Optimizing Atmos for long distance replication?

Hi everybody, I was looking for a little information. I am currently running a pilot with Atmos VE edition across several geographically disperse sites. The sites are connected with high bandwidth links (oc-48 and greater), however there is a high latency on the links (>100 MS).

In the past, I have tuned linux boxes to use different tcp window scaling and different TCP collision avoidance algorithms.

Is there any guidance out there from EMC who can suggest optimizing Atmos to really push high bandwidth-high latency links. Or is it already optimized?

Thanks!

Saunders