Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have a client in a managed office, and part of that service is the provision of internet, with static IPs for all desktops provided by the building, (I know it's a recipe to have all your data stolen, and you not even know!). They insist on saying that traffic is not managed, but we have different bandwidth available on different sockets.

Two desktops have blisteringly fast internet, and the rest have dodgy crappy internet access.

Is there anyway that I can prove that the traffic is being shaped, so that I can convince the powers that be to bring networking inhouse?

share|improve this question
add comment

3 Answers

up vote 0 down vote accepted

It's not possible to tell if traffic shaping is taking place, but why would anyone bother doing that on part of a LAN?

I agree with @massimo's first sentence, it is more likely to be a cabling issue. The trouble is that moving a laptop about might not prove anything. What might be worthwhile is setting the NIC on the PC (laptop or otherwise) to fixed speed & duplex (if it is currently auto/auto) as that might match with the switch port setting whereas auto/auto will cause a duplex mismatch.

share|improve this answer
    
"why would anyone bother doing that on part of a LAN?" "managed office" arrangements rent individual office spaces and may charge additional charges for things like network access. –  user48838 Jun 14 '11 at 14:20
    
@user48838 If it is part of the managed office deal then I can see that they would have limited Internet speed for all of their office but the only reasons I can think of for part of their office having different speeds are either incompetence or switch port issues (which would also actually be down to incompetence, assuming it is on the network side, not the PC side). –  blankabout Jun 14 '11 at 14:44
    
They could have things mis-provisioned if it is a new move-in, where they may offer "buy-ups" for more speed across their client base. –  user48838 Jun 14 '11 at 14:55
    
Someone would do it when they have around 100M connection to the internet, and about 3K endpoints inside the same building :) –  Mister IT Guru Jun 14 '11 at 15:09
add comment

This could very well be a switch/cabling issue, or a configuration problem on the desktop themselves.

Anyway, a quick and easy test: take a laptop, plug it in socket A, download something big from a well-known high-bandwidth site (like Microsoft download center), measure time, move laptop to socket B, repeat, compare speeds.

Also, while you're at it, you can compare LAN speed (f.e. copying something to/from a file server): this will help you identify if the speed issue is related to LAN traffic also (which would definitely point in the direction of a switch/cabling issue), or only to Internet traffic (which would strongly support the traffic shaping theory).

share|improve this answer
    
Keep in mind that Massimo is right...this could be a switch or cabling issue still. All you're doing is eliminating if it's the PC that's the bottleneck or not, not whether there's an issue with the port or drop. –  Bart Silverstrim Jun 14 '11 at 12:40
    
Also by shaped, do you mean that all traffic is limited just by port, or shaped by protocol? –  Bart Silverstrim Jun 14 '11 at 12:41
    
Did all the above already - I got speed results with too much deviation on the links that are crappy. Please bear in mind that the client is in a managed building, and I have no control or access to the network infrastructure. I was hoping to be able to send a packet to a remote server, and tell if it had been altered by a router shaping traffic –  Mister IT Guru Jun 14 '11 at 15:09
1  
No, routers never "alter" packets, they only prioritize them; there is no way to tell if a router made a packet wait a little more because the link was busy/faulty or because it deliberately queued it for a while. –  Massimo Jun 14 '11 at 16:27
    
They may modify the TOS/DSCP bits in the IP header and they will ALWAYS modify the TTL. They shouldn't be touching the payload, though (except, possibly, splitting it if teh packet needs to be fragmented and the DF bit isn't set). –  Vatine Jun 15 '11 at 10:16
add comment

Sometimes Ethernet flow control may be a contributing factor. Please have a look at this article and see if it applies to your situation.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.