VeriTest report: Clariion CX3-80 vs NetApp FAS3070

Now now Mario. I never said EMC was working on RAID 6, they may or may not be that's for the company to disclose, what I said was that they were *evaluating* it. Which is inline with their statements on the subject.

Of course, you said "evaluating", Storagezilla, and it's a matter of record now.

My problem is that when "evaluating" things I convince myself that I am "working", so you will probably understand my confusion there.

Thank you for your comments, always welcome.

Well I can confirm the VeriTest findings. We conducted an RFP a few months ago which NetApp and EMC (among others) participated in. Comparing a CX3-40 (RAID 5) and FAS3050 (RAID 6), we found almost the same relative performance in our SQL Server testing as VeriTest reported today.

Why is it NetApp only publish performance benchmarks for systems using less than 15% of total capacity? Why not spec out the entire array and test against that?

Ah yes...

http://www.emc.com/techlib/abstract.jsp?id=1754

Why is it that people believe these "commissioned" tests? Why is it that these "tests" alway use a LOT of spindles, without using much of the capacity? Load a filer up to 30% full on a volume, then run the tests, see what happens then.

I would assume that you have a bad experience with ONTAP FC LUNs for OLTP? Your statement is quite generic to understand what's your trouble with the NetApp filers?

Sometimes, it's also wise to realize that NetApp boxes are true multiplatform controllers, which is what a real environment has. A real world test would then be to setup a box with FC-SAN, NFS and CIFS services with varying loads, and see performances you can monitor from every 'pore' of that box.

I don't know what conditions that EMC whitepaper was written under, but it directly contradicts our experiences with NetApp SAN for Oracle and Exchange.

We're at over 60% utilization (remember we don't waste half our spindles on RAID 10 so we're already ahead of the game) and are seeing perfectly predictable high performance.

Will EMC consider having an independent 3rd party audit their curious findings about a competitor's storage?

Regarding the slanted EMC "whitepaper"
http://www.emc.com/techlib/abstract.jsp?id=1754

There is another viewpoint:
TR-3521: Cooking the Numbers: A Revealing Discussion of EMC's Measurements of NetApp SAN Performance (PDF)
http://www.netapp.com/library/tr/3521.pdf