congested European cities. If my memory deserves me right, in the late 90s many mobile operators were very afraid that with out new networks, mobile networks would fulfill and the rate of dropped calls would skyrocket. So having a 3G network and 3G phone are definitive pluses. I also don't agree that there wouldn't be need for data heavy applications like mobile radio and TV, and mobile shops for music and video.

It should be also noted that it's not 3G versus EDGE, as in any good 3G network and phone you can use all the networks GSM, GPRS, EDGE and UMTS. In urban areas you get UMTS and when you enter countryside you get EDGE or GPRS. In that sense it's quite futile to even discuss about 3G versus EDGE as they are complementary to each other.

This is bunk

[photon317]

I switch from being a longtime EDGE user to basic 3G (just plain UMTS, no HSDPA) on a Nokia N75 device. I'm able to switch back and forth at will for testing. The bottom line is that UMTS gives me double the bandwidth and less than half the latency. Better all over. I haven't tried HSDPA, perhaps that in particular is the problem. It has much higher transfer rates than plain UMTS and EDGE, and I have no idea what the CPU usage or latency is like. Mostly, I use EDGE/UMTS for four things:

1) Very rarely, browsing the web via Opera Mini on my phone to look up a movie showtime while I'm out or something like that
2) Google Maps for Mobile
3) Gmail Mobile
4) Tethered data access over bluetooth to my MacBookPro, where I use it for browsing, corporate email, and lots of interactive SSH sessions.

My subjective experience is that all four are better on UMTS than EDGE.

Article is fanboish-crap

[WhiteWolf666]

Latency?

Has the author _EVER_ used an EDGE network?

Latencies over 1000 ms are not uncommon. On EVDO RevA, Latencies under 100 ms are expected.
3G Wins

Battery Drain?

For any given quantity of data, EDGE will take an order of magnitude longer to transfer it. As such, the radio is transmitting/receiving 10x as long. Also, he's assuming some sort of theoretical, 100% efficient (per information theory) protocol is being used on EDGE, which is far from the case. Newer protocols are actually more efficient, not less; there's a reason that ALL new Sprint and Verizon phones are 3G; indeed, there's a reason most new phones worldwide are 3G. Their battery consumption and frequency usage are a good deal better.
3G Wins

Error Prone?

Ever try to use an EDGE data device while travelling over 60 mph? Constant retransmits. Ever try to use an EVDO RevA device over 60 mph? Works just as if it were stationary.
3G Wins

Processor Speed?

Bwaahaaahaa. Beside using a "real" internet device, like connecting your laptop or PDA via bluetooth to a phone, many new Windows Mobile Phones (which I absolutely *hate*) have processors that would put a 4 year old PC to shame. You can get dual-core WM phones now; and Nokia's latest and greatest is nothing to sneer at, either. The main issue here is browser optimization; install Opera on your WM phone, and you'll have snappy browsing. The built-in browsers on the firmwire tend to suck, which is a software problem, not a hardware, or frequency problem.
3G Wins

Not matter what, however, I'm certain that the author has not used a modern 3G network. You cannot compare the latencies of EDGE and 3G; 3G is fantastically better; at least an order of magnitude, and in many cases two orders of magnitude. You cannot compare 3G versus EDGE when you are travelling; and you cannot compare the error rates.

I'm not saying the iPhone isn't a wonderful device, but an EDGE network? It *sucks*. I've had to work over an EDGE network doing interactive tasks, and it made my life miserable. Upgrading to EVDO RevA was a blessing; I can run Skype over it! And combined with a decently powered WM5/6 device (Motorola Q), I've had a vastly better experience, even admitting that the phone's UI sucks terribly compared to an iPhone.

I spent the weekend camping with a friend of mine who has an iPhone. And it knocked the socks off my phone in terms of UI. But I was streaming video, downloading large files, and had snappy access to websites; not to mention VOIP.

This blog is a troll.

Re:Article is fanboish-crap

[NNKK]

Amen. I don't know what the hell he's talking about, but every EDGE link I've ever used, best ping was over 700ms. My EVDO card from sprint never breaks 200ms, usually hovers between 120 and 150, emphasis towards the low end. And SSH is actually practical over it!

Re:Article is fanboish-crap

[Anonymous Coward]

Dont confuse EV-DO to 3G... 3G is based on GSM.... EVDO RevA is CDMA... Cingular(ATT), T-Mobile, are GSM... Verison, Alltel, and i think Sprint are CDMA... Not 100% sure on Sprint

I understand where they are coming from

[gravis777]

Ever tried using high speed internet over a line that has a lot of noise in it? Up until I replaced the wiring in my house, there were many times I saw my 6 Mbps DSL line drop to speeds painfuly slower than dialup. Remember back with analog cell phones, how awful the quality was? Okay, different technology, but I am trying to go somewhere. You got static, you got echos, it was awful. Digital cell phones have dramatically improved quality, but have you ever talked to someone while they were driving or something? There seems to be quite a huge drop of data packets. This is fine for voice (well, for most of us, lets say we put up with it), but could be fatal for data. Could you imagine the number of data errors that must happen on a cell network?

I have experienced web browsing on an Edge Datacard, and I can tell you it is painffully slow, but, as the article describes, this seems to be from latency, not from bandwidth. I will sit here sometimes waiting anywhere from 5 to 30 seconds for it to send its page requests, then it has to acknowledge that the server has recieved the requests, then the page loads up. The actual load time are not that bad, its just the latency.

Now I have been able to compare Edge VS 3G on Blackberry BES activations. I must say, I HATE 3G Blackberry activations. Yes, the activation itself is usually about 5-10 times faster than the Edge, however, the Edge activations usually never fail, the 3G usually fails several times before it finally goes through. In reality, the Blackberries I have activated on 3G networks tend to take me a couple of hours to do, whereas the Edge ones I can normally knock out in about 10-20 minutes.

UMTS supports concurrect voice and data

[stefanb]

Haven't followed the latest EDGE specifications closely, but AFAIR GPRS and EDGE are limited to either sending and receiving data, or having an active voice connection. So no looking at an incoming mail or at some document while you're talking. UMTS allows both data transmission and voice circuits to be active concurrently.

I don't have practical experience with EDGE, but latency on GPRS just blows, compared to UMTS.

hmm...

[hitmark]

i could have sworn that i recently read about a enhanced EDGE system from ericsson.

Not really what he said, but...

[cirby]

Of course, you just misrepresented what the guy actually notes in his article, but one other thing about EDGE vs 3G: a lot of the time, I get solid EDGE connections with good throughput, where the 3G guys are waving their phones around trying to get ANY connection whatsoever.

Re:Not really what he said, but...

[jrockway]

That doesn't make any sense. In the US anyway, 3G HSDPA degrades to EDGE and then down to GPRS data.

Latency

[simpl3x]

I have had experience with this issue in terms of cell cards. My Verizon CDMA card was much "faster" and more responsive than the EDGE card due to the fact that one gets an actual connection through an IP number in CDMA. Somebody add to this, but that was my recollection in terms of online applications and such. Edge really was not great in my opinion, but I stuck with it nonetheless for a couple of years.

If UMTS/HSDPA is worse, I'm not looking forward to it. EDGE on my iPhone truly blows most of the time.

Re:Latency

[nxtw]

If UMTS/HSDPA is worse, I'm not looking forward to it.

It's not worse. It's much bette than EDGE.

My Verizon CDMA card was much "faster" and more responsive than the EDGE card due to the fact that one gets an actual connection through an IP number in CDMA.

That has little to do with it. CDMA is simply a superior technology (1xRTT has lower latency than EDGE, even though it has less total bandwidth; EVDO blows both away). HSDPA is a lot better than EDGE & directly comparable to EVDO.

384 kbps is 3G?

[Solandri]

I've used Sprint's EV-DO 3G service for about two years now. I average about 800kbps - 1.5mbps. Best I've gotten was slightly over 2mbps. It was actually faster at file transfers than the shared T1 at my previous workplace.

cost of speed?

[tgibbs]

Faster is always better. But the speed of my iPhone is adequate for my purposes, and I wouldn't be willing to trade a speed bump for less battery time or a bigger phone.

3G

[Anonymous Coward]

3G should out perform Edge because 3G is newer technology... Both are based on the GSM system which in my opinion sucks balls.. I have never experienced so many dropped calls while on the GSM system... Where i had a 3G phone and 3G wasnt offered just yet so it was still on edge so i cant tell you speeds between 3G and edge but whenever i did get on a 3G system is still lacks to EV-DO its far superior plus based on the newer CDMA2000 technology.. plus when i comes to using the phone i have not dropped a single call... GSM is built with a narrow bandwidth so when you have several people using 1 tower i dont think it matters if you have edge or 3G its going to be slow..... EDGE speeds are equivelent to maybe an ISDN 128K line... 3G speeds are approaching maybe 200-300K EV-DO speeds are at DSL speeds 768K... Until they GSM gets their act together I wont go near their system again... Besides I think the future is WCDMA but thats for another topic...

Experience

[lymond01]

Here's my experience:

1) When surfing the web on Edge, webpages turn up very slowly, if at all, regardless of signal strength.
2) When surfing on 3G, webpages come up in seconds, if at all.

The article seems to imply that bandwidth isn't the bottleneck. In some cases it isn't, in most of the cases you're using the data portion of your phone plan, it is. If we were charged by time downloading instead of just a flat rate, no one would buy an iPhone.

HSDPA has lower latency than EDGE

[punka]

I dont know where this article got its facts, but it is dead wrong on the latency issue.

EDGE Latency = 500ms - 1000ms (yes, that is 1 second!)
HSDPA Latency ~ 50ms - 200ms

HSDPA is lightning fast not because of the increased bandwidth (which yes is nice), its because its lightning FAST in response!

And btw, I think that if an EDGE radio has to be on for 60s download a page, it will consume more power than a 3G radio on for 5s.

I know how to fix this

[Anonymous Coward]

Move to a country that doesn't suck donkey balls

Re: Strange reasonning

[nxtw]

I don't get the argument that high speed networks are more error prone. Following this rationale, this would mean that our broadband or wifi network would be much less reliable than a good old modem.

Yes; high speed wireless networks have extensive error correction. For example, 54mbit/sec is the maximum raw data rate for 802.11a/g, and doesn't consider error correction. This is why you don't actually get those speeds in real-world applications.

Also, I fail to see why EDGE technlogy would be less prone to radio interference and reverberation than UMTS

EDGE is more prone to interference & reverberation. The things that CDMA did better than GSM/TDMA ten years ago still apply; WCDMA-based UMTS is better at dealing with adverse RF conditions than GSM/EDGE..

I don't know if EDGE retains the same principle

It does. EDGE is just a way to pack more bits into the same timeslot.

Dear Blackfriars Communications,

[usv]

When next time writing about EDGE vs. 3G, try taking your head out of your collective iAss first. Thank you.

Point by point

[Anonymous Coward]

"People confuse network bandwidth with latency." Yes they do. However HSDPA and UMTS both have much better latency than EDGE and higher bandwidth.

"High bandwidth radio networks are more error-prone." They have error correction. Also, with fancy stuff like EVDO, UMTS, HSDPA, and even CDMA 1X, they can trade off data rate versus additional error control info dynamically. Also actual data rate can be stepped up and down; a 14.4mbps HSDPA network isn't going to try to toss 14.4mbps at you when you have a 0-1 bar of data service (if it's set up right.) With EVDO, I've gotten like 800kbps+ while near a site, then 320kbps while just barely in range of the same site; no packet errors whatsoever, just lower data speeds.

"Phone processors and software don't necessarily keep up with fast data transmission." Yes. But when they do, they are faster with the faster transmission.

"High bandwidth networks drain batteries" Yes they do. But Job's excuse is just that, an excuse. The difference isn't that high. For Verizon's EVDO phones at least, there's a difference in battery life when an area is marked as having EVDO in the roaming list, but doesn't actually have EVDO yet (because the phone scans regularly for this non-existent signal.) I'd say 10-20%. Then when Verizon is TESTING EVDO there's a huge difference, the phone might last 1 day instead of 5, because it tries to connect to the EVDO multiple times a minute but is kicked off. Then when EVDO is up, I'd say the difference is 5-10%. On most phones EVDO can be forced off, and on the existing UMTS/HSDPA phones as far as I know they give explicity "GSM only", "auto" and "3G only" choices, so they can have the 3G forced off too to save battery life.

I get Zero latency using my blackjack.

[jamie(really)]

Latency and bandwidth are issues. Which is why I don't rely on a browser to get my shit done. My email is pushed to my phone and when its ready, I look at it. Zero latency: its on the phone. When I update contacts, or my calendar, its synced with my server. When I make changes online, or on my laptop, they are synched to my phone, automatically. My exercise program? On my phone using Documents To Go.

OTOH, when I want to browse and download tunes or watch movies, I get about 1Mbit/s. Try that on EDGE. I know, because when my phone can't get 3G (HSDPA) it drops down to EDGE speeds.

This article is a joke. How the hell did it make it onto /.?

Alphabet

[trovak]

It looks like the alphabet threw up all over this discussion.

This article's problems have been solved long ago

[Sultan]

The problems discussed in this article are well known in the industry and there have been solutions to most of them for years now. Secondly as many others have noted 3G/HSDPA operate in a totally different mode and with HSDPA the latency has been cut to 10s of ms, instead of 100s of ms.

People like Bytemobile - http://www.bytemobile.com/, Flash networks - http://www.flashnetworks.com/asp/main.asp, 724 solutions - http://www.724.com/, etc. have been working on these problems for years and if the operator has their product the latency should be a minor issue even when using GPRS.

What this article doesn't talk about is how inefficient EDGE is when it comes to implemeting it for an operator. It's no surprise that most euro operators don't have EDGE systems as expanding GSM to cater for EDGE is an extremely expensive excercise.

You're the one who's wrong

[scgops]

Carl,

We know why EDGE vs. 3G doesn't matter much to you -- you have an iPhone.

You can rationalize it however you want by trying to talk about latency, but then you bump up against the fact that you haven't actually tested latency on an EDGE network and, when you do, you'll find out that it's actually pretty bad.

I've got measurement data that shows latency is typically around 130 to 160 ms on 3G networks, including AT&T, Sprint, and Verizon. 2G networks, by comparison, usually have 250+ ms latency, and are often in the 300 to 600 ms range.

Even if you're right, and latency matters more than bandwidth, EDGE doesn't win the competition.

Okay, so what about real-world application performance? You've pointed out that someone who can optimize their applications to make maximum use of network and phone resources will be able to deliver the best applications.

Well, on average, wap.google.com takes 11 seconds to load over 2G networks vs. 8 seconds for 3G. That might serve to strengthen your point, because performance isn't significantly worse on EDGE.

However, wap.mapquest.com takes 27 seconds to load over 2G vs. only 8 to 11 seconds for 3G. Perhaps you think that's Mapquest's fault. Or the phone carriers'. Either way, the difference between 2G and 3G networks clearly does matter to end users, particularly when being limited to 2G networks can make connecting to some sites take 2 to 3 times longer.

Since not all sites are lightweight enough to perform equally well on both platforms, the easily perceived performance differences between EDGE and 3G networks will continue to matter to end users.

-DaveU

EVDO and Skype

[meehawl]

Ever try to use an EVDO RevA device over 60 mph? Works just as if it were stationary.

This is true. EVDO is annoying because like EDGE it won't do net connects during voice chats. So you can't, for instance, chat on a voice line and use Google Maps (unless you're on Wifi). But there's a way around it. I did a cross-US drive a while back. Used the Windows Mobile phone (Sprint Mogul AKA HTC Hermes) to run Skype and chat a lot while also grabbing network data. Worked great. Even managed to get some crappy quality chats in the middle of the Utah and Nevada deserts. When EVDO is working (ie, >=1Mbps), you can even do video conferencing using something like Windows Portrait.

EVDO Means No Sponging

[meehawl]

Maybe with iphone 2 you won't have to sponge off random wifis for a consistent, nonlocal signal. Also, bit hard to use Google Maps on the phone if you're driving and relying on wifi.

Further information, running Opera Mobile 8.65 on my Sprint Mogul (HTC Hermes) for redhat.com:

EVDO
9 seconds

1xRTT
28 seconds

802.11g
2 seconds

I just ran a speedtest from DSLReports, and it says my EVDO is currently running at 600 Kbps with 350 ms latency. That's bad. However, I have tethered this phone to a notebook and used it to download bittorrents and for video conferencing, and gotten speed tests up to 1.3Mbps, so I suspect that the phone's 400 MHz CPU is the limiting factor here.

Sadly This is Simply Not True

[Novarum]

I have been now actually measuring various cell data networks with various clients .. including an iPhone. And the truth is that a real 3G network (HSPA or 1xEVDO Rev A) has about 300 msec latency and usually will deliver 2-300 kbps upstream TCP throughput and 700 kpbs downstream TCP throughput. On average. YMMV and it does vary based materially on time of day and network load. None of these networks does very well in an airport around 5-6PM. And of course .. you should preferably be in a major metro area to get 3G at all. EDGE - either on an iPhone or a separate data card ... gets north of 500 msec latency, often approaching 1000 msec, and averages around 100 kbps throughput. The good news sorta .. is that that throughput is roughly the same up and down. EDGE is really not very good. HSPA is really quite good. And is getting better. WiFi compensates a bit for EDGE's performance shortcomings ... but frankly it is the delicious quality of the web browser image on the iPhone that distracts from EDGE's inadequacy. Now ... an iPhone with HSPA and an 802.11n client? Rock and roll guys ... this would be truly high performance at the state of the art. And a world phone as well.