General discussion


WAN Congestion Caused by SUS Download of XP SP2

By ShitForBrains ·
We have a partially meshed VPN WAN, with 2 large sites connected by T1's and 3 small sites connected by ADSL or, in 1 case, a fractional T1 (384k). The VPN tunnels are all PIX-to-PIX connections. We normally see ping rates averaging 65 to 75ms. Beginning on August 17, these rates would fluctuate from these normal rates to rates as high as 1,200ms! It was quite intermittent, with no recognizable pattern. You can imagine the frustration of remote users trying to hit SQL and Exchange at headquarters!

We have been using SUS for quite a while now, without any problems. Our network engineer has it setup in a hub and spoke model, with the downloads coming into the SUS server at our headquarters, and then the "spoke" sites pull their updates from the hub SUS server through the VPN.

We found the root of the problem by using Object Planet's Network Probe 1.1.1 ( and creating a mirrored port on our Cisco 4507R switch to monitor the traffic to our Cisco PIX515. It didn't take long before we started seeing conversations between the hub SUS server and the spoke SUS servers ALL at once. Conversations were running about 50 to 60MB before we stopped it. What we concluded was that the initial and subsequent SUS updates to the spoke SUS servers failed. As a result, it kept retrying to update after some period of inactivity. Thus, the cycling of the WAN latency rates.

I hope this can be useful to someone out there. There may be other structures or SUS settings we could have adjusted to have avoided it. We haven't taken the opportunity to look into this yet, as we have stopped the SUS services on our hub SUS server while the spoke SUS servers pull their updates directly from MS.

This conversation is currently closed to new comments.

0 total posts (Page 1 of 1)  
| Thread display: Collapse - | Expand +

All Comments

Back to IT Employment Forum
0 total posts (Page 1 of 1)  

Related Discussions

Related Forums