T-Mobile LTE & Network Discussion

[Conan Kudo]

Let's turn this discussion somewhat around. Let's say you're on a congested sector. Would you rather your speedtest show you what you could be getting if the sector was not congested or what you're actually getting? Since they have DPI they could momentarily elevate your priority to trump everybody else's so that your speed test results do not screw up T-Mobile's average. Would you be OK with that?

If the cell is congested, the speed test will reflect that. T-Mobile's treatment of speed tests was pretty much done at the request of its customers, in which they didn't want speed tests to count towards the data cap and they wanted to use speed tests to see what the network is capable of, barring the current status of their account, which is why I find it hilarious that people are complaining about it now...

 

The only condition in where speed test will not match what you experience is when you've used up your full speed data bucket. Every other condition will match (congestion, etc).

 

Contrary to popular belief, this is not really handled by DPI. T-Mobile isn't examining the data portion of packets to figure out whether it's a speed test packet or something else. The packet header (which is read for routing purposes) has information on the source and destinations. T-Mobile uses a whitelist of destinations that it acquired from companies that offer speed tests to exempt it. It's the same as how Music Freedom works.

 

DPI is computationally expensive (thus slow) and prone to errors (as not every speed test application tests it the same way). Destination whitelisting is computationally simple (thus very fast) and hard to screw up if you have the right lists. The lists may even be sourced by the companies T-Mobile is working with to whitelist in a dynamic form (such as a server list script that T-Mobile pings every 24 hours and dumps into its whitelist).

[derrph]

Ok so with this whole whitelisting of the speed test app and they have to show customers what their actual speeds are. What does it mean for T-Mobile as far as their rep being the fastest? And what does this mean for the customers?

 

 

Sent from my iPhone 6+

[Conan Kudo]

Ok so with this whole whitelisting of the speed test app and they have to show customers what their actual speeds are. What does it mean for T-Mobile as far as their rep being the fastest? And what does this mean for the customers?

 

 

Sent from my iPhone 6+

Insofar as the reputation is concerned, it will always show what you can get as if you had unlimited full speed data. No more, no less. For customers, unless they use the upcoming speed test application being provided by T-Mobile for the purpose of evaluating throttled speed, they will see the true performance of the cell site(s) they are connected to.

 

Since T-Mobile now relies on this data to check network performance in real time, speed tests that show weak performance consistently will trigger action on T-Mobile's end. So, if something indicates that T-Mobile isn't the fastest (or fast at all) consistently, then T-Mobile will evaluate the location and work to improve the situation.

[Keitarou]

They should have their own speed test if they want to exempt it from data allotments. Kind of similar to how Sprint Zone will test your SMS/MMS/voice quality. They could have their own TMO app to also show speed test results. Other speed tests should be showing what you actually get, not what you could potentially be getting if you paid more.

I get 85/85 on my home ISP, and that's what shows when I run speed tests. This would be no different than them giving me their 500/500 package, but only for speed tests. I could pay more for the higher throughput but I don't. Consumers could pay for more wireless high speed unthrottled data, but they don't. How is that any different?

That is in essence what they are doing by providing tools to allow people to check their real speed. There is no right or wrong way in doing what you describe. It's just a different way for them to do it.

 

Again the example with home ISP. That example doesn't make sense compared to what T-Mobile is doing. ISP does not sell unlimited data with limited high speed data after certain GB usage. They are not paying for high speed data but high speed data up to certain gb's. Simply put, consumers are paying for data.

[travismheim]

That is in essence what they are doing by providing tools to allow people to check their real speed. There is no right or wrong way in doing what you describe. It's just a different way for them to do it.

 

Again the example with home ISP. That example doesn't make sense compared to what T-Mobile is doing. ISP does not sell unlimited data with limited high speed data after certain GB usage. They are not paying for high speed data but high speed data up to certain gb's. Simply put, consumers are paying for data.

Not all home ISPs offer unlimited data, neither do all wireless providers. If T-Mobile is including throttled data over the data bucket cap (which is something I support BTW), speed tests should reflect what the customer is actually receiving. What's the purpose of a speed test? I know when I run one, I want to know what I'm getting. Not what someone else with a larger wallet could be getting.

[richy]

Not all home ISPs offer unlimited data, neither do all wireless providers. If T-Mobile is including throttled data over the data bucket cap (which is something I support BTW), speed tests should reflect what the customer is actually receiving. What's the purpose of a speed test? I know when I run one, I want to know what I'm getting. Not what someone else with a larger wallet could be getting.

That is the speed you would have been getting before you hit the cap so your comment makes no sense? A person with a larger wallet gets the same speed, just for longer. 

A speed test serves more than one purpose.  It's entirely correct to believe it should show the speed a person is getting at a given point in time even if they are capped,  I get that.  However,  the tests are also used to measure network performance and in that respect the uncapped rate is a more realistic measure. Ideally both should be given or the capped results should be discarded when considering relative network strengths. I understand why tmo did what they did,  I agree it isn't right for people who are capped to be shown a faster rate but I also agree with why tmo allowed the tests to run unthrottled. You don't obviously, I'm sure we will survive

[richy]

I get your point, to some extent.  But shouldn't a network who does not throttle and provides a more truly unlimited experience also get a chance to show its superiority in the rankings?  If someone like Tmo gets to pretend their users devices are not throttled, it gets to unfairly compete in speed rankings against a provider who does not throttle.

 

If Tmo wants to throttle, there is a drawback to that.  And this is it.  Can't have it both ways.  And does Tmo need all the help it can get on network speeds?  Hardly.  If that's the case, Sprint needs all the help it can get.

To your first point. I guess it depends on what you want to get from the results. Personally I want to evaluate potential new providers so I want to see an apples to apples comparison. Is rather have it represent capacity and congestion rather than be influenced by capped plans which is a situation I don't intend to find myself in. 

 

Re the network,  perhaps but it's my impression that Sprint is in a far better position. It already has all the spectrum assets it needs, it just needs time to build them out. Tmo has a network which is already suffering in some areas (at least that's my experience here) and whilst it has plans it needs to find a hell of a lot more cash to get where it needs to go. Buying more 700 lower A, the aws3, 600MHz, building out it's 4x2 mimo network across all those bands and upgrading is edge network plus expanding its footprint,  all on 1/3rd the network spend of at &t or Verizon and managing to do it all before it piles in so many subs its network can't cope. The comments legere made about Sprint subs leaving and not coming back applies just as much to tmo subs, tmo has some great areas and some areas collapsing due to load. The whole reason I'm on this forum is that my bet is Sprint will be the better provider just as soon as NV gets fully rolled out here. I could be wrong but I hope not!

Plus after the aws3 had raked in as much as it has so far what's the bets cavalier et al start wanting a lot more for that 700 MHz? 

[milan03]

tmo has some great areas and some areas collapsing due to load.

 

When you say collapsing due to load, are you talking rural/suburban areas or major metros?

 

Because in NYC around the busiest areas of Times Square, in the middle of the day or night, T-Mobile LTE will definitely slow down a bit, but you definitely won't see it collapsing. You'll still stay on LTE, get data rates in the low teens, and latency will still stay in the 30's. In my book that's still a decent user experience overall.

[danlodish345]

When you say collapsing due to load, are you talking rural/suburban areas or major metros?

 

Because here in NYC, on Times Square, in the middle of the day or night, T-Mobile LTE will definitely slow down a bit, but you definitely won't see it collapsing. You'll still stay on LTE, get data rates in the low teens, and latency will still stay in the 30's. In by book that's still a decent user experience overall.

tmobiles lte gets 60 megs a second here and their hspa+ gets in the low 20s 

[jamisonshaw125]

tmobiles lte gets 60 megs a second here and their hspa+ gets in the low 20s

Thats similar to my experience whenever I'm on b41 in an area wit upgraded backhaul. It's pretty awesome.

 

 

Sent from my iPhone 6 on Crapatalk

[Destroyallcubes]

tmobiles lte gets 60 megs a second here and their hspa+ gets in the low 20s 

And im stuck in an area with Tmobile stuck on with low teens as a max for LTE speeds ive seen, and hspa+ that well isnt very good at all. The market share here for Tmobile is very low due to them not expanding sites to meet demand. As well as the other 3 out performing them throughout East texas, kinda sad when the word Tmobile comes up at target or elsewhere and everyone cringes, amd laughs 

[S4GRU]

When you say collapsing due to load, are you talking rural/suburban areas or major metros?

 

Because in NYC around the busiest areas of Times Square, in the middle of the day or night, T-Mobile LTE will definitely slow down a bit, but you definitely won't see it collapsing. You'll still stay on LTE, get data rates in the low teens, and latency will still stay in the 30's. In my book that's still a decent user experience overall.

 

That's right.  10-15Mbps is a more than satisfactory experience.  And if that is what it looks like when well burdened, that's admirable.  And that's where Sprint needs to get.  10-15Mbps as consistent low end experience.  Some people will only be happy with peak speeds in every corner of the network.  But that is not achievable now with current technology and realistic budgets.

[richy]

Low single digits lte, 1-3 mbps in Wailuku Town. If I'm wrong wrong I'm happy but it seems either air side or backhaul is hurting them.  Plenty of other places it's between 6 and 30mbps. Then there's North and East Maui where you get no signal but that's another story. 

[milan03]

Low single digits lte, 1-3 mbps in Wailuku Town. If I'm wrong wrong I'm happy but it seems either air side or backhaul is hurting them.  Plenty of other places it's between 6 and 30mbps. Then there's North and East Maui where you get no signal but that's another story. 

Sounds like the backhaul could be the issue in your area. I know that Honolulu for instance is a 20MHz FDD LTE market, and the backhaul seems to be in place there as HoFo users have been reporting peak rates north of 100Mbps for a while.

[richy]

Sounds like the backhaul could be the issue in your area. I know that Honolulu for instance is a 20MHz FDD LTE market, and the backhaul seems to be in place there as HoFo users have been reporting peak rates north of 100Mbps for a while.

Yeah, we are on 5x5 here. They perhaps aren't throwing the switch on a wider channel until they can get the backhaul in. Given they can deploy the backhaul here quickly  (one of the few things that does happen reasonably quickly,  especially since they have fiber drops anyway so it's just a provisioning boost) I'm curious as to why the wait,  perhaps keeping their opex low,  boosting FCF prior to the Aws3 auction. It's going to be interesting to see what they end up with from that,  given the pricing they may not have many big wins. The NY 10x10 was around .3 cents per MHz per cockroach, 2.5bn is nearly tmo entire budget for the auction so I doubt they got it,  unless they are trying to make a statement.  Maybe they grabbed some 5x5 in important markets? I think Sprint was smart to sit out. 

 

My concern over tmo isn't so much that is terrible currently,  I can just see where it is starting to slip. Now if they keep adding subs at their current rate and can't keep up with the required investment they are going to fall hard. They need to buy and build,  Sprint just needs to build right now.

[danlodish345]

Yeah, we are on 5x5 here. They perhaps aren't throwing the switch on a wider channel until they can get the backhaul in. Given they can deploy the backhaul here quickly (one of the few things that does happen reasonably quickly, especially since they have fiber drops anyway so it's just a provisioning boost) I'm curious as to why the wait, perhaps keeping their opex low, boosting FCF prior to the Aws3 auction. It's going to be interesting to see what they end up with from that, given the pricing they may not have many big wins. The NY 10x10 was around .3 cents per MHz per cockroach, 2.5bn is nearly tmo entire budget for the auction so I doubt they got it, unless they are trying to make a statement. Maybe they grabbed some 5x5 in important markets? I think Sprint was smart to sit out.

 

My concern over tmo isn't so much that is terrible currently, I can just see where it is starting to slip. Now if they keep adding subs at their current rate and can't keep up with the required investment they are going to fall hard. They need to buy and build, Sprint just needs to build right now.

t mobile just needs more low band spectrum...for coverage

[EvanA]

t mobile just needs more low band spectrum...for coverage

They need capacity in some markets as well. Cincinnati being a notable one.

[maxsilver]

Tmo has a network which is already suffering in some areas (at least that's my experience here) and whilst it has plans it needs to find a hell of a lot more cash to get where it needs to go. Buying more 700 lower A, the aws3, 600MHz, building out it's 4x2 mimo network across all those bands and upgrading is edge network plus expanding its footprint, all on 1/3rd the network spend of at &t or Verizon and managing to do it all before it piles in so many subs its network can't cope. The comments legere made about Sprint subs leaving and not coming back applies just as much to tmo subs, tmo has some great areas and some areas collapsing due to load.

 

My concern over tmo isn't so much that is terrible currently,  I can just see where it is starting to slip. Now if they keep adding subs at their current rate and can't keep up with the required investment they are going to fall hard.

 

I don't think this "starting to slip" is actually happening in any significant amount, and I don't expect to see much of it in 2015.

 

T-Mobile LTE as it exists last month (fully loaded with a nation full of subscribers) won fastest average network speeds in Denver, NYC, Philadelphia, Minneapolis, and Detroit (and tied with Verizon AWS LTE in Memphis). Those are just the RootMetrics reports released this past month, from some of the largest and/or most heavily loaded markets in the nation.

 

Barring a small handful of areas they are truly spectrum crunched in (like Cincinatti), any area they *want* to do well in, they can, and so far have done so.

 

This includes other areas in your state --- T-Mobile took first place in Honolulu for data speed, data reliability, data performance, and overall performance just 4 months ago - http://www.rootmetrics.com/us/rsr/honolulu-hi 

 

It *really does* sound like they just haven't gotten around to upgrading the backhaul in your particular area yet, and that may look like they are "slipping" for you because of that.

 

Luckily, that's typically the easiest upgrade to make happen. Providers in Michigan can literally double bandwidth on my lines in about 48 hours, with a single phone call and a faxed signature (if I had the cash to do so).

 

I don't know much about the situation in Hawaii, but I really can't imagine it would be that hard for T-Mobile to get you better backhaul on those sites, if they decided to do so.

[richy]

 

I don't think this "starting to slip" is actually happening in any significant amount, and I don't expect to see much of it in 2015.

 

T-Mobile LTE as it exists last month (fully loaded with a nation full of subscribers) won fastest average network speeds in Denver, NYC, Philadelphia, Minneapolis, and Detroit (and tied with Verizon AWS LTE in Memphis). Those are just the RootMetrics reports released this past month, from some of the largest and/or most heavily loaded markets in the nation.

 

Barring a small handful of areas they are truly spectrum crunched in (like Cincinatti), any area they *want* to do well in, they can, and so far have done so.

 

This includes other areas in your state --- T-Mobile took first place in Honolulu for data speed, data reliability, data performance, and overall performance just 4 months ago - http://www.rootmetrics.com/us/rsr/honolulu-hi 

 

It *really does* sound like they just haven't gotten around to upgrading the backhaul in your particular area yet, and that may look like they are "slipping" for you because of that.

 

Luckily, that's typically the easiest upgrade to make happen. Providers in Michigan can literally double bandwidth on my lines in about 48 hours, with a single phone call and a faxed signature (if I had the cash to do so).

 

I don't know much about the situation in Hawaii, but I really can't imagine it would be that hard for T-Mobile to get you better backhaul on those sites, if they decided to do so.

Yes it is that simple here as well, assuming fiber is already lit there (which it is) and theres no capacity issues. That was pretty much my concern, if its so easy and they obviously are having peak issues, why not pull the trigger? Too many legacy non lte devices that they don't want to cramp into less refarmed spectrum? Just saving money? Honolulu probably is fast, they have 6 times our population  so they tend to get more lovin when it comes to any kind of infrastructure. They're even getting their own disney style monorail thing The tmo thing just confused me honestly, its not like the boonies were slow where it only affects few folks, theyre actually upgrading the whip aerial edge only deployments in the boonies to lte. Its 'town' that was suffering which by our standards is a lot of people. If it was so easy to do (which it should be) why not just do it?

 

Overall tmo is in a fairly decent situation right now, theres a good chance they can continue to grow and invest, however my worry was that the big two would try and drive the price of acquiring the low band spectrum they need through the roof. Theres a chance we may see a new sherrif in town in 2016 and if the auction date gets moved again we could see a change in rules to a more 'business friendly' (read bought and sold officials) situation where its a free for all over all the spectrum, in which case tmo doesn't really have the financial might of the big two. Sprint has spends, it's sugar daddy has deep pockets, would DT leverage itself even more to do the same for tmo. As I've said before, I am happy tmo is the right place for me, in my market, right now, but I have doubts about their future. The recent 'uncarrier' moves have largely been fluff, they have enough mid band right now but I have a feeling the low band is going to cost them dearly. AWS3 is approaching 40bn. If google or amazon decide to play in the 600MHz, alongside Dish, then it could get really messy.

 

Either way I am happy we have choices in who we can use! In some respects I am glad that the tmo sprint merger didn't happen as at least I have a choice besides the bad two! 

[danlodish345]

Yes it is that simple here as well, assuming fiber is already lit there (which it is) and theres no capacity issues. That was pretty much my concern, if its so easy and they obviously are having peak issues, why not pull the trigger? Too many legacy non lte devices that they don't want to cramp into less refarmed spectrum? Just saving money? Honolulu probably is fast, they have 6 times our population so they tend to get more lovin when it comes to any kind of infrastructure. They're even getting their own disney style monorail thing The tmo thing just confused me honestly, its not like the boonies were slow where it only affects few folks, theyre actually upgrading the whip aerial edge only deployments in the boonies to lte. Its 'town' that was suffering which by our standards is a lot of people. If it was so easy to do (which it should be) why not just do it?

 

Overall tmo is in a fairly decent situation right now, theres a good chance they can continue to grow and invest, however my worry was that the big two would try and drive the price of acquiring the low band spectrum they need through the roof. Theres a chance we may see a new sherrif in town in 2016 and if the auction date gets moved again we could see a change in rules to a more 'business friendly' (read bought and sold officials) situation where its a free for all over all the spectrum, in which case tmo doesn't really have the financial might of the big two. Sprint has spends, it's sugar daddy has deep pockets, would DT leverage itself even more to do the same for tmo. As I've said before, I am happy tmo is the right place for me, in my market, right now, but I have doubts about their future. The recent 'uncarrier' moves have largely been fluff, they have enough mid band right now but I have a feeling the low band is going to cost them dearly. AWS3 is approaching 40bn. If google or amazon decide to play in the 600MHz, alongside Dish, then it could get really messy.

 

Either way I am happy we have choices in who we can use! In some respects I am glad that the tmo sprint merger didn't happen as at least I have a choice besides the bad two!

i actually have verizon i get service in the pocono mountains and also i get swrvice on top of mountains

[richy]

i actually have verizon i get service in the pocono mountains and also i get swrvice on top of mountains

Same here,  I use a Verizon Lte tablet as a hot spot for the places I can't get a tmo signal. Works nearly everywhere but even the LTE is fairly slow in many places. Could just be the tablet though. 

[maxsilver]

Yes it is that simple here as well, assuming fiber is already lit there (which it is) and theres no capacity issues. That was pretty much my concern, if its so easy and they obviously are having peak issues, why not pull the trigger? Too many legacy non lte devices that they don't want to cramp into less refarmed spectrum? Just saving money? Honolulu probably is fast, they have 6 times our population  so they tend to get more lovin when it comes to any kind of infrastructure.

 

The tmo thing just confused me honestly, its not like the boonies were slow where it only affects few folks, theyre actually upgrading the whip aerial edge only deployments in the boonies to lte. Its 'town' that was suffering which by our standards is a lot of people. If it was so easy to do (which it should be) why not just do it?

 

I think they just haven't gotten around to caring about your area yet.

 

I feel your pain though -- we have a similar situation over here. T-Mobile just won 1st place on data speed in our largest city (Detroit), and the network over there is fantastic. But in the 2nd largest city (Grand Rapids) I still can't get T-Mobile service in one of the malls, or the county courthouse, or the suburban WalMart, or the south end of Downtown, etc. We're not out in the boonies -- and it's only a handful of areas that are broken. But it is our #2 city with 1.1+million people that's got these struggling areas (and it's doubly-frustrating, since legacy MetroPCS has CDMA+LTE running sites leased in each of these areas, that T-Mobile is throwing away presumably since they don't *have* to have them once they get Band 12 up and running...) 

 

I suspect they'll get to your backhaul. Eventually.

[Thomas L.]

More than anything else I'm curious as to when and where we will start seeing T-Mobile deploy on their 700mhz spectrum. I love S4GRU because I know when/where Sprint is deploying certain bands/technologies. Right now T-Mobile apparently has 1900mhz LTE and 700mhz LTE in some places but no one knows where because there is no site like this one for T-Mobile!

[milan03]

More than anything else I'm curious as to when and where we will start seeing T-Mobile deploy on their 700mhz spectrum. I love S4GRU because I know when/where Sprint is deploying certain bands/technologies. Right now T-Mobile apparently has 1900mhz LTE and 700mhz LTE in some places but no one knows where because there is no site like this one for T-Mobile!

Their L700 project just recently started but reports are slowly coming from different parts of the country. TmoNews is a decent general T-Mobile related news source, although nothing comparable to S4GRU community http://www.tmonews.com/2014/11/t-mobile-700mhz-lte-spotted-in-houston-texas/

[Ascertion]

That's right.  10-15Mbps is a more than satisfactory experience.  And if that is what it looks like when well burdened, that's admirable.  And that's where Sprint needs to get.  10-15Mbps as consistent low end experience.  Some people will only be happy with peak speeds in every corner of the network.  But that is not achievable now with current technology and realistic budgets.

10-15 is even high for my standards, I'd love to have even a consistent 5mbps so I can load pictures/videos at anytime.  Although I do like to show off B41 speedtests to my cousins who have ATT and Verizon.