• The average number of accesses expected over a given timescale.



  • Any peaks and troughs, when they are expected to occur and how much fluctuation is involved.



  • The response times required for each process.