on 28-10-2010 16:31
Why dont VM make tests with their own kit on seperate "test" networks? Failing that, why don't they try setting up alternate ports for their "test" machines and not affect actual people's downloads?
Don't they have a couple of client computers knocking about to do this with? If not I have one or two old machines they can have it if stops people suffering like this.
on 28-10-2010 16:38
on 28-10-2010 17:04
It's a good question, something I have pondered over many a time.
The problem is their network is so complex, (due to the mergers of three companies), it would be impossible to create an accurate test network.
In their defense things that work in test don't necessarily work in the real world. I did extensive tests once on a 155 fibre connection between Amsterdam & Dublin, we had test networks each end. Everything seemed to work okay and what didn't was resolved. Putting it live caused all sorts of issues and had to ve reverted back because of it.
There is indeed equipment and software which generates traffic and could be used for testing but again it would never cover the scope of a real network.
Although I accept there will be problems I think the major flaw is the lack of communication. Take the recent mail migration as an example. There were a lot of problems, (again, I think some due to lack of testing), but at least most people were informed about it.
A few mods have explain about some testing post midnight yes the service status & maintenance pages have said nothing about it. It is this lack of communication that frustrates some of the end customers.
If the service status page listed maintenance and testing schedules it may even lighten the load on the forums. I think this would be welcomed due to the huge backlog I'm hearing about.
on 28-10-2010 20:09
on 28-10-2010 20:33
I disagree, if it is communicated and the tests were done at appropiate times then it wouldn't be an issue.
Virgin could do a number of things. Firstly send a mass mail to everybody or something through the post as a reminder of what to do when things go wrong. It's covered on their webpages but only if you know what you are looking for.
Direct people to the service status page explaining that any service issues or maintenance schedules will be on that page.
Keep the existing service status page but bring back the option when you can see issues in other parts of the country as you could before. At least then if I see my service labelled as working but I can see a lot of surrounding areas as faulty I know my area hasn't had it's service updated or there is a knock on effect happening.
At least then when my system comes to a halt at 02:30am I can identify if it's a known issues or if there is maintenance work being carried out.
Their maintenance work, (in most cases), would have been planned weeks in advance unless ofcourse it was emergency maintnenance.
Surely there must be a way of keeping a record of what mac addresses are on what network segment. The UBR can tell you what IP addresses are connected to it, the IP addresses can be resolved to MAC addresses. The Virgin staff here can tell who you are from your IP address the same as the call center when you ring them up.
Having said the above, why can't there be a system that emails the customers concerned about planned outages?
The answer. If maintenance is being done after 02:00am most people won't know about it. The only people who suffer are the night surfers and people working late. This percentage of people are probably minimal and it's not worth the effort to notify them.
on 01-11-2010 21:48
Well, it's about 9:45 on a Monday evening and I'm getting between 30k/s and 100k/s trying to download anything from news.virginmedia.com
All my other traffic is fine; I can download via HTTP, FTP, SCP or any number of increasingly obscure protocols at "up to 20Mbit" (i.e. I'm not STM'd, not that it would matter if I were), including NNTP servers not running on port 119, so it seems like it's specifically restricting that port.
More surprise hardware testing?
01-11-2010 21:50 - edited 01-11-2010 21:51
More surprise hardware testing?
No it is the published policy
We moderate the total volume of file sharing traffic on our network between 5pm and midnight on weekdays and midday and midnight on weekends. This policy is restricted to Peer to Peer ("P2P") applications and Newsgroups (which are commonly used to distribute large amounts of data)
This policy does not impact any applications other than Peer to Peer and Newsgroups, so things like watching iPlayer, online gaming, making calls via Skype, downloading music tracks from iTunes or streaming them from Spotify and sending an email or normal browsing are unaffected.
It's important to remember that these traffic management policies only apply at peak times when speeds are most likely to be affected by people using more than their fair share. Outside of peak times we do not manage traffic.
on 01-11-2010 21:55
Fair enough; obviously a fairly new one - when did it come into effect?
Don't suppose there's any more detail as to what "We moderate the total volume of file sharing traffic on our network between 5pm and midnight on weekdays and midday and midnight on weekends." actually entails? Is it simply QoS or is there a hard cap involved?
on 01-11-2010 22:07
It's the old policy still with the P2P & New's Groups bit added to it, and yes it's quite recent.
So there is still the usual restrictions during the morning, during the afternoon & during the evening.
In addition to the above restrictions P2P & Usernet will be limited until midnight. It won't be QOS as that would still allow the traffic to passthru at the same bandwidth just as a lower prioirty. It would be a bandwidth control on certain protocols and packet inspection to identify P2P. It was probably setup initially with default settings, hence all the people that play games that uses P2P technology are stuffed at the moment.
I imagine a lot of the above was done to free up some bandwidth that a lot of areas seem to be short of at the moment. It seems like a lot of people have 'high upstream utilisation', the new managment is a desperate attempt to sort it out.
Still count yourself lucky, look at my speed for a 20mb connection all I'm trying to do is regular browsing. :\
on 01-11-2010 23:42
Right-o. This is beyond a joke now! I have NOTHING AGAINST traffic management policies in practice. I work around these to prevent things from happening.
When measures are this DRACONIAN it becomes a farce. I have no problems with my individual bandwidth being limited to 25% at peak times. When the bandwidth that is being supplied to me is one twelth of the total it is wrong. No matter what that traffic is, it should at least be a REASONABLE level.
I have done some testing and I am able to connect to an NNTP source and get the full bandwidth through a paid server. The way I can do this, I am keeping to myself but it costs money to both myself and VM for external bandwidth usage.
Something that should only take 10 minutes to download is currently taking 120 minutes. I would appreciate some kind of response from a VM staff member in an official capacity.