Yes. It is theoretically feasible. But you have to come bak to the question, what are you gaining (or hoping to gain) by going the route of create threads on the fly?

There are legitimate reasons for doing this, but it is an expensive process that adds complications, some of which can be extremely difficult to resolve. So before I would consider going this route, I would need to have good reasons for doing so, and on the basis of what you told us so far, I do not see that reason?

What are your reasons for thinking this is a good idea?

Also, it would be easier to advise you if you woudl describe more details of the processing you are doing. For example, you say you are querying clients, but what type of clients and what type of queries?

And the second part of your processing is "comparing"...but comparing what? This is a comaprison:

if( $x == $y ) { #do A } else { #do B }
[download]

but it certainly wouldn't be worth starting a new thread in order to perform that operation. However, if you are comparing several megabyts of data, it might make sense.

For better advise, you need to explain more of what you need to do. And for serious consideration of your architectural questions, you need to explain why you are thinking of implementing them that way.