Scott Drummonds on Virtualization

First Ever TPC Result on VMware, New Records


A new era has dawned on VMware virtualization: the Transaction Processing Performance Council (TPC) has posted an audited result on a virtual platform.  That platform, VMware vSphere 4, ran ParAccel’s Analytic Database (PADB) to set new records for the TPC-H benchmark using a 1,000 GB database.  You can read more about ParAccel’s work in their recent press release.

Here are a few interesting stats about the published TPC-H score:

  • The score of 1,316,882 QphH is the highest ever reported throughput for the 1,000 GB results.  It is 13% higher than the next greatest performance score.
  • The $0.70 price per QphH is the lowest price/performance ratio ever submitted on the 1,000 GB results.  It is 40% less than the next closest price/performance score.
  • ParAccels results are a mere 1/8 the price/performance of the second best performing result.
  • The vSphere 4 configuration used 40 HP servers with Xeon 5560 processors running 80 4-way virtual machines.

ParAccel’s amazing work represents a large number of firsts for VMware: our first audited TPC result, our first proof point with big data applications, our first published partnered work in the rapidly growing data warehouse space.  One more amazing thing about this accomplishment: ParAccel and VMware set this record on commodity x86 hardware, which means the infrastructure can be shared with other enterprise applications.

Congrats to ParAccel.  They have shown a product that can set records not in spite of virtualization, but because of it.

4 Responses

Interesting, it was 40 servers and HT was disabled. My biggest problem is that they used Standard, I can’t see realistically using Standard for such a mission critical application. Even if they had included Enterprise it would have been highly competitive on $/TPC basis.

    • Well, it was a benchmark submission. Most benchmark results are produced without following production best practices. I think the choice of Standard was reasonable. But you are right that ParAccel was so far ahead that going all the way to Ent-plus would not have stopped them from setting a record.

      However, as a fan of performance records, I wish they had done it on a free version of ESXi!


  • I’d be curious to see the same test on a standard database platform like oracle, db2, sybase or SQL Server.
    PADB isn’t representative of any real dB workload

    • If your goal is to compare PADB to other DB platforms, I think you will see that PADB is much faster. ParAccel uses is a columnar database that is much faster for data warehouse workloads like TPC-H.