dslreports logo
 
    All Forums Hot Topics Gallery
spc
uniqs
18
Dr Gary
join:2001-01-13
Geneva, IL

Dr Gary to MonnieRock

Member

to MonnieRock

Re: [Connectivity] Extreme 50/15 but getting 10

I may be missing something here but your ShaperProbe upload measurement seems way off. It is showing only 1,232 kbps? That is only a little over 1 Mbps...nowhere near 10 Mbps.

On my Comcast 27/7 Business Class service my ShaperProbe upload is 7,914 kbps.

Why would your ShaperProbe result look so much slower?

plencnerb
Premium Member
join:2000-09-25
53403-1242

plencnerb

Premium Member

I was seeing the same thing. This leads me to believe that maybe it is a modem configuration issue (wrong config file sent to your modem), or as GSTEXEC1 See Profile has said, your local CMTS not being upgraded yet to support that speed plan.

For reference, I ran a ShaperProbe test just now. My results are below. By the way, my plan is "Performance w/powerboost", which in my area is 25/4.

-----------------------------
DiffProbe release. January 2012. Build 1008.
Shaper Detection Module.

Connected to server 83.212.4.23.

Estimating capacity:
Upstream: 4160 Kbps.
Downstream: 25267 Kbps.

The measurement will take upto 2.5 minutes. Please wait.

Checking for traffic shapers:

Upstream: Burst size: 5001-5076 KB;
Shaping rate: 2128 Kbps.

Downstream: Burst size: 9576-10065 KB;
Shaping rate: 15978 Kbps.

--Brian

AnonMan
@sct.com

AnonMan to Dr Gary

Anon

to Dr Gary
ShaperProbe is not accurate by any means.

People say use it and what not but like any other service it all depends on the hosting server and which server you connect to.

ShaperProbe has a list of IPs and sometimes mine connects to a slow one and sometimes a fast one.

It takes a few runs across different IPs to know an decent avg. He probably just ran it once.

EG
The wings of love
Premium Member
join:2006-11-18
Union, NJ

EG to plencnerb

Premium Member

to plencnerb
said by plencnerb:

By the way, my plan is "Performance w/powerboost", which in my area is 25/4.

-----------------------------
DiffProbe release. January 2012. Build 1008.
Shaper Detection Module.

Connected to server 83.212.4.23.

Estimating capacity:
Upstream: 4160 Kbps.
Downstream: 25267 Kbps.

The measurement will take upto 2.5 minutes. Please wait.

Checking for traffic shapers:

Upstream: Burst size: 5001-5076 KB;
Shaping rate: 2128 Kbps.

Downstream: Burst size: 9576-10065 KB;
Shaping rate: 15978 Kbps.

--Brian

Looks like Performance is 16/2 sustained with Powerboost bursts to 25/4 in your area.

pflog
Bueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA

1 recommendation

pflog to AnonMan

MVM

to AnonMan
said by AnonMan :

ShaperProbe is not accurate by any means.

People say use it and what not but like any other service it all depends on the hosting server and which server you connect to.

ShaperProbe has a list of IPs and sometimes mine connects to a slow one and sometimes a fast one.

It takes a few runs across different IPs to know an decent avg. He probably just ran it once.

Obviously it can vary. But the variance, in my experience, has been very minimal.

Johkal
Cool Cat
MVM
join:2002-11-13
Pennsyltucky

Johkal to plencnerb

MVM

to plencnerb
Maybe I'm not alone with ShaperProbe reporting inaccurately. ShaperProbe has been borked for me for months.

My tier: Blast 50/10

DiffProbe release. January 2012. Build 1008.
Shaper Detection Module.

Connected to server 80.239.168.233.

Estimating capacity:
Upstream: 1660 Kbps.
Downstream: 21223 Kbps.

The measurement will take upto 2.5 minutes. Please wait.

Checking for traffic shapers:

Upstream: No shaper detected.
Median received rate: 1624 Kbps.

Downstream: Burst size: 17692-18157 KB;
Shaping rate: 11312 Kbps.

pflog
Bueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA

pflog

MVM

Interesting. Have you compared traceroutes to the server IPs where you get funky results? Or do you get low results every time?

Johkal
Cool Cat
MVM
join:2002-11-13
Pennsyltucky

Johkal

MVM

I get low results all the time. I can download/upload large files and see correct transfer rates for my provisioning. I have not figured out why ShaperProbe is inaccurate for me now.

pflog
Bueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA

pflog

MVM

said by Johkal:

I get low results all the time. I can download/upload large files and see correct transfer rates for my provisioning. I have not figured out why ShaperProbe is inaccurate for me now.

Hmm, weird. Have you emailed Partha about it? Very responsive in my queries and very knowledgeable. If you care to figure it out of course. I would hate it if I couldn't rely on one of the only speed tests runnable on non-windows/mac computers.

Johkal
Cool Cat
MVM
join:2002-11-13
Pennsyltucky

Johkal

MVM

I've been lax with looking into it further. I should investigate since there may be others who may be seeing this. Might be worthy for the benefit of others.

EG
The wings of love
Premium Member
join:2006-11-18
Union, NJ

EG

Premium Member

I think that these are isolated cases and more so the exception and not the rule. IMHO, it is still a gold standard measuring tool and I will still be recommending its use when I want a more accurate reference than any of the run of the mill Flash and Java based speed testing sites.

pflog
Bueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA

pflog

MVM

said by EG:

I think that these are isolated cases and more so the exception and not the rule. IMHO, it is still a gold standard measuring tool and I will still be recommending its use when I want a more accurate reference than any of the run of the mill Flash and Java based speed testing sites.

You would think these speed test sites would come up with a way to measure the same kind of thing, then display the peak (powerboost) speeds as well as the sustained speeds.

But as we know, it would take more time and more bandwidth on the part of the servers hosting the speed tests, so that's probably a deterrent. But given many cable providers have implemented powerboost, I'm surprised there are no good web-based speed tests sites that will break it down like shaperprobe does!

Johkal
Cool Cat
MVM
join:2002-11-13
Pennsyltucky

Johkal to EG

MVM

to EG
I agree. I've not looked into because it works for the majority. I just use other resources.

pflog
Bueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA

pflog

MVM

said by Johkal:

I agree. I've not looked into because it works for the majority. I just use other resources.

If you know of a web-based test that can detect and report PB and also sustained speeds, please share!

Johkal
Cool Cat
MVM
join:2002-11-13
Pennsyltucky

Johkal

MVM

This one isn't bad for Download check: »www.thinkbroadband.com/download/

pflog
Bueller? Bueller?
MVM
join:2001-09-01
El Dorado Hills, CA

pflog

MVM

said by Johkal:

This one isn't bad for Download check: »www.thinkbroadband.com/download/

Ah ok, so these are just large files to download? I have a couple of mirrors I use to do that, but this is a nice link, thanks.

I was thinking of something more like shaperprobe, where there's a flash/html5/java UI and it runs a longer test or uses udp as shaperprobe does to give an idea of PB and post-PB speeds.

Johkal
Cool Cat
MVM
join:2002-11-13
Pennsyltucky

Johkal

MVM

I wish there was a ShaperProbe equivalent. I haven't found another. I just check with large file samples.

EG
The wings of love
Premium Member
join:2006-11-18
Union, NJ

EG to Johkal

Premium Member

to Johkal
said by Johkal:

This one isn't bad for Download check: »www.thinkbroadband.com/download/

FWIW, I get download speeds that are off by 30 Mbps with the largest DL file there. Guess that that server does not open up multiple concurrent sockets / streams to really reveal the true capability of the pipe ?

telcodad
MVM
join:2011-09-16
Lincroft, NJ

telcodad to Johkal

MVM

to Johkal
For speed tests with large file downloads/uploads, I use the TestMy.net site: »testmy.net/

To get past the PowerBoost effects, you can custom select test file sizes of up to 200MB on the downstream, and up to 33 MB on the upstream.

tshirt
Premium Member
join:2004-07-11
Snohomish, WA

tshirt to Johkal

Premium Member

to Johkal
That one seems to top out around 20Mbps for me at home (but I'll try it at some other locations)
»www.linode.com/speedtest/ works great if you open multiple saves, either the same server or around the world (which mimics real web pages and/or P2P sources) I consistently hit 37.50Mbps on Blast! during the boost phase here.

EG
The wings of love
Premium Member
join:2006-11-18
Union, NJ

EG to telcodad

Premium Member

to telcodad
said by telcodad:

For speed tests with large file downloads/uploads, I use the TestMy.net site: »testmy.net/

To get past the PowerBoost effects, you can custom select test file sizes of up to 200MB on the downstream, and up to 33 MB on the upstream.

FWIW, that site has never ever given me anywhere near accurate results for over five years now telcodad See Profile.

Johkal
Cool Cat
MVM
join:2002-11-13
Pennsyltucky

Johkal to telcodad

MVM

to telcodad
I've had that one bookmarked for a while. It's good as you say, but I don't use it. I just like simple down & dirty tests.

telcodad
MVM
join:2011-09-16
Lincroft, NJ

telcodad to EG

MVM

to EG
said by EG:

said by telcodad:

For speed tests with large file downloads/uploads, I use the TestMy.net site: »testmy.net/

To get past the PowerBoost effects, you can custom select test file sizes of up to 200MB on the downstream, and up to 33 MB on the upstream.

FWIW, that site has never ever given me anywhere near accurate results for over five years now telcodad See Profile.

Yeah, you've told me that before, but it seems to work pretty good for me. For my Performance 16/2 connection, it just gave me a result of 14.8/2.5 (with 100MB/25MB test files).

But to each their own, I guess.

Johkal
Cool Cat
MVM
join:2002-11-13
Pennsyltucky

Johkal to EG

MVM

to EG
said by EG:

said by Johkal:

This one isn't bad for Download check: »www.thinkbroadband.com/download/

FWIW, I get download speeds that are off by 30 Mbps with the largest DL file there. Guess that that server does not open up multiple concurrent sockets / streams to really reveal the true capability of the pipe ?

A bit of a crap shoot using anything other than ShaperProbe. Fortunately that one works for me. I have to figure this issue out.