Intro

[Donald DaCosta]

I've used this service to test a newly subscribed TWC, "up to 20 mb/s" ISP service. The "up too" should have ben a red flag but the at&t (bellsouth) "fast access DSL" infrastructure in my neighborhood could not provide > 1.5 mb/s, my chosen at&t long distance phone service was NOT unlimited and with the packaged Directv service included, with no premium channels, cost about $185/mos.

 

TWC "up to 20 mb/s" internet with unlimited long distance and cable, including Showtime is costing about $153/mos. A promo price but a pretty good deal huh? Aside from the relative immunity to heavy rain and maybe snow (don't get much of that here) TWC sucks in comparison to Directv; in channel choices, the reliability of the CISCO cable interface box, their menu, channel choice, search and record functions are awkward, unwieldy and antiquated when compared to Directv or Comcast, others I've had extensive and limited experience with respectively. You get what you pay for.

 

In addition there are numerous, daily reminders of what the "up to" means. Using Time Warners speed test function, download speeds considerably below 1 mb/s are often encountered; especially annoying when the video that's captured your interest suddenly freezes. Measured download performance upwards from 10 mb/s on a single user, home wireless network is convincing evidence that these wide ranging speed anomalies are not caused by in house problems. Performance has also been tested with a direct to cable modem connection with identical results.

 

Upload speeds are adequate, consistently around 1.9 mb/s, and have not been a problem.

 

That long winded introduction is why I joined testmy.net. I intend to run automated speed tests on a daily basis and present the data to TWC technical. My expectations are very low but, for what it's worth, they will be shared with the members on this forum who may be having similar problems.

 

Don D.

 

 

 

[TriRan]

Welcome to testmy.net Don

Sounds like your on the right track! Keep us updated on your situation I would love to hear how it plays out

-mark

[Pgoodwin1]

I'm on TWC Cincinnati. The first of the year they upgraded their services. I'm on the turbo 15 Mbps down, 1 Mbps up plan. Since the first of the year, I've had pretty consistent results testing here showing that they are meeting the 15 Mbps down. I get about 0.8 Mbps up speeds measured here. Prior to the first of the year upgrade, I was only getting about 8 Mbps down and the same 0.8 up speed I got prior to the upgrade. Their old plan said up to 10 Mbps, and I very rarely ever got 10 except for an occasional burst. The newer service is consistently right around the 15 Mbps speed they advertise. This upgrade was likely only for the TWC Cincinnati area though. Your results may be worse if the equipment and service they're providing you is like the older equipment was here. They didn't change my cable modem, only the service on their stuff outside my house.

[CA3LE]

Hey Don,

 

Welcome!

 

If you like DIRECTV for television... and you like Time Warner for Internet... why not just do it that way.  You don't have to bundle... saving money and getting a worse product isn't really saving money in my eyes.  Personally, I like the cable service in my area (Cox Communication, fully upgraded DOCSIS 3.0 area), but we want Sunday ticket.  DIRECTV bundles with DSL providers, varies by area.  Now, there is no way that I'd do DSL... I don't care what level of service they have, if it's not coax or fiber to the home I don't want it.  So I go with Cox for my home Internet and DIRECTV for television.

 

Plus, my handle would make much less sense without a CABLE connection.  Then again, if I end up with fiber... that's a cable too.  

 

The variation in the speed during your testing displays the issue you're having.  Gathering more information from the auto speed test like you have planned is a great idea.  It will take readings throughout the day and night and you can see the variation during the day... as well as deeper within each result.  It provides a lot of evidence that can clue them into the fact that you're having a real issue.   ... they WILL make you use their test as well, if you surf around and read what's written here you'll understand why the results here are more representative of your actual speed than the tests they force their customers to use.  Don't let them fool you into thinking it's working right just because it's working right to their servers... the Internet doesn't work that way, my idea is that your speed test shouldn't either.

 

 

See those dips... good, clean, uninterrupted connections have less variance. (examples taken out of the live database, random users)

 

See how the middle holds out steady through the test, good result.  But a flat line like that can also mean bottleneck.  Luckily for this person it's a fast result.

 

This one is still good, there is variance but it's bursting and still fast... never really dropping out.  Little network protocol blips.

 

Very typical result on a cable connection.  See the powerboost losing its steam.  Although there's a pretty big variance this person would be expected to have a great online experience.  As with the other examples his network probably has enough bandwidth to stream HD to a bunch of devices without an issue.

 

Make sure that it's not your computer or network.  If you have a router, remove it and test directly to the computer (you need to unplug and reset the modem to get the new Mac address from the computer... do the same to hook it back up to the router).  Also make sure that your TCP stack is tuned, especially if... well, now I see you're running Ubuntu, cool... you should be fine.  For others, TCP Optimizer (free from speedguide.net) really helps with Windows TCP settings.

 

Test from more than one location with speed test mirrors and also try the multithread speed test.  I offer experimenting in many ways, utilize them... testing in different ways to different locations can give a broader picture of what's going on.  For example, if you get nearly identical results to many locations you or your ISP might want to look for the problem closer to home (because the common connections closer to home are probably at fault).  

 

You're at the right place to test it.  Other speed tests out there are not nearly as scientific as TMN... this site is truly meant to help you get more out of your connection not just output a number.  

 

I'll keep an eye on your results as they come in...

[CA3LE]

I was looking at those screenshots and realized that there needed to be a modification to the variance percent.  It now omits the first and last readings.  This update corrects variance on all previously logged results as well.

 

example from above

 

 

corrected

 

 

there are many instances that great connections have a variance at the beginning and end of a test... the middle is what we want to know, so that's what's taken into account in that calculation now. Also realize that the TiP data itself is omitting the first 5% and last 5% of the result.  As always the overall result logged to the database takes everything into account.

 

[Pgoodwin1]

For reference, here's some fairly typical performance with TWC (Cincinnati). Dallas server, not multithreaded

TIP Summaries 07/27/13 ~ 7:30 PM. 15 MB download test, Dallas server, not a multi thread test

Minimum :: 7.07 Mbps | Middle :: 14.79 Mbps | Maximum :: 25.86 Mbps

Minimum :: 5.73 Mbps | Middle :: 14.38 Mbps | Maximum :: 24.97 Mbps

Minimum :: 6.88 Mbps | Middle :: 16.58 Mbps | Maximum :: 27.16 Mbps

Variations on all three runs were 65-70% - pretty typical for my TWC during prime time evening hours. In off hours, there is a little less variation with averages in 14-16 Mbps range. But TWC has always had a lot of variation. I can usually tell by how slow or fast Netflix or Amazon Prime Instant video downloads/starts shows.

My average results for those three runs above:

Sat Jul 27 2013 @ 7:32:07 pm 25 MB 14.47 Mbps 1.81 MB/s

Sat Jul 27 2013 @ 7:31:03 pm 25 MB 12.7 Mbps 1.59 MB/s

Sat Jul 27 2013 @ 7:28:33 pm 25 MB 13.77 Mbps 1.72 MB/s

[Pgoodwin1]

I offered the above data to show what I think is normal and typical for TWC. Obviously the TWC in your area won't be the same, but if you see performance that looks like mine, I wouldn't call it atypical of Time Warner's service. I've never seen TWC's cable service flat like CA3LE's showing. Even before this site had the graphic display of the TIP, you could watch the progress bar during the test and see it speeding up and slowing down. With TWC, I'm happy that they now at least provide me with an average that is close to what they advertised. Before their 1st of the year upgrade here, I was always 20% to 50% below what they advertised as their minimum.

[Pgoodwin1]

This is TWC here over a two week period at various times of day.

[CA3LE]

Update on Donald's results.  6 hours right now.

 

 

up to ~ 5x slower in the valleys than the peaks.  That's significant for sure.