A journey on the evolution of the G

[kckid]

I thought this was an interesting article with an informative graphic about the evolution from 1G.

 

http://www.telecomstechnews.com/news/2013/aug/16/journey-evolution-g-infographic/

 

 

 

[lynyrd65]

4G is 100mbps eh? Not yet.

[The Dave]

Once LTE-A becomes the norm in the next year or so, I imagine we will start seeing that.

[koiulpoi]

I'm glad this infographic actually mentioned the fact that 4G vs 3G is a move from a switched network to an all-IP one, which is one of the biggest, real broad changes between generations (along with increases in spectral efficiency, etc.). Sadly, it falls in to the "numbers" trap, quoting the (often touted and completely, utterly arbitrary) "100 Mbps" number.

[Nickel]

Once LTE-A becomes the norm in the next year or so, I imagine we will start seeing that.

I'm not convinced, I think you have a better chance of seeing those on Verizon with 20 MHz of AWS.

[The Dave]

I figured with the spectrum that Sprint now has on band 41, it'd be possible with TD-LTE-A

[gnoj]

I'm glad this infographic actually mentioned the fact that 4G vs 3G is a move from a switched network to an all-IP one, which is one of the biggest, real broad changes between generations (along with increases in spectral efficiency, etc.). Sadly, it falls in to the "numbers" trap, quoting the (often touted and completely, utterly arbitrary) "100 Mbps" number.

 

Well, the speed benchmark is kind of arbitrary, but when they were writing the standard, a line had to be drawn somewhere. At the end of the day, speed (and ping) is what effects the experience of using data on a phone, and users don't know or care whether the data is circuit or packet-switched. If an "all-IP" network only ended up being capable of 4 Mbps instead of 3G Ev-DO rA's 2 Mbps, I wouldn't really consider that to be enough of a technological advancement to be worthy of being considered a new "generation."

 

I figured with the spectrum that Sprint now has on band 41, it'd be possible with TD-LTE-A

 

Yes, a single 20 MHz TDD carrier is I believe capable of 90 Mbps downstream, so two of those aggregated together should get the job done. I'm pretty sure Sprint has at least 40 MHz of Clearwire spectrum in every market, although unfortunately it's not usually contiguous. Once battery life limitations are overcome (carrier aggregation apparently draws a lot of power), I can see Sprint deploying additional TDD carriers pretty quickly, along with whatever firmware upgrades are needed to update from LTE r9 to r10.

[koiulpoi]

Well, the speed benchmark is kind of arbitrary, but when they were writing the standard, a line had to be drawn somewhere. At the end of the day, speed (and ping) is what effects the experience of using data on a phone, and users don't know or care whether the data is circuit or packet-switched. If an "all-IP" network only ended up being capable of 4 Mbps instead of 3G Ev-DO rA's 2 Mbps, I wouldn't really consider that to be enough of a technological advancement to be worthy of being considered a new "generation."

It's completely artbitrary, because it depends on the amount of spectrum you're using and the spectral efficiency of your technology. The first depends entirely on licensing schemes, and has nothing to do with technology. The second involves new engineering, and is indicative of a new generation. Are you seeing the problem? By assigning an arbitrary speed cap, an arbitrary assumption of (iirc) 67 MHz (for 1 Gbps fixed) worth of spectrum being the minimum to run a network was put in place. Having more spectrum to use doesn't make your technology new, it just means... you have more spectrum. Spectral efficiency? Yes, that matters, a lot. It also is something end users don't know a thing about and don't care about. Engineering standards should not be measured by how an end user perceives it. That's dumb and silly and we should stop. Providing a good user experience is a worthwhile goal, but that's not what my first post was about.

[kg4icg]

Something to think about. I am on a 20mb fiber connection at 7ms ping and can do all I want to with hd video and all, so why is it everyone with a phone wants to have 100mb connection knowing that there phone doesn't have the capabilities there normal comnputer have? It makes no sense?

[gnoj]

It's completely artbitrary, because it depends on the amount of spectrum you're using and the spectral efficiency of your technology. The first depends entirely on licensing schemes, and has nothing to do with technology. The second involves new engineering, and is indicative of a new generation. Are you seeing the problem? By assigning an arbitrary speed cap, an arbitrary assumption of (iirc) 67 MHz (for 1 Gbps fixed) worth of spectrum being the minimum to run a network was put in place. Having more spectrum to use doesn't make your technology new, it just means... you have more spectrum. Spectral efficiency? Yes, that matters, a lot. It also is something end users don't know a thing about and don't care about. Engineering standards should not be measured by how an end user perceives it. That's dumb and silly and we should stop. Providing a good user experience is a worthwhile goal, but that's not what my first post was about.

 

100Mbps/1Gbps is a speed floor (for theoretical peak speeds) not a cap. If it takes 67 MHz of spectrum to deploy a true 4G network, then that's what it takes. I agree with you that spectral efficiency matters more than raw throughput (as long as that throughput noticeably exceeds the top speed of the previous generation), but that's not how the ITU chose to define it. They own the trademark to the "4G" moniker, so they can define it however they like. I found it frustrating to see private industry abuse the term and defining it however they wanted to, which led to consumer confusion as each "faux-G" network performed very differently.

 

Hopefully the ITU has learned from that debacle and will be a bit more realistic about constraints on spectrum availability when writing the next set of IMT standards.

[koiulpoi]

100Mbps/1Gbps is a speed floor (for theoretical peak speeds) not a cap. If it takes 67 MHz of spectrum to deploy a true 4G network, then that's what it takes. I agree with you that spectral efficiency matters more than raw throughput (as long as that throughput noticeably exceeds the top speed of the previous generation), but that's not how the ITU chose to define it. They own the trademark to the "4G" moniker, so they can define it however they like. I found it frustrating to see private industry abuse the term and defining it however they wanted to, which led to consumer confusion as each "faux-G" network performed very differently.

 

Hopefully the ITU has learned from that debacle and will be a bit more realistic about constraints on spectrum availability when writing the next set of IMT standards.

Whoops, misspoke. Meant speed floor.

 

I'm glad you agree that the ITU requirements are a little ridiculous, and as "they can define it however the like", that you agree that it's completely arbitrary.

 

My point is, when it comes down to it - "faux G" or no... It's one thing entirely to say "your technology must have a minimum peak downlink spectral efficiency of 15 bits/s/Hz" (to quote Wikipedia), and another thing altogether to say, as I've heard so many tech afficionados and media sites say, "ITS NOT REAL FOUR GEE UNTIL I GET A HUNDRED DOWN!!!1111". The first is a real technological advancement, the second is a dick waving contest, and has nothing to do with generational advancement.

[Mobilesolutions]

So my 68Mhz minimum in every market is good enough for now!

[Nickel]

I figured with the spectrum that Sprint now has on band 41, it'd be possible with TD-LTE-A

Well, they can but I think they'll wait until more devices are LTE-A compatible, but you never know. 

[The Dave]

All your spectrum are belong to us, mwahahaha!

[lynyrd65]

So my 68Mhz minimum in every market is good enough for now!

Let me know when you get your Tulsa FIT on.

 

Sent from my One using Tapatalk 2