Speed Issues

[ecksmen]

Hi, new to the forum so be gentle! [wink]

I've been on cable now for a couple of years on and off. I'm in Cardiff and are on the 10mb connection.

I'm a little concerned over the speeds I've been getting recently, they seem to vary considerably.

I know NTL use proxies, which is my main concern in determining my actual speed of my connection in the real world. What tests can I conduct to prove whether my connection is running as it should be?

Let me explain, I used to work for 3rd line tech support for a certain dsl company. Users with slow connections could connect using speedtester@speedtester.bt.com (IIRC) and could navigate to speedtestet.bt.com where they could conduct a speed test, if the results fell within certain margins would prove whether the connection was running as it should. May I add those margins are terribly slow, 2mb connection IIRC could run as slow as 400kbps!

What sites can I go to that will be a true indication of my connection?

If I goto the likes of microsoft and say download sp2 (250mb ish) it'll come down at a decent rate, I guess it's cause it's cached in the NTL proxy.

If I go to dslguide.org.uk speedtest I struggle to get over 200kbps. I know the bandwidth isn't garunteed on that site, so fair play dslguide.org.uk could simply be overrun.

I jumped onto speedtester.bt.com and again about the same speed of dslguide.

Reboot the modem your saying? Done.

Let me explain my home network.

I have a linsys wrt54g wireless router, into which is connected a server using two 100mb ethernet cables (don't ask). Then on the wireless side I connect to a macbook (brandnew), two PC's both running xp pro and sometimes a dell pda.

All the machines are well maintained no viruses etc... I've even connected the macbook and the server (suse10.1) independantly straight into the cable modem bypassing the router and the results are the same. Poor.

What I'm looking for is some sort of tests I can conduct that I can take to the "wonderfull" ntl tech support to get something sorted. Between proxies, peak times etc... It's not really helping.

What I suspect is that if it's cached by the proxie then I will get a good download speed. But what about all the stuff that isn't cached?

What test can I conduct that will give me a true indication of the speed on my connection?

---------- Post added at 23:47 ---------- Previous post was at 23:16 ----------

Ok so I am now trying out the different speed test's on that are listed on this website. Guess I should have read more. Will post results... :Z

[James Henry]

What kind of speed variation are you seeing?

Is it worse at any particular point in the day?

Is it possible to isolate it to one country perhaps? IE if you are seeing poor downloads from a couple of sites doing some traceroutes to see if there's a common point in all the munted sites?

[ecksmen]

What kind of speed variation are you seeing?

Is it worse at any particular point in the day?

Is it possible to isolate it to one country perhaps? IE if you are seeing poor downloads from a couple of sites doing some traceroutes to see if there's a common point in all the munted sites?

I'm a work at the moment (working nights, grrrr) - I justed remote desktopped into the server, so that would use a little bandwidth (all settings on the lowest). I started up the test

http://homepage.ntlworld.com/robin.d.h.walker/speedtest.html

And it's showing as full speed, 10mb. Bare in mind I've never tested this late before.

Think what I'll do is run several tests, dslguide, speedtester.bt.com and the robin.d.h.walker one maintaining records over the next week.

During the day, I can't remembet the last time I saw a good download speed. I'm going to run a few more tests remotely (not ideal I know)...

I know it's not much info to go on, but I'll be running that test now for a couple of days from the server with no other machines connected and see if I can get to the bottom of this.

At least now I have a few more bandwidth tests at my disposal.

---------- Post added at 01:46 ---------- Previous post was at 01:40 ----------

Wed, 14 Jun 2006 00:46:19 UTC
1st 128K took 110 ms = 1191564 Bytes/sec = approx 9914 kbits/sec
2nd 128K took 130 ms = 1008246 Bytes/sec = approx 8389 kbits/sec
3rd 128K took 130 ms = 1008246 Bytes/sec = approx 8389 kbits/sec
4th 128K took 81 ms = 1618173 Bytes/sec = approx 13463 kbits/sec

[Chrysalis]

use ftp for proxyless testing.

[BrightonJohn]

http://www.ntlworld.com/toolsservices/tiermigration/

From my reading of this page it looks as though NTL WON'T be doing anything about the older set top boxes from PACE - they'll just carry on at 1 meg - is that how others are reading this?

[James Henry]

http://www.ntlworld.com/toolsservices/tiermigration/

From my reading of this page it looks as though NTL WON'T be doing anything about the older set top boxes from PACE - they'll just carry on at 1 meg - is that how others are reading this?

No idea what that has to do with the original poster's thread and speed problems, but yes that's correct. 1M isn't being upgraded for anyone it's nothing personal to those on a Pace.

[ecksmen]

Out of interest, what do NTL class as an acceptable speed for 10mb before they consider doing anything?

I don't expect to get to get 10mb all the time, but currently I only see that at night. This morning at 09:30 the only time I could test due to other obligations (like sleep lol) I was benching at arround 4mb and less.

Not being funny or anything, but I could save a lot of cash and downgrade if this level of service is acceptable by ntl...

[James Henry]

Out of interest, what do NTL class as an acceptable speed for 10mb before they consider doing anything?

I don't expect to get to get 10mb all the time, but currently I only see that at night. This morning at 09:30 the only time I could test due to other obligations (like sleep lol) I was benching at arround 4mb and less.

Not being funny or anything, but I could save a lot of cash and downgrade if this level of service is acceptable by ntl...

I don't think they have any set in stone acceptable minimum however 4Mb/s at 9:30am is poor. What are you testing with? Try http://www.adslguide.org.uk as well as this site.

Seeing bad speeds due to congestion that immediately get better after midnight is weird considering that network load would generally be lower at 9:30am than at 12 midnight...

How's the latency? Any packet loss?

[ecksmen]

Thu, 15 Jun 2006 00:04:51 UTC
1st 128K took 160 ms = 819200 Bytes/sec = approx 6816 kbits/sec
2nd 128K took 160 ms = 819200 Bytes/sec = approx 6816 kbits/sec
3rd 128K took 160 ms = 819200 Bytes/sec = approx 6816 kbits/sec
4th 128K took 161 ms = 814112 Bytes/sec = approx 6773 kbits/sec

---------- Post added at 01:09 ---------- Previous post was at 01:08 ----------

Reply from 82.133.85.65: bytes=32 time=15ms TTL=51
Reply from 82.133.85.65: bytes=32 time=15ms TTL=51
Reply from 82.133.85.65: bytes=32 time=18ms TTL=51
Reply from 82.133.85.65: bytes=32 time=16ms TTL=51
Reply from 82.133.85.65: bytes=32 time=15ms TTL=51
Reply from 82.133.85.65: bytes=32 time=16ms TTL=51
Reply from 82.133.85.65: bytes=32 time=15ms TTL=51
Reply from 82.133.85.65: bytes=32 time=17ms TTL=51
Reply from 82.133.85.65: bytes=32 time=17ms TTL=51
Reply from 82.133.85.65: bytes=32 time=28ms TTL=51
Reply from 82.133.85.65: bytes=32 time=15ms TTL=51
Reply from 82.133.85.65: bytes=32 time=19ms TTL=51
Reply from 82.133.85.65: bytes=32 time=15ms TTL=51
Reply from 82.133.85.65: bytes=32 time=15ms TTL=51

Ping statistics for 82.133.85.65:
Packets: Sent = 50, Received = 50, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 14ms, Maximum = 42ms, Average = 18ms

---------- Post added at 01:13 ---------- Previous post was at 01:09 ----------

Broadband Speed Test Results
Results from broadband speed test recorded on Thursday, 15 June 2006, 01:11.

Your Connection
Speed Test Results

Downstream 5,330.7 Kbps ( = 5.2 Mbps )
Upstream 325.8 Kbps ( = 0.3 Mbps )

Upstream has always been good - I'm conducting these remotely, using vnc. I've turned off all images within internetexplorer, so it's only updating my screen with the text changes. I've disabled the wireless side to my router, so we know it's not someone wireless hacking and stealing bandwidth.

---------- Post added at 01:21 ---------- Previous post was at 01:13 ----------

1.5 megabits per second - On bandwidthplace.com, selected cable and the uk, so should be using all the correct servers etc... I think this result is a little off tho. Further testing... going to do the old sp2 from M$... brb

---------- Post added at 01:24 ---------- Previous post was at 01:21 ----------

791kb/sec from microsoft.

Screen shot - http://hosting.flamingeck.com/sp2downloadspeedtest.jpg .

---------- Post added at 01:29 ---------- Previous post was at 01:24 ----------

SUSE 10.1 download, again arround the 700kb/sec download.

http://hosting.flamingeck.com/suse10.1downloadspeedtest.jpg

---------- Post added at 01:36 ---------- Previous post was at 01:29 ----------

Right, well I'm certain that I'm not getting all my bandwidth that I should be. the only test I havent done is to disconnect the tv box from the cable line and remove a splitter which is in the house. I'll connect the macbook via ethernet directly to the cable modem at the first point of entry into the house to ensure it's nothing router / cabling ala dsl style.

Although benchmarks tonight have shown a bad upload for me, they really have been much better, usually the full 60kb/sec upload so I'm not sure if that's a transient blip or what.

I know my speeds are still higher than the average UK household, and webpages since the DNS issues of recent have been fast so it's not major. I just want what I pay for, otherwise I'll be migrating or downgrading.

Thanks guys, you've given me a lot more to be working on. Thanks!

[ecksmen]

Well I've confirmed that the speeds are crap during the day.

I spoke with NTL today, turns out my flat mate hasn't notified them of our change of address (recently moved). They're going to be sending out an engineer to ensure the cabling is at is should etc... I feel like a bit of a fool, I should have double checked he had done all of this. Oh well, atleast when the engineer turns up in the day I'll get him to run through some of the speed tests etc...

[James Henry]

Sadly you pay for 'up to' 10Mbit not a guarantee of it so we'll see what they can come up with. Still gack when you can't get it at any point during the day though.

[ecksmen]

Sadly you pay for 'up to' 10Mbit not a guarantee of it so we'll see what they can come up with. Still gack when you can't get it at any point during the day though.

Yeah, well if I only see 10mb at night time then I;ll be downgrading to 4mb package. I'd rather save cash and have a solid 4mb connection than paying nearly £40 for a 4mb day / 10mb night connection.

Don't get me wrong, 4mb is more than I really need, but still when you pay for something, it's nice to have it.

[Chrysalis]

ntl are shooting themselves in the foot if they dont garantuee a 10meg 34.99 connection to always outperform a 4meg 24.99 connection.

[imran]

ntl are shooting themselves in the foot if they dont garantuee a 10meg 34.99 connection to always outperform a 4meg 24.99 connection.

I have a 10Mbps connection and can download at 10Mbps 24/7 without any performance drop.

I suppose the issue lies within certain areas which are over crowded with students downloading like crazy :p:

[Chrysalis]

and the static routing with some area getting poor peering routes.

but if an area has a lot of students or just a lot of people on 10meg rather then 1 meg, that isnt the customers fault, they would expect ntl to commision more bandwidth for the area.

how do you know ntl commision he same bandwidth in worser areas as in your area? you dont so dont think oh they cant be blamed because its good for you.

hey at least you not claiming every link you click on gives you 10meg throughput.

[imran]

hey at least you not claiming every link you click on gives you 10meg throughput.

No I cannot claim that, however the places where I need to download at the full 10Mbps are giving me that throughput and thats what counts for me.

Hopefully NTL will sort out problems within other areas and all will ne happy.