News

posted on 23 April 2008 15:16

In an independent test commissioned by Pillar, its Axiom 500 storage array rebuilt drives up to 71 percent faster than arrays from EMC and NetApp with minimal application performance impact.

The Demartek independent test laboratory compared an EMC CX3-40 and a NetApp FAS3050c with the Axiom on drive re-build times and looked at application performance during the rebuild. The Axiom array re-built its drives faster and showed the least effect on application performance than the other arrays under varying I/O loads.

Each storage system had four disk shelves populated with 500-GB SATA disk drives and configured into RAID-5 disk groups (RAID-4 for NetApp) of six drives, with a few open drives for hot spares. Each one had at least 52 500GB disk drives. Lastly, each system was configured with as many 600-GB LUNs as possible, with a few LUNs of “leftover” sizes, depending on the particular system.

In more detail the findings were:-
- The Pillar Axiom system was up to 30 per cent faster than the EMC CX3-40 and 26 per cent faster than the NetApp FAS3050c in terms of drive rebuild on idle LUNs.
- It was 60 per cent and 71 per cent faster than EMC and NetApp, respectively, on busy LUNs,
- Pillar had higher performance during all tests,
- Pillar’s performance was only slightly impacted during rebuilds; EMC and NetApp saw substantial performance decreases,
- Pillar had the shortest drive rebuild times for all scenarios tested

Pillar's performance was attributed to its use of multiple controllers per disk shelf, with the report stating: "the Pillar system design allows for higher performance for many functions, including the rebuilds that were performed for this report."

Download the report here. 

^{[Paul Roberts, news editor.]}