DSXchange

We are migrating our "local" infrastructures to a remote site. We moved our DEV and INT instances recently, and the first fallout was an egregious increase in delay between the engine/database and both Designer and Director clients.

We are discussing the issue internally, and have an IBM resource on tap. My question to the group is have you experienced this sort of thing, and do you have any stories to tell about latency and how you addressed it. I'm not asking for detailed technical things -- we've identified ip address on servers as a big part of the problem -- but more along the lines of did you solve it, decide to tolerate it, or something in between.

Unless there's something more new-fangled and better out there, we've used a Citrix server for hosting the clients in the past.

We are implementing Cisco Any Connect, replacing Citrix. The issue is not on that level, I'll clarify.

The big change is geographic distance. During the long transition to relocate the operational side, the firewall/gateway remains "local" to the developer workstation, but the engine and database servers are now (mumble can't reveal exactly) many miles further away. It results in a double handshake between client and server.

Putting the LATE into latency, eh?

In 2006 I worked in Japan with a Citrix connection to Switzerland. I think their Citrix people knew what they were doing, because delays were minimal to non-existent.

Yeah, late bordering on never.

The trace made it rather clear. The physical location of the main gateway, where our guardians of security reside, is the culprit. The distance itself will impose an unavoidable increase in latency, but it should be tolerable. What we have now is basically a 10 to 20-second save and compile taking 2 minutes or more. I'm baselining things, especially export and import. I expect it to be rather ugly.