Skip to end of metadata
Go to start of metadata

Meeting held at UPC after the first TUC meeting.

Present:

  • Miquel, Norbert, David, Larri (UPC); Peter (VUA); Orri, Hugh, Antti (OGL); Barry (Ontotext), Irini (FORTH)

Summary:

  • start with the Social Intelligent Benchmark (SIB) data generator (S3G2)
    • investigate also the Myriad data generator generator from TU Berlin
  • assess the queries in the SIB W3C page
    • Peter/Orri will take a look at the queries to make sure they are affected by data correlations
    • we will investigate a mechanism to obtain parameter bindings with predictable selectivites (and write a paper about that!)
    • ask all task force members to look at these queries (including Santiago and Carlos)
  • there will be two benchmarks:
    • transactional
      • updates (add user, friend, discussion, discussion post, photo, photo tag, add/delete friend.
      • lookups and simple analytics
      • see SIB W3C page
    • analytical
      • BI question, see SIB W3C page
      • global graph analytics - to be defined
      • data load and enrichtment??? not discussed yet
  • outcome on discussion on the metrics
    • the transactional benchmark has a Throughput score
      • measure  the max amount of queries/sec, given a concurrency level of choice
        • the read/write ration in the transactions should be fixed
        • or maybe limited amount of combinations??
      • does it also have a Power Score?
        • fixed query latency may make this meaningless
        • maybe power score as being the read-only score
    • the analytical benchmark has both a Power and Throughput Score
      • benchmark bulk load time may be incorporated into the scores

 

 

  • No labels