Hi list, We have an environment that is using 7604 SP 5 and currently no load balancer in between mid tiers and app servers and are evaluating introducing one.
Using windows sever 2003 and F5 load balancer. There are a few questions. Currently, app servers are serving stand alone requests from web servers and are seeing ~400 users on each server during peak load, however are configured via hardware to handle ~1000 users concurrently. Usage of incident managament is the highest and no other modules are considerably used. Assuming there are 6 app severs, current peak usage number of users would be around 2400. If we introduce a load balancer in between MT and app servers and set "enable lifespan" on the MT config page, per documentation by BMC, any user request from any MT can hit any app server. Questions below: 1. Does that mean at peak time, all app servers could see ~2400 users connected to it? 2. If #1 above is true, this could potentially cause performance issues as each app server is configured to handle only 1000 users 3. How does enable lifespan work in the background? Could not find any documentation on this 4. If each request is load balanced, load on load balancer tends to grow exponentially, is this a valid assumption? 5. Number of hops between mid tier and app would increase and could also cause slowness but this could be a tradeoff considering we will achieve HA. Is this a valid assumption? Regards, Karthik _______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org "Where the Answers Are, and have been for 20 years"
