WAEs optimizing but application data is slow

Unanswered Question

I have two Sepaton VLT archieving servers on either end of the network. The WAN consists of 100 Meg MPLS connection. On either side of the WAN are Gig interfaces connecting the router to the in-line module WAE and then gig to the Sepaton. When we run the archieving process, we see full optimization with good numbers on TFO, DRE and LZ; however, our max WAN throughput is only about 45 Meg and our LAN througput on the WAE-Sepaton interface (LAN) only max'es at 80 Meg. We have default policies applied on both WAEs and the Sepatons are the only devices on either end. We previously did an FTP test between Sepatons and it ran optimized with good end-to-end performace.

It almost seems that the WAEs (running 4.0.7) are pacing/throttling back on the Sepaton traffic.

Any ideas? I'd appreciate it.


I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4 (1 ratings)
Zach Seils Wed, 05/02/2007 - 23:18


What model WAE's are you using? What is the latency of the WAN link? How many concurrent TCP connections are involved in the archiving process?



Zach, Typiclaly there are only two active TCP sessions at once. We have a 7936 at the core and 612 at the edge. It is actually a test topology and our "WAN" consists of two 6500's (trunked together) and we are doing CAR rate limiting on the WAE/Sepaton VLANs to mimic our WAN link; therefore, our latency is minimal.

Zach Seils Sun, 05/06/2007 - 20:38

Have you tried tuning the TFO TCP buffers? For this test, I would recommend that you max them out:

tfo tcp optimized-receive-buffer 8192

tfo tcp optimized-send-buffer 8192

tfo tcp original-receive-buffer 8192

tfo tcp original-send-buffer 8192



This Discussion