Basic IOPS

I was asked about IOPS, so that a new server could be sized, this could help. Your results might look something like: Remember that statistics contained in the GV$ views are cumulative and therefore should be sampled. You then subtract your first statistics from the last statistics to give you the throughput. It is now easy to calculate the IOPS for small and large I/O and also MBPS. The following calculations will assist in determining the workload for the current database system. Small Read IOPS =(333474-106883)/(10*60) = 377 IOPS Small Write IOPS =(500211-205903)/(10*60) = 490 IOPS Total Small IOPS =(226591+294308)/(10*60) = 868 IOPS I/O Percentage of Reads to Writes = 44:56 Large Read IOPS =(8010-2791) /(10*60) = 8 IOPS Large Write IOPS =(142981-40298)/(10*60) = 171 IOPS Total Large IOPS =(5219+102683) /(10*60) = 179 IOPS I/O Percentage of Reads to Writes = 4:96 Total MBPS Read =((5586081648-2009381888) /(10*60))/1048576= 5 MBPS Total MBPS Written =((22232604961-4188587008)/(10*60))/1048576= 28 MBPS Total MBPS =((3576699760+18044017953)/(10*60))/1048576= 34 MBPS We could conclude from these figures that this is more of a OLTP as there is more writes occurring.
Posted in Blog Tagged with:

Affiliations

Engagements

  • Oracle OpenWorld 2014
  • Gartner BI Summit 9/3/15-10/3/15
  • UKOUG Business Analytics SIG 12/3/15
  • UKOUG Conference 7/12/15-9/12/15
  • Oracle Ireland Conference 2016 3/3/16-4/3/16
  • UKOUG Systems 2016 12/5/16
  • Oracle OpenWorld 2016
  • POUG 2016 7/10/16
  • DOAG 2016 15/11/16
  • UKOUG 2016 5/12/16

Videos

Posts

  • 2017 (8)
  • 2016 (22)
  • 2015 (22)
  • 2014 (8)
  • 2013 (2)