US Telco Support > AT&T > AT&T U-verse > ATT U-Verse: Actual Internet Speed vs Advertised

ATT U-Verse: Actual Internet Speed vs Advertised

I wouldn't say it's typical. I have the 12Mb also, at speakeasy, I'm showing from Dallas:
-Download Speed: 10868 kbps (1358.5 KB/sec transfer rate)
-Upload Speed: 1409 kbps (176.1 KB/sec transfer rate)
from New York:
-Download Speed: 11577 kbps (1447.1 KB/sec transfer rate)
-Upload Speed: 1410 kbps (176.3 KB/sec transfer rate)

Inside the modem, I see:
-------------- Download ---- Upload
User Rate --------13152 kbs 1536 kbs
Max User Rate --- 35436 kbs Not Available

On my U-verse I would get roughly 5-10% more than the advertised speed.

reply to WildOne
What are they playing, PC or Xbox Live, PS3? Online shooters (COD BO2)? You might have line issues causing problems and also to note if you are on VDSL2 (This is the TV version of U-verse) it is interleaved so you will have higher latency (ping) than cable which usually will cause lag. Though that speed is good enough to play online they might be getting alot of lag/latency due to line issues/routing/high pings etc. so I would contact Tier 2 here or call them. I would not settle for those speeds. I have had U-Verse for 4+ years and my speeds are always fairly close to advertised speeds and I play on Xbox Live and it runs fairly smooth for me.

»AT&T Direct

I would also run a ping test

reply to Frodo
Frodo,

thanks for the reply.

I'll try the test at speakeasy also.

When you say "Inside the modem, I see", to what are you referring to? Do you need a special modem to "see" this? Is that another/different test? Please explain.

Best,
Bill

reply to WildOne
Are you running these tests over wired or wireless? Try wired if you were trying it over wifi.

In your area we always ran short on the 3801s, the black modem, the 3800's, silver modem, have degrading wifi issues.

reply to Nuckfuts
Hey N,
they play mainly Battlefield and League of Legends on PCs.

How do I find out if I'm on VDSL? I notice that my tests since switching to UVerse do, in fact, show higher latency (~25ms with CC vs ~55 with UV).

Who/what is "Tier 2"? Sorry for all the questions; just trying to get this figured out.

Best,
Bill

reply to UverseTech
Hey UVTech,

thanks for the reply.

We have Cat5 and wireless through the house. I'm running the tests on a computer with a wired connection.

The kids play on computers w/ a wireless connection. I'll get them to try wired. But, to me, the bottom line is that they were playing on the wireless connections prior to the switch from Comcast...nothing's changed except the provider. Now, with UVerse, they are complaining about slow connections. Now, a complaining teen isn't something I take too seriously but the data backs it up; we're significantly slower than we were.

The modem that came w/ UVerse is a silver modem; a 2Wire model 3800 HGV-B. How severe a problem is the "degrading WiFi? Is this something that needs to be replaced/upgraded?

Best,
Bill

The 3800s should have been removed 3 years ago when the 3801s came out. The signal goes out slowly and eventually just goes.

If possible see if the support folks would ship you a 3801 or make them roll a tech and get you another modem. You will be serviced out of the Presidio garage and make sure they bring the 3801. This garage was out of these sporadically. You have pretty good outside plant there in the Sugarland area, that was my fav place to go when i was there, never had line issues.

reply to WildOne

reply to UverseTech

reply to WildOne
I put the AT&T Direct link in my post. They are very good at making sure your line all the way to the RG is good. They can run a line test to make sure. They are much better than Tier 1 support. Do you know which modem you are on (model #)? Like I said, you will get higher pings with U-Verse. If you have VDSL you can download this program and post your error table and bitloading graph so we can see if you are having issues.

»uvrealtime.com/

reply to WildOne
This was roughly my experience with them, getting 40-60% of what I paid for. When you called, they insisted everything was fine, and then would try to upsell you on the 24MB package.

I did upgrade at one point from 12 to 18 and the speeds did increase, which made me very suspicious why they couldn't manage better speeds on the lower package.

reply to WildOne
I get the 18 and regularly get between 15-16

It isn't 18 but....frankly, I will take it since so many here complain of barely getting half of what they are promised....and yes, I know it is not technically promised but hell, we are paying for it.

Thanks for the post Van.

Are you using DSLREPORTS for measuring your speed?

I'm just astounded by the differences in speed measured by the various sites. I retested again today (after I moved to 24 mb service) DL/UL (Mbs):

DSLREPORTS: 2.7/2.15
ATT.com/speedtest: 22.6/2.7
speakeasy: 23.0/2.9

I've always thought of DSLREPORTS as being the most objective; they have the least to gain by doing any number "fudging". I'm always suspect of a site that has a vested interest in the outcome (ie att.com).

reply to WildOne
You should see roughly 92% of your speed when doing speed tests due to ethernet overhead (not AT&T's fault). Of course it comes down the size of the packets really in the end and what your doing. There is no way AT&T would be giving your a quarter of your speed on purpose.

Looks like the speakeasy and AT&T test speeds are the most reliable and line up with what i said above.
--
My Blog. Because I desperately need the acknowledgement of others.

The Judd Family site!

reply to WildOne
I had U-verse in the past,and decided to switch to Comcast because I couldn't stand the fact that I was paying for a 24mbps service and was getting a fraction of that. The internet part on ATT is shared with the tv side and eats some of your bandwidth when you are watching HD programing on multiple devices. Since I switched to Comcast I been a happy camper and my speed is always steady with no powerboost.
I may consider switching back to ATT when they offer something like this.

reply to WildOne
you need to test when no one is playing or downloading.

reply to Nuckfuts
jitters matters the most in games. if latency is like 30ms but jitter only 1-3ms thats pretty good!