I am looking to build a disaster recovery strategy for our MOSS farm. The stretched farm configuration looks appealing, but the MS documentation warns not to use it when there is more than 1ms latency between data centers.
Our WAN link has an average 6ms latency between our data centers. Has anybody ever built a stretched farm over a WAN link? Are Microsoft's warnings sincere or over protective?
Reference: http://technet.microsoft.com/en-us/library/cc748824.aspx
They are sincere, and you put yourself into an unsupported environment by doing so (CSS will not be able to help you when you run into weirdness between the farms).
If disaster recovery is your primary concern, your best option would be to configure disaster recovery across SharePoint farms by using SQL Server log shipping. If you are more interested in providing the ability for users connected to different datacenters to use a single (logical) SharePoint implementation, then you would want to look into globally deploying multiple farms.
You are better off going with the sql server log shipping route with a secondary farm. Your probably already familiar with the link but here is the SharePoint availability guide
I'm not a big fan of the multiple farm idea for disaster recovery, I greatly prefer the log shipping option.