How does TimesTen support geographic redundancy
Hi:
We are using TimesTen 7.0.3 in our product. We deploy an active
standby pair in the system. There are several(1-4) application servers
deployed. The system is a OLTP system. Application server execute Db
trasactions to the Db server. So the network traffic between application
server and DB server and the netwoek traffic between DB server are heavy.
Now we are considering the possibility that the product support
geographic redundancy. That is the system are deployed in two site,
if one site is destroyed, the part of system in the other site can
still work. The two DB server shall be deployed in the two site.
ANd I think the application servers shall be deplyed in both sites
also.
Could any person can suggest if there is success story of timesten
on the geographic redundancy?
I am wandering what are the issues that needed be considerred. The
network bandwith is apparently an issue. If Gb switch can be used,
what kinf of network equipment shall be deplyed to reach certain
network bandwith. Another issue is the network latency. I think
that will not a big problem, as long as the latency is controled
in certain level(say, less that 1 second). Maybe the throughput
is the only trouble.
This task is new to us. I wish I can have some discussion here and
get some ideas.