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

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.

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: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...

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

[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? ;)

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: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: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: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.