Reflex : Conclusions

• Under extreme loads, MQ

  • Virtually eliminates lock contention
  • Reduces lock hold time by 30-50%

• PLS

  • Almost the same as DSS upto 32 threads
  • Better above 32 threads

Notes: