We just deployed WXC for a customer (Hub and spoke topology with 10 remote sites and all in off path mode --- all sites are connected via VSAT - 128kbps links). Everything seems to be working fine now except I am not convinced with the compression and especially acceleration performance. I know from experience that the WXC could do a lot better. I however decided to study the flow diagnostics for different sites and found a few ‘errors’ that I need clarification on. Please note that most of the compression issues come from the exchange (2007) and a few other applications (about 3) while the others are working fine. The following errors are seen after magnifying the diagnostic flow of these applications on the hub and some spokes:
Flow diagnostics summary under General Flow
1) Compression status Pass through – No assembler
2) Compression status Pass through – Filter Applied
These errors occurred randomly in the flow diagnostics especially for exchange e.g. Exchange sometimes displays the ---- ‘Application name Exchange’ and ‘Compression status Compressed – Defined App’ while the network sequence caching of that same flow might then give an error ‘Saw Reliable Packet FALSE’. There are about three other apps that exhibit this random behaviour too.
Also, the TCP Acceleration errors occur for more applications. Some applications do not even show any acceleration in all the branches. In general, the acceleration of this particular deployment is very poor.
I would appreciate your explanation of these errors with any recommendations you can offer.
Also, two sites have been problematic since the deployment. A particular site keeps having issues with their core applications and exchange intermittently (applications stop working). Once the traffic re-direction (WCCP) to the WXC of this branch has been removed, everything starts working well. Please note that this also applies even after the tunnels between that site (spoke) and the HQ (hub) have been disabled. The other problematic site works very fine most times, as a matter of fact, it is the branch that gives the best compression and acceleration statistics when it’s up. The issue with that branch is that the exchange application stops working sometimes. There was a particular situation in which only one person could send mails in the branch while others could pull their mails but could not send. Traffic re-direction to the WXC of that site is usually removed by the customer (network administrator) to ensure it works, before we get in to monitor the situation.
What could be the cause of these problems for the two sites?
Lastly, I noticed before close of business today that I couldn’t access 9 out of the 10 remote WXC’s from the HQ (hub) via web UI. It gives the error ‘internet explorer cannot display the webpage’. This is very strange to me. I pinged the WXCs and default gateways of these sites and they all responded. I also monitored the links and everything seems okay. The compression and acceleration statistics of these sites can still be viewed from the HQ and everything seems normal except I can’t access them for configuration changes.