As interesting as funny puppets are, and if someone actually took the time to read my Blog entries as a cumulative collection of posts and not just what is on the front page, they will see I’m usually un-biased. In fact I specfiically call out NetApp as one of the lead NAS vendors in several areas, including having the SPEC-SFS results that showed the OnTap GX cluster as by far the IOPs leader for clustered file systems. That artcile, by the way, is far from “second rate blog” material, thanks for making it personal, not all of us are paid to pick on others products, some of us actually have the experience in technological reality, developing, delivering, and seeing products within customer accounts.
Now let’s take a look at Alex’s blog response. It is funny that I took a statement about NetApp picking on 3Par’s solution, and it turned into a “They must be very worried indeed”, the THEY in this being Dell of course. This was never about PS-series storage or Dell for that matter. So, yes of course I do what anyone with literary license does, I took a graph from a public NetApp document and posted it showing clearly that the WAFL has an issue when utilization is increased in a volume. (hey Alex, yes this was originally about utilization claims in a performance benchmark). So from that same document Alex pulled the following:
With the “claim” that concurrence fixed the problem. Well what isn’t shown, and was neatly left off this graph was what the “performance” looked like “with concurrence” at utilizations prior to 50% (notice the green line for those following). So, if NetApp would like to publish what this graph looked like before it was edited (or maybe those tests were never run)…
Leaving me to defend Dell’s, the leading iSCSI storage vendor (by market share according to Gartner), PS-series storage arrays and documented benchmarking…….which I am un-willing to do, as this isn’t about Dell, or Equallogic. This was my personal defense of Marc Farley, who I’ve known for awhile and felt his post was being picked on.
So lastly, as I’m kind of on a graph roll here, Alex posted the following:
Which was so blury you can’t even make out what is bei’ng represented. However, what this is, is showing that “over-time” IO has remanded constant in a “high utilization” environment with Snapshots enabled. So what is missing here, is, not a change in utilization, because we’ve already proven that with NetApp you loose performance as you increase utilization. It also shows that NetApp has LITTLE TO NO OEVERHEAD with Snapshots!!!!!!
Hold on your your hats here….YES IT IS TRUE NETAPP HAS SNAPSHOT TECHNOLOGY THAT IS TRUELY LOW/NO OVERHEAD! I said it, the pain from the Dell implant almost kept me from typing those very words. Why is this true? Because, NetApp DOES NOT USE a COPY-ON-WRITE based snapshot technology. This is a very good thing, since copy-on-write based snapshots incur a pretty hefty performance penalty.