CA Gen TISRVMLS listener optimization

Idea created by TughanC on Jun 20, 2016
    • LanceBrown
    • mgj
    • TughanC
    • staffan.sjogren
    • Britt-Marie



    We would like to enhance the performance of the CA Gen multisocket listener through an optimization of TISRVMLS listener code. In earlier studies with IBM, we have found out that the select() function causes significant increase in CPU time and IBM modified their CSKL and CSKE listener programs without the select() function. After the tests, the CPU time of the program is decreased significantly. As a parallel study, we would like to perform a similar listener upgrade with CA to optimize the performance of the TISRVMLS program if possible. If not, we request guidance on replacing the TISRVMLS listener program with the IBM listener programs to test the response of the CA Gen sockets to such a listener program modification.