hello All,
I need to tune up a Linux server and network which communicated with the Exchange. Due to latency the servers are not performing well.I have a hard time to figure out what is the cause and how to recover form it.
Lot of websites suggest to increase the buffer,setting sysctl parameters but no output.
Can anyone suggest me step by step how i can improve on this?
I will be very grateful.

I'll start with the "physical" side. Where are located the servers ? It's a LAN, WAN, Internet ? Have you checked your cables, switches, routers ? What kind of traffic are you transmitting ?
Once you have discarded all hardware/environmental issues you can start checking your software.

Server Details:Server are Co –located Users Engines run On server ,user Connecting to these server Over Lease line,my observation is Hardware Server ,Linux ,Nic ,Switch, Ultimately Host Server Could be a culprit, Cpu And Memory utilization is Nominal 2gb Out of 32 gb Cpu Utilization 3-4% ,Getting latency on Nic2 Connected to Nse
Hardware Details:
IBM Blade 4
Greek Manager
IBM Xeon Dual 5687 4core Ht enable @3.47Ghz/2X8 48 GB 146gb
Option cm2:98
Switch
Dell 2848 Systimax Cat 6 2 Mtrs Igbps Links <10 Server Per Rack ,I Gig Main Link From Nse
Nic:
Broadcomm 5709 1 Gbps