Bored of Megalaunches
It's storage. Ok so it's less boring that it used to be, but it's not the bloody Olympics. Why do they have to put on these stupid massive marketing eventgasms?
Redefine Possible? What a joke.
EMC's been keeping quiet on the whole "how fast can this baby go" front when it comes to the VMAX3 range. But the storage vendor used its megalaunch – with its "Redefine Possible" tagline and futuristic look – to break its silence. So what's under the hood? EMC has extended VMAX to run more and different workloads. Each of “ …
Lol. Yeah.
Separately it's annoying to see the processor tiering between the 3 models. What if someone wants a lot of CPU for very intensive workloads but doesn't need to spread it widely across 2PB in a single box.
Seeing what SSD enabled x86 machines are doing nowadays, it doesn't take a large footprint to start pushing 500k+ of IOPs.
These arrays (my opinion only) are for people who want data protection (SRDF/snaps etc) and the very high uptime the Symm gives - which is still relevant for a lot of places.
The world generally though is moving to lots of small machines doing stuff that's dispersed. It's going to take time, but the arrow is pointed in that direction.
I think what you meant to say was :-
"Overall this launch demonstrates the complex minefield of acquisitions EMC’s storage portfolio has become."
It's no wonder they need Customers to pony up more cash for VIPR in order to to pull this mash up of products together.
Because it works? EMC is the juggernaut of the storage industry, and it has arguably done this with it's clever marketing and robust products. They've continued to eat NetApp and others lunch by claiming innovation as vendors of their own (iSCSI and multiprotocol/unified storage to name a couple) and then dominating the market.
Companies spend millions on marketing because, in conjunction with suitable products, it sells more stuff - even to very clever people that think they are not susceptible to the message.
So in prime fashion, we are left to assume what kind of performance results we could see from this. Sure, bigger processors, more 16Gbit ports, more drives etc, but they still didn't tackle how fast the thing will go. To assume by the components the thing goes really fast would be an error in judgement.
Hi Chris,
Great job on the coverage. One point of clarification on the material above. VCE began shipping in Q4 of last year the purpose built VDI appliance with XtremIO. There is a 1 Brick/Rack and a 2 Brick/Rack system available which support up to 3500 and 7000 desktops respectively. The announcement for later in the year are the larger systems (4 and 6 Brick configurations) for mixed workloads, OLTP, large scale VDI & VIrtual Server environments.
Cheers,
Michael Wing
EMC
Hi Chris,
Thank you for this nice post. (**DISCLOSURE** I work for EMC) I would offer a couple of notes:
On the 100K/200K/400K, you discuss that they support 2/4/8 DAEs. I am sorry we were not more clear in the session. Each engine in each of the arrays can have up to 6 DAEs, each of which can be dense (120 x 2.5" drives) or standard (60 x 3.5" drives), or up to 720 x 2.5" drives per engine. The 100K can have up to 2 engines, the 200K can have up to 4 engines, and the 400K can have up to 8 engines. It seems that we presented this in a way that allowed confusion between the supported engine counts and the supported DAEs.
I would also note that the older (10K/20K/40K) systems support up to 16 x 8Gb or 8x 16Gb FC ports per engine. The new engines support up to 32x 16Gb FC ports per engine, providing 4x the host line rate (rather than double).
Again, thank you for the note of our new systems.
-Vince Westin, Technical Evangelist, EMC (www.linkedin.com/in/vincewestin/)