What operating system are you running on ?

unix flavour

What are you expecting to be your performance bottlenecks?

processing speed and Memory utilization

What kind of hardware are you working on?

minimum CPUS available 4 max - 12

Are there significant initialization costs and How much data needs to be passed around

i have to read many queries which are into very big files around 500 in no.The output of queries can also be bulky.Then i need to compare them.Maximizing speed with minimum memory overhead is what i m trying to achieve

Is there any possibility of moving this to a cluster?

not sure right now...

Well i have received some valuable advice from various monks in the thread " Problem in Inter process Communication" though i still cannot decide