dslreports logo
 
    All Forums Hot Topics Gallery
spc
Search similar:


uniqs
2386
Mele20
Premium Member
join:2001-06-05
Hilo, HI

Mele20

Premium Member

[TWC] Hawaii has 20000kbps now? Nah, speed tests are wacko

Click for full size
Oceanic TW speed test

Speedtest.net SanFrancisco
 

Web100 SantaCruz, Calif.
 

OOL FTP test
Click for full size
MySpeedAdvanced San Jose, Calif
Surfing has been slow all afternoon today. This site is painfully slow and Avira forum in Germany almost won't load. The modem has tripped off line twice this afternoon. Modem stats are fine though (upstream power level at 44dBmV).

So, I decide to do some speed tests. Oceanic's test claims my speed is 20867kbps!!! I have standard RR at 5000kbps down and 1000kbps up. So, I do a few more tests. Some show much higher speed than my cap. The ONLY ACCURATE TESTS ARE:

MySpeed (Visualware both from MySpeedPCAdvanced software I own and online test to San Jose, Calif.) and OOL FTP download test. The rest are showing junk reports.

swintec
Premium Member
join:2003-12-19
Alfred, ME

swintec

Premium Member

Do you have powerboost maybe??? Sure looks like it by the TW provided speed test.
Mele20
Premium Member
join:2001-06-05
Hilo, HI

Mele20

Premium Member

We have not had Powerboost but I suppose it could have just been turned on today. But wouldn't Powerboost affect ALL speed tests? MySpeed is not affected and it is a java test. Even OOL's FTP test should be affected in the beginning and the beginning of that test had the speed at 133KB/sec and it gradually got higher. Powerboost would have the opposite effect...high in the beginning and then dropping down. So, I don't think that is what is going on. Plus, there is nothing on Oceanic's website, or through Googling, about Powerboost being introduced to Hawaii. Could be they are testing it though but I still don't think that is what is causing the weird speeds because that should affect Visualware's MySpeed as well as the first few seconds of OOL download.

amysheehan
MVM
join:1999-12-21
Chula Vista, CA

1 edit

amysheehan

MVM

Have you tried Linkline's test via dslr ???
It has always been the most accurate for me.

Here's the link: »/speed ··· premium=

EDIT TO ADD: Latest speed test:



Mele20
Premium Member
join:2001-06-05
Hilo, HI

1 edit

1 recommendation

Mele20

Premium Member

Linkline gave me a blank page for results. But it did record on my tools test page:
28-02-2009 08:47 PM java speedtest 7579 Kbps 941 Kbps

I downloaded the trial of the latest version of MySpeedPC Advanced (now called MyConnectionPCBusiness) and wow! I saw no reason to upgrade in the past from the version I have (1.1) until this 3.0b version. It is a huge improvement and will be worth the cost I think. I've been playing with it for a couple of hours now testing at all four locations (3 Mainland and 1 in London, England).

What is happening is RAMPING UP on all tests beyond California. No ramping up to SanJose which explains the speed I was seeing earlier on all the background tests I did to San Jose using othe older version of MySpeed Advanced that I own. There is no Time Variation test on my older version of MySpeed. Visualware came out with it very recently specifically to test connections that have some form of packet shaping. The San Jose test showing no ramping up could be false, and due to some problem with that server currently, but it makes sense that because California is closer than Chicago, Virginia and England that Oceanic would see no need to ramp up for the West Coast but would for locations further away from Hawaii.

No ramping up to the West Coast though means that my speed to San Jose is much slower than my speed to Virginia or London even with excellent QOS to San Jose and less so to further locations. That is a bit screwy. I hope there is something wrong currently with the San Jose server and that the ramping up will be seen on all tests.

When Powerboost was first available to some RR divisions I think I recall everyone posting that the increased speed is at the BEGINNING of a download, whereas, for me it is TOWARD THE END. I don't know if this is Powerboost and it is being applied to the Hawaii network in a different manner from the way it is on the Mainland or if this is just Oceanic messing around while they ready locations outside Oahu for Turbo and Turbo Plus.

Summary of Time Variation test results
Variation From 2722 Kbps
Variation To 4091 Kbps
Variation window 45 Seconds

"This is what Visualware says about ramping:

The basic need to regulate network traffic only comes about because networks are often oversubscribed by the provider and congested. Just as cars are affected by traffic lights because of the need to regulate traffic flow, especially prevalent during peak loads such as the rush hour. Similarly Internet connections also use flow policies to manage load. Data flow control (packet shaping) is common across the internet and one of the most common ‘tricks’ used is ‘ramp-up’ and ‘ramp-down’.

Regulation is not always a good thing, by its nature and intent regulation is there only to balance networks and is often used to hide oversubscription problems.
Analysis of Time Variation test results

1. Time variation from 2722 Kbps to 4091 Kbps has climbed by nearly 50% over 45 Seconds which may indicate ramp-up regulation is in effect. The time variation test measures a connection to detect if the speed of the data is being regulated to go faster or slower over time. This is often referred to as packet shaping although there are many different types of packet shaping methods. If ramping either up or down is detected it implies that the ISP is attempting to regulate traffic to provide a better experience or to hide a bad experience. As a typical example, High latency networks, such as satellite, often implement ramp up policies. This has the affect of speeding up the data flow when a lot of data is detected."

The above was part of a full test to Visualware's server in Virginia .

Application speed was measured at:

3.83mbps down
986kpbs up

Connection bandwidth capacity: 29600Kbps
Capacity download speed: 13800 Kbps

Application speed is the one that matters and as you can see Visualware's test is the only one not fooled by packet shaping being used by Oceanic.

After writing the above about no ramping up to San Jose, I just did another test to there and the Variation test result shows SEVERE problems with my connection indicating extensive corruption. Whether this means that there is currently a problem with that server or something else I don't know but I should not be getting higher speeds to Virginia and London, England when the route speed and average response times for each hop are fine to San Jose according to MyConnection test. So, there may be ramping up also to the West Coast but no way to tell currently.

edit:typos
Mele20

Mele20

Premium Member

Click for full size
I've lost 50% of my speed to Oceanic's test on Oahu.

Now the Linkline test is showing:

28-02-2009 10:57 PM java speedtest 3258 Kbps 819 Kbps

That's less than 1/2 the download speed it was showing a couple of hours ago.

MyConnectionPC to Virginia is showing about the same download speed as earlier but even clearer indication of Oceanic using Ramp-up:

"Analysis of Time Variation test results

1. Time variation from 2477 Kbps to 4176 Kbps has climbed by nearly 69% over 3 Seconds which may indicate ramp-up regulation is in effect."

Tweaked
@charter.com

Tweaked

Anon

How is the computer you are testing tweaked?

»/tweaks
Mele20
Premium Member
join:2001-06-05
Hilo, HI

Mele20

Premium Member

I was testing 3 computers: VMWare Workstation 5.5 with XP Pro SP2, VPC 2007 with Vista Ultimate and my host computer with XP Pro SP2. Right after I wrote the above, I thought about the RWIN. The VMWare XP computer was the one I got the speed at Oceanic's own test at 50% lower so I looked at RWIN for it (used Tweaks here) and it was set at 75000. I knew on my host computer that it was set at 125560. So, before I changed it on either of those machines, I went back to the Vista computer and did the Web100 test to Santa Cruz, California so I could see in the statistics what RWIN Vista was using. It set the RWIN for that test at 315136 and gave me a high speed (a little over twice what I usually get on that test) indicating ramping.

So, I played around with the Tweaks test here on both XP machines. If I told Tweaks that my cap is 5000kbps then Tweaks told me that when I set the RWIN to 315136 like Vista had it that was way too high. So, I told Tweaks I had 10,000kbps down and Tweaks still said that was too high an RWIN. Tweaks said no higher than 20020 so that is what I set it at on the VMWare XP virtual machine. On the Host XP machine, I have RWIN at 315136 like Vista sets it for West Coast speed tests.

When I set it to 20020 on the XP virtual machine then the Oceanic speed test showed the higher speed. But on the host XP machine, I am still getting HORRIBLE speed to San Jose California. I get faster speed to Virginia.

I just did OOL speed test twice back to back and on the first one I saw ramping up so the final speed at 81% done was nice at 711KB/sec and the second one was ramped down and it looked awful toward the end at 70% done at 328KB/sec.

Whatever this is, Powerboost I don't think because my speeds are much worse now, my opinion of Oceanic and Time Warner in general has dropped greatly. This is a way to CHEAT AND LIE to the user and actually give them worse speed. Sadly though, since most will not get Visualware's MyConnection nor read their new white papers that expose the sad state of most speed tests on the internet the average user is going to believe they now have fabulous speed.

hobgoblin
Sortof Agoblin
Premium Member
join:2001-11-25
Orchard Park, NY

hobgoblin

Premium Member

"my opinion of Oceanic and Time Warner in general has dropped greatly."

Thats funny coming from you.

You almost full time trash Time Warner and Oceanic!

Hob