Forum Discussion

2 Replies

  • Classis case of 'it depends'. Are the rules complex? Are you storing a large number of variables and data in memory? Are they well written? And so on.

     

    Generally speaking, as long as they are well written and commands are executed in appropriate events then the overhead should be minimal and not something you have to think about. iRules are a core TMOS/TMM component and designed for high performance.

     

  • Hamish's avatar
    Hamish
    Icon for Cirrocumulus rankCirrocumulus

    You can find out the performance impact by adding the

    timing on
    

    statement at the beginning of the iRule. That'll generate timing stats for each event.

    If you want finer grained you can use the [clock ] command to get the timings at high resolution and work out the performance by logging the clock at various points (Via HSL. Although of course adding these commands will affect the timings against the iRule without them). 🙂

    There have been a few discussions here on devcentral over the last few years about what's more efficient. And IIRC there were some real examples as well.

    H