Servlet CPU Scheduling

Question

Using Weblogic 5.1 SP3 under Solaris 2.7
          We are using the native Performance Pack which allocates 3 Posix readers
          can
          we or should we tune this???
          We have an unusual situation where we want to essentially do an long
          running
          post so we can stream data back to a client much like a file transfer.
          What we are seeing is that depending on the number of Execute Threads we
          see the first
          say 12 Clients - (Using Execute Thread Count 15) start and we dont see
          our other
          say 288 clients until each of our 12 clients are almost finished
          processing.
          It would appear that the cpu allocation timeslice is not switiching
          between concurrent http clients effectively if the servlets are long
          running.
          Is there some tunning which can be done to change this???
          I know this is an unusual situation but given the nature of
          firewalls/proxyservers/load balancing hardware like ArrowPoint,F5 etc we
          need to try to use http vs native
          t3 protocol out to our clients who will be both jfc and web based.
          Any help in this area would be greatly appreciated.
          

Answer 1

naruto8 wrote:
write a program in any language that will simulate cpu scheduling SRTF.
1.the program will ask the user how many process to run
2.then the program will let the user input the cpu burst time for the following processes
3.the user will also input the arrival time for the processes
4.the program will then simulate the cpu scheduling srtf
5.the program should show he gantt chart, the ATT and the AWT
pleas help me with this program...We won't do your homework. What have you done so far? What's the problem?