Slashdot Mirror
The Many Faces of 3G
An anonymous reader writes "Did you ever notice how each new generation of cell-phone tech gets branded '3G,' and the previous thing is retroactively downgraded to some lesser number of Gs? An MIT engineer explains why in this brilliant essay about '3G' over the last 10 years, showing how the cell carriers have kept offering it and swiping it away to sell more stuff. He cites numerous Cingular/AT&T and Sprint press releases showing how the companies have made '3G' into a brand name ideally suited for amnesiac consumers. Meanwhile, no cell carrier is foolish enough to sell you bottom-line throughput like an ISP in 1996 — you could actually hold them to that (PDF)."
No, can't say I noticed that at all.
Next question?
I want the 3 G's ... and the WiFi's ... and the G-B's ... obligatory youtube video
Somebody drop some straight science on this geezer. Y'all yak "3G" like y'all know what it means. What the heck does it mean? I only vaguely get that it's provides bit faster data service. Rummaging through Wikipedia only muddle things up further by delving into various optional protocols.
Fuck systemd. Fuck Redhat. Fuck Soylent, too. Wait, scratch the last one.
...no cell carrier is foolish enough to sell you bottom-line throughput like an ISP in 1996
Metro PCS? It's cheap, but I dropped every single call I ever made on their antique phones before I switched.
Drill baby drill - on Mars
Isn't 3G on a recent network like an automatic bandwith limitation? There are modems that operate at another "G" that also compete on the Service Network for bandwidth, and they're generally cheaper than buying a Cell Phone to make phone calls. Sprint has a couple made through Sierra, some USB dongles for the Cell "G" networks and a couple others that are WIFI-only like a HUB without LAN.
From the article:
>What we really ought to care about is the same as with any Internet service provider -- the throughput
>and latency and reliability you get to the endpoints you want to reach. That's what matters, not the
>sophistication of one piece of the puzzle.
I have often wondered about all the marketing jargon floating about cell phones, and about people who go ga-ga about how their cell phone browses the internet.
Every phone I've tried browsing the web on makes me just about cry with frustration - I feel like I'm back in college with a 2400 baud modem again.
When you shop for an ISP you shop based on best-effort advertised upload and download rates.
Cell phones should be the same way.
A work that expires before its copyright never enters the public domain and thus enjoys eternal copyright protection.
how many people know what UMTS, HSPA, HSPA+,HSUPA, CDMA, WCDMA mean? even a lot of the tech enthusiasts who think ATT is on GSM don't know that AT&T is using a form of CDMA. i saw it in the Nexus One settings on the internet where the network says WCDMA.
Did you ever notice how each new generation of cell-phone tech gets branded '3G,'
No. Especially since this iPhone 4G thing came out. It was in the news, you might have heard about it.
I've got some news for you... ...and make sure you've got a black turtleneck to wipe your manboy fanboi tears with...
You might want to sit down...
Uhhh, wasn't that the iPhone 4, not the iPhone 4G?
Do not argue with an idiot. He will drag you down to his level and beat you with experience.
They didn't "make 3G into a brand"; it has always BEEN a marketing label. There is no such thing as a "3G" wireless signal, rather there are various (existing and emerging) modulation techniques which collectively exist under the 3G label. http://en.wikipedia.org/wiki/3G#Overview
Moreover, the signal is the phy layer. The fact that you have a 3G signal doesn't guarantee any minimum performance, any more than having a gigabit NIC guarantees a fast internet connection. It only defines the upper boundary of performance.
https://www.eff.org/https-everywhere
The Sprint HTC EVO is 4G, but the iPhone 4 is definitely not 4G. You might want to pay more attention, especially if you plan on being a sarcastic douche about it later.
Sprint, at least is calling its LTE network "4G", as it rolls it out.
As I understand it:
As I see it, the xG shorthand is a way to track the evolution of the network, link level, and physical layers. Every time one of those changes, you get a new "generation" of cell phones.
We all know what to do, but we don't know how to get re-elected once we have done it
No. Especially since this iPhone 4G thing came out. It was in the news, you might have heard about it.
The iPhone 4 supports 3.5G (HSPA+)
http://www.google.com/search?q=3.5g+iphone+4
4G is the new 3G
It's all just marketing talk and the details are buried in the fine print
[Fuck Beta]
o0t!
I'm sorry, did you mean the HTC EVO 4G?
Or the Iphone 4, which in spite of its shininess and hype, is still 3G?
Welcome to the Panopticon. Used to be a prison, now it's your home.
I think that the overuse of 3G (and subsequent use of 3G as an advertised speed) is a result of locked phones being tied to carriers. When Joe Average Consumer goes out to buy landline internet, there really isn't a whole lot to choose from that differentiates comcast, att, and whoever else. The main thing he decides on is speed; the hardware that comes with is usually irrelevant. What we have in the cell phone market is 3G being used as a sort of loose guarantee that internet will be somewhat fast. The whole using a protocol as a speed definition is stupid, but the reason Joe doesn't notice is that he is too busy choosing which phone to use, which determines the carrier. It seems all carriers have realized that it is significantly easier to advertise "3G enabled" and not put a speed on it, and let the phone pull in sales, rather than the network. If we lived in a world (or nearly any foreign country) where unlocked phones are the norm, you'd pick your phone, then comparison shop for either the fastest or cheapest (or balance of the 2) network.
tl;dr version: Overuse of 3G is caused by locked phones
"Going to war without the French is like going deer hunting without your accordion." ~General Norman Schwarzkopf
*psst: So's the HTC EVO. It's just newer 3G.*
Do we really need to point out that 3G doesn't actually mean ANYTHING? Hell, I'm surprised we're not at 10G or higher now, nothing stops any carrier from one-upping the competition by simply saying "Sprint may have 4G, but we have 5G!!". that's what happens when you make up terms that don't mean anything.
"Have you ever noticed..." No, i haven't noticed that, nor do i care. What i do know is, is that the shit's still overpriced.
haha you dumb apple fanboi cunt bitch! you got servedddddddd
These half-generations are retarded. Who the hell is the idiot behind those terms?
The Sprint HTC EVO is 3G [Wimax not being genuine 4g], but the iPhone 4 is definitely not 4G. You might want to pay more attention, especially if you plan on being a sarcastic douche about it later.
Fixed that for ya!
The answer to slow cellular data rates is Opera mini. Browsing raw interweb on cellular is just horrid.
The Sprint HTC EVO is 4G...
Except it isn't. You should read the article, it's good.
Eagles may soar, but weasels don't get sucked into jet engines.
At least here (Israel) iPhone IS marketed as 4G...
I'm not one to procrastinate. I'll be a sarcastic douche about it now!
Sometimes, life itself is sarcasm...
>and what we'll REALLY get is 200-300k speeds for data on a good day, and one active voice connection that works ... sometimes.
But this is exactly how it works with home internet.
You can't hold an ISP to download/upload speeds because in the end we are all on a shared pipe. But we should at least have a ballpark to work with. It's generally understood when you buy home ISP service that the advertised rates are _maximums_.
I would like to see phones advertised similarly.
A work that expires before its copyright never enters the public domain and thus enjoys eternal copyright protection.
Hahaa! You're all absolutely right! I stand corrected and I'll get my coat!
Drill baby drill - on Mars
Meh, the business bullshitters in charge also use the G, mostly to stand for "Generation". So the parent is mostly correct in that context.
Verizon / Sprint started as CDMA (code division multiple access, GPS satellite signals also use this) networks, vs. everyone else who started as GSM use TDMA (time division multiple access).
In CDMA, all units basically transmit on the same wide frequency, but have a unique code to distinguish their signal from others. In TDMA, all the units get timeslices (~120 per second) and narrower frequency bands, and transmit bursts of packets during their allotted timeslice. That's why you can hear GSM phones cause audible humming (apparently at 217Hz) when you place them near an amplified speaker. Also made GSM phones theoretically more energy efficient, since they don't have to broadcast the carrier wave the entire time they transmit.
http://www.smartdevicecentral.com/article/that+crazy+gsm+buzz/199379_1.aspx
Originally CDMA might have been a bit better for sparse country networks, and GSM better suited for densely populated city networks, but with the new standards emerging around the "3G" timeframe, they both pretty much incorporated each other's technologies into the newer WCDMA standards.
Here's my dicey understanding of the GSM generations:
(See the bottom of http://en.wikipedia.org/wiki/Enhanced_Data_Rates_for_GSM_Evolution for the actual standards lumped under each G)
POTS (Plain old telephone service) : 64kbps line that carries uncompressed 8kHz 8-bit mono audio (that's why phone calls sound like crap when they're on TV / Radio talk shows). Also sort of explains why the fastest dialup modems were around 56k (after data protocol and error correction overhead).
1G AMPS: the old analog(ue) brick mobile phones the rock stars used in the 80s
GSM 2G voice: 8kHz 8-bit mono audio compressed using some codec that allows them to fit it in a ~7kbps - ~13kbps stream at roughly the same audio quality, except with compression artifacts. Different phones would support different codecs, but would fall back to some lowest common denominator. That's why some mobile calls sounded like crap while others sounded somewhat better. The data mode was called 3GPP (ha! 3G before the new meaning of 3G!) and could give you roughly the same data rates.
GSM 2.5G data: Then GPRS came along and gave you 56-114kbps, I think mostly by allowing you to use more timeslots if no one else was using them.
GSM 2.5G data the second: EDGE came along and gave you ~240kbps, mainly by squeezing more data bits into the carrier wave.
GSM 3G makes it more CDMA-like, where it can also use more of the frequency spectrum to pack bits. So now you can hog up to 1Mbps from the spectrum by using more timeslots and more of the available spectrum bandwidth.
GSM 4G: I'm not really sure where they're going with LTE Advanced, but more of the same, probably. Which means we'll probably have bunch more 3.5G brandings running around for a bit.
I have the EVO 4g, and the 4g is somewhere between worthless, and a handicap. If I leave it on, it sucks down the battery VERY quickly, and turning it on and off gets old quick. I can do everything I need it for with 3g. I have not jail broken it as I don't need tethering. 3g EVDO works just fine. I do know about the 5gig cap on 3g, where as 4g does not have a cap. I got it for the 1ghz, Android, and 4.3 screen to replace my palm pre. The 4g-wimax chip also does the wifi, and has a max power usage of 350ma. Wimax on a phone just can not go through buildings well enough to be a turn it on and forget it without killing a battery. Now at home I have a larger more powerful Wimax modem for home use, and I have 5 hardware computers, along with 10 or so VM's all sharing a single 3Mpbs/1Mbps Wimax connection, and other then ping times in the 100-350ms range, it works just fine.
>>>POTS (Plain old telephone service) : 64kbps line that carries uncompressed 8kHz 8-bit mono audio (that's why phone calls sound like crap when they're on TV / Radio talk shows). Also sort of explains why the fastest dialup modems were around 56k (after data protocol and error correction overhead).
>>>
POTS is actually only 7 bits, because the 8th bit is used for control signals. Hence 56k. Also the sample rate is 8000 times but the actual frequency width is only 4 kilohertz.
"I disapprove of what you say, but I will defend to the death your right to say it." - historian Evelyn Beatrice Hall
Except it isn't. You should read the article, it's good.
You mean. Except, it isn't yet. The Wimax Forum has applied for the 4G label from the International Telecommunications Union - Radiocommunication Sector (ITU-R). That application/proposal hasn't been accepted yet, but there is really no reason it shouldn't. 4G may be a so-called standard, but the International Telecommunication Union and its working group does treat it more like it's a brand than a real standard.
Oh cool... I always thought there was a good reason for POTS to sound worse than 8-bit / 8kHzsamples from my old ISA Sound Blaster card.
I thought there might already be some 8b/10b encoding going on too like in ethernet that brought you down from 64kbps, but I guess not... maybe that's why the best you could practically expect from a 56kbps modem was 40+kbps with any error correction overhead.
I did enjoy learning about the Viterbi decoder and other forms of forward error correction used in wireless mobile networks. Seems like there are still plenty of other applications you could apply that kind of thing towards...
Yup.
But the iPhone 2 was "iPhone 3G", and iPhone 3 was "iPhone 3GS".
I imagine the only reason Apple didn't call it "iPhone 4G" is because almost nobody has AT&T "4G" service available. Few enough have 3G service available with AT&T as it is.
Security is mostly a superstition... Avoiding danger is no safer in the long run than outright exposure. - Helen Keller
Don't let the door hit you in the ass on the way out!
Security is mostly a superstition... Avoiding danger is no safer in the long run than outright exposure. - Helen Keller
Obviously the ISP can't guarantee 500kbps to arbitrary websites.
However, they should be able to guarantee 500kbps to arbitrary high-bandwidth sites. (kernel.org, or microsoft.com, or various university sites)
Of course they don't *want* to do this because that would require them to advertise real speeds, which would force them to actually spend money to upgrade their infrastructure.
the PDF makes me think.
there is allot of talk about the Debian Freedom box where the apps and data live on peoples little sheeva plug at home.
SO, why not use the same idea for mireless data ?
I know for example that the FON system in Spain is quite popular. But they screwed it up by charging for it .
But considering what a huge waste of money is spent on paying these mobile operates, which are simply paying off the HUGE license fees that the government charged them.
When you just think about it logically its CRAZY.
1. I don't kind if someone uses a little bit of my bandwidth. Skype DOES that anyway !!!!
2. Peer to peer is secure at home because most people have it NATTing through their wifi router anyway.
its funny how we all work against each other, and the fat corporates reap all the the benefit...
I get 53 k out of my phoneline modem - that's the maximum limit allowed in the US, due to FCC speed limits. Otherwise it would be a solid 56k as advertised and per the V.90 spec.
The up speed is 48k.
"I disapprove of what you say, but I will defend to the death your right to say it." - historian Evelyn Beatrice Hall
POTS is actually only 7 bits, because the 8th bit is used for control signals. Hence 56k. Also the sample rate is 8000 times but the actual frequency width is only 4 kilohertz.
That applies to the robbed bit T1 interfaces, which can be debatably called POTS vs pure analog that was the original standard. More modern PRI-based T1s are a full 64 kbps since you lose a channel for signaling (23 vs 24 in old robbed bit land.) We use modern PRIs for that get broken down into POTS lines where I work, so no robbed bits. Of course, modems can't do any better than 53 kbps or something lame due to the telcos' petitioning the FCC so they don't have to explain why their network is stuck in the 70s.
I don't know, but it works for me.
First off, there's no technical article that's going to be worse than a person who has no idea what he's talking about, and who's basing his observations on technical details as given by a salesperson. From what I RTFA, it's the basis of this discussion, but it can't be. Using the typical car anaology, you've built a race car with no tires.
First off, 3G is a generic term. If I say 3G for wireless telecomm, I'm referring to CDMA2000, which is a 3rd generation of wireless data protocol. 2.5 G was never really accepted as 3G because it didn't implement all the standards, such as real time allocation, and it was circuit switched packet data (laymans terms: wireless modem). Getting back to "3G", a third generation of iPhone can be called 3G, but still work on the 4th generation of wireless standards, right next to a 4th generation iPhone (4G) running on an older 3G data network (Sprint, whoever). In an attempt to keep this discussion simple, we'll just stick to the wireless, 3rd generation data format when saying, "3G."
Before 3G data was sent over the air on dedicated channels. If you wanted to have more data, you set asside more time, or codes (TDMA or CDMA). However if network modeling was bad, you either banged on the headroom for data (surfing Google took longer), or voice (calls didn't go out or come in). Carriers in the US from my observation are always giving priority to voice. So the common configuration was to give them more "pipe" and higher priority.
Enter 3G CDMA, aka CDMA 2000. Initially there was only one way to implement 3G, later developers came up with newer formats that were backwards compatible in most cases, such as EV, EV-D, and EV-DO. These all have meanings, feel free to Google. In a nutshell though, they're all different implementations of 3G. 3G, or CDMA2000, allows the cell site to allocate pipes by usage and type. So, if you're data surfing at 1am, when no one is around making voice calls, you get the full pipe and your data screams. Use the same phone, on the same cell site at 12 noon, and you get the minimum pipe, and if everyone's on voice calls, you may not get out at all until a slot opens up. This is not to be confused with "breathing" (where cell sites expand and contract RF coverage according to usage). That's at the RF, or Layer 1 if you will.
When you start mingling WiMax and other technologies, you're now blurring the usage of the term. WiFi is not typical CDMA (I'm only hedging with "typical" because I don't know what modulation method wifi uses). Back to car analogies, it's like buying a 2009 car, putting a 2010 engine in it, and calling it a 2010. Yes part of it is a new generation, but it's still a 2009. Adding Wifi to a CDMA phone didn't take it from 3G to 4G, so from a logical techology standpoint, going WiMax isn't either. It's a different format, frequency, and usage.
Eventually, all these technologies will blur and the author will be correct in being confused. The telecom manufacturers (lucent, nortel, etc) have been moving the "ip up the train." In the beginning, they went out a specific trunk to a rack mounted shelf of modems (2.5 G, circuit switched packet data) which either went into another backend, or out a plain old telephone line (POTS). With the original implementation,, data shared RF with voice, came in the tower, went through the switch, which then split out data out a PRI interface (T1) to a server which converted over to TCP/IP and then used Home/Foreign Agents to manage real-time changing points of connection within a network. In laymans terms, you could jump in your car in San Diego, fire up your laptop, and drive from SD to New York without changing the IP address your laptop was assigned. When I left telecom (early 2000s), they were rolling out IP from the Site Controllers back. Meaning, the Mobile Switch back at the main office didn't break it out. The eventual plan back then was IP from the cell site. Everything coming out was TCP/IP, regardless of data or voice. 3G still all applies, because
>>>telcos' petitioning the FCC so they don't have to explain why their network is stuck in the 70s.
Nope. They discovered that modems operating at peak speed (7 bits and 56k) caused crosstalk on neighboring lines. So the FCC limited the *power output* of the modem to prevent that.
"I disapprove of what you say, but I will defend to the death your right to say it." - historian Evelyn Beatrice Hall