kitz
Posts: 4323
|
 |
« on: November 27, 2006, 08:53:22 pm » |
|
It is on the other portals. As has already been made aware PUG were shown a draft of the new network improvements, on Friday afternoon. There were some questions on this that still havent been answered. I apologise but I wasnt here this weekend and if I hadnt been so distracted by a certain event that occured on Sunday night when I got back and had to waste time on...  I really would have had time to pick up on this much sooner and been way more on the ball today. Anyhow Ive just been doing some serious testing.. at it would seem that the following is now occuring. HTTP traffic being limited to 2Mbps FTP being limited to 60 KB/Sec YouTube etc streaming is being limited. This occurs on numerous gateways. Im well under my SUP/FUP limits and BTw Speedtests confirm I should be getting more. Anyone else seeing the same thing? |
|
|
|
« Last Edit: November 27, 2006, 11:11:54 pm by kitz »
|
|
|
|
|
|
wildmind
Guest
|
|
« Reply #1 on: November 27, 2006, 08:58:11 pm » |
|
HTTP is definitely affected atm.
Windows updates are crawling down on a new PC - lucky if I see 100KB/s rather than the 400KB/s I should get.
|
|
|
|
|
scarymonkey
Posts: 1085
|
|
« Reply #2 on: November 27, 2006, 08:59:34 pm » |
|
I can watch youtube videos without buffering problems. Its been a while since I tried so not sure if the buffering is slower etc.
|
Vince Marsters
|
|
|
|
|
|
Oldjim
Posts: 1016
|
|
« Reply #4 on: November 27, 2006, 09:34:55 pm » |
|
Interesting comparison with previous figures from here which justified the 1%ers Average usage of 0-90% Premier for Oct/Nov 05 was 5.36GB and for all Premier was 10.98GB The latest graphs for the equivalent times give these numbers Average usage of 0-90% Premier for Oct/Nov 05 was 8GB and for all Premier was 15GB Interesting discrepancy isn't it However what is even more interesting are the overall usage numbers The latest graphs show a value of approx 7GB for Oct/Nov 05 but the above chart gives a value of 5.28GB. In the above I am assuming that both sets of data exclude the PAYG users |
|
|
|
|
|
mikeb
|
|
« Reply #5 on: November 27, 2006, 09:52:33 pm » |
|
Yup, see this thread : http://usergroup.plus.net/forum/index.php/topic,3818.0.html and other details on another thread somewhere as well - it's been going on for me for more than a week now. Prior to that everything was just about fine most of the time and had been for absolutely bl**dy ages ... well, barring the occasional hiccup of course, but in principle everything was fine until a week or so ago. |
|
|
|
« Last Edit: November 27, 2006, 09:56:52 pm by mikeb »
|
|
--
WARNING: The e-mail address on my profile is not my usual address, all messages sent via this site have been redirected elsewhere for test purposes. This could result in messages not being received in a timely manner or potentially not being received at all.
|
|
|
|
godsell4
Posts: 397
|
|
« Reply #6 on: November 27, 2006, 09:53:32 pm » |
|
This could be a coincidence, have just enabled VPN to my office at work where we have an uncontended 8Mb SDSL connection ... I get *exactly* a 512kb/s download speed.  SW. |
BBYW1/10GB
|
|
|
kitz
Posts: 4323
|
|
« Reply #7 on: November 27, 2006, 11:11:36 pm » |
|
Seems like this started tonight... or it did for me at least - its certainly the first time Ive seen probs on http.
Ive put a load of tests and facts on the internal forums and given someone a nudge to get Si to look at this tomorrow.
|
|
|
|
|
dgdclynx
Guest
|
|
« Reply #8 on: November 27, 2006, 11:19:42 pm » |
|
Interesting comparison with previous figures from here which justified the 1%ers Average usage of 0-90% Premier for Oct/Nov 05 was 5.36GB and for all Premier was 10.98GB The latest graphs for the equivalent times give these numbers Average usage of 0-90% Premier for Oct/Nov 05 was 8GB and for all Premier was 15GB Interesting discrepancy isn't it However what is even more interesting are the overall usage numbers The latest graphs show a value of approx 7GB for Oct/Nov 05 but the above chart gives a value of 5.28GB. In the above I am assuming that both sets of data exclude the PAYG users Jim. There is something called Pareto's Law which says that in any population 10 to 20% will grab 80 to 90% of the resources. It is an economic law but is found to apply to most situations. So according to it whenever PN get rid of the 5% another 5% will take their place. PN cant blame the current situation on the 5%. They have to accomodate the entire customer base. |
|
|
|
|
scarymonkey
Posts: 1085
|
|
« Reply #9 on: November 28, 2006, 12:15:55 am » |
|
There definately appears to be a max cap on http downloads of just under 2Mb here, whether using 1 connection or multiple ones (I use freedownload manager).
Do I take it that some of those who had been tunneling through the gaming and voip policies are now using http, so PN are again taking action by rate limiting to prevent over abuse?
|
Vince Marsters
|
|
|
|
NB
Usergroup Member
Posts: 2073
|
|
« Reply #10 on: November 28, 2006, 12:39:15 am » |
|
It's just sticking a finger in the dyke, it'll solve nothing long term. There are as many clever people working on ways round things like this as there are people trying to stop it. The problem is the defenders are reactive but those trying to get round it are proactive and using many different methods.
How long will it be before someone comes up with a new approach. It'll be like virii constantly evolving and be impossible to write rules into the ellacoyas to deal with.
|
|
|
|
|
|
Inactive
Posts: 101
|
|
« Reply #11 on: November 28, 2006, 12:45:44 am » |
|
I apologise but I wasnt here this weekend and if I hadnt been so distracted by a certain event that occured on Sunday night when I got back and had to waste time on... I really would have had time to pick up on this much sooner and been way more on the ball today.   |
|
|
|
|
|
mikeb
|
|
« Reply #12 on: November 28, 2006, 01:00:06 am » |
|
It's just sticking a finger in the dyke, it'll solve nothing long term. There are as many clever people working on ways round things like this as there are people trying to stop it. The problem is the defenders are reactive but those trying to get round it are proactive and using many different methods.
You're almost certainly right in which case it's obviously time for Plan B then ... Introduce a realistic and sustainable hard cap above which your access is totally denied unless more money is forthcoming or simply charge all customers by-the-byte to start with. Sorted ... well kinda  D@mn all those P2Pers  |
--
WARNING: The e-mail address on my profile is not my usual address, all messages sent via this site have been redirected elsewhere for test purposes. This could result in messages not being received in a timely manner or potentially not being received at all.
|
|
|
neilarmstrong
Posts: 733
|
|
« Reply #13 on: November 28, 2006, 07:36:13 am » |
|
ALl, for your info a cut and paste of a posting I made in the NDA PUG forums last night:
----
Guys,
As Simon has said we've changed *nothing* on the network today that wasn't changed on Friday so it looks like something has gone seriously wrong.
Only the heaviest of users are supposed to experience traffic shaping at all on Premier accounts.
Product specs supposed to deliver:
Plus - designed for web and email, not designed for downloading
PAYG - do whatever you like
Premier - excellent voip, gaming, excellent browsing, usable downloads on-peak, excellent off-peak
I think we all agree that clearly this isn't being delivered - your feedback has been received loud and clear and we're going to have a full wash up on this first things in the morning and back out whatever the hell has borked this so badly. No way 95% of customers should be affected by something caused by 5%.
|
Neil Armstrong
Products Director
PlusNet
|
|
|
|
jelv1
Posts: 2130
|
|
« Reply #14 on: November 28, 2006, 10:02:54 am » |
|
I'm guessing someone entered a new figure for the amount of capacity to be allocated to the Gold queue and missed a nought off the end - that would explain both the dire browsing speeds and the appallingly slow POP3 last night.
|
jelv
|
|
|
|
