If you're a technophile, stop reading this post right now :-) This is a nit, but since I posted earlier on this little saga, and some folks LOVE getting a peek into the insider sillyness , here goes....
To recap - EMC was using a slide that all of a sudden we we asked to stop using by IDC. I talk about that here. So, we got the formal approval from IDC what we can use, and I've passed the edict to our field to update. These are all public docs, and ones EMC folks and EMC partners present to thousands of customers every day, but still... want to see the before and after? Based on what IDC told us we can and can't do - it's a huge difference. It also strongly suggests where the complaint came from. I have to get this off my chest, and then back to technical action (I'm working on a DOOZY of a post)..... Read on.
Ok - here's the old one.
Here's the COMPLETELY different new one:
Umm... Ok - one is a pie chart, and one is a bar chart?
Ok, one change is very material which I pointed out earlier: it's inaccurate (and clearly misuses the data) in the old one to call it "Market Share" - it was the IDC study sampling US-based customers. As soon as IDC told us that (hey, to me "a sample of 311 customers" is a "market share study"), we changed it everywhere to the title they suggested, and cited the source literally.
What was interesting was what they wouldn't let us change and what they were OK with:
- No year to year comparison. Ok that makes sense to me - I mean, if it's not a market share study, but a sample, the only real accurate year over year measure would be to survey the EXACT same customers - which wasn't done. OK, fair enough.
- We had to use a bar, not a pie chart. Ok, I get that too. Pie implies percentages of a whole. bar charts are numerical responses to the question. That's totally fair.
- They asked us to omit our "Dell included Dell-Branded EMC" comment. Ok, I'm good with that too. Even though in 2007 the majority of Dell storage sold as shared storage in general was heavily CLARiiON biased (there was some MD3000i - the LSI Logic box they rebrand, and loads of PowerVault, my understanding is most VMware configs were CLARiiON) - it's a statement that wasn't explicitly asked in the survey (i.e. "is that Dell, or Dell/EMC?"). It will be interesting in 2008 to see the Dell/EqualLogic and the Dell/EMC split - I hope the IDC folks ask (but would be surprised - it's not really the most exciting part of the survey unless you sell arrays.
- The explicitly forbade us when we asked to call out some of the other folks in "other". HUH? This one was odd. We wanted to pull out the 3% HDS and 4% NetApp results from the survey, as they are strong competitors, and not lump them in with "Other". This request clearly highlights who was complaining to IDC in the first place. Don't take this as arrogance. I respect all our competitors - we're all trying to innovate and serve our customers. In particular, I'll reiterate two things I've said in the past:
- about NetApp: "I appreciate having as formidable a competitor as Network Appliance. With other competitors, we compete with them, and it’s a slow waltz, but with you folks, it’s a tango. It makes us better, win or lose with any individual customer, and in the end, it’s good for the industry and the customers." (you can see that publicly here: http://blogs.netapp.com/dave/2007/04/emcs_celerra_si.html)
- about using this sort of data: "Now - there's two things that if your EMC salesperson (they are wont to hyperbolic chest-beating) tells you - you should correct him fast: 1) "EMC dominates worldwide storage attach for VMware" (study was US only) ; 2) "EMC has dominant market share for VMware" (this isn't market share, it's a sample - and more importantly - who gives a rat's a$$ - in the end share is useful, but what really matters is who has the best solution for you, at the best value - the EMC team better focus on that part) BTW - I stated both of these earlier in What's in a study-
But likewise - if another vendor tries to tell you "If you don't use storage virtualization with VMware, you're missing out" - a common HDS refrain (I'm SO on the same page as Alessandro here) or "NFS is the only right choice for VMware storage, and everyone is migrating en masse" - a common NetApp refrain.
Alessandro says on that same post that "what's the right storage infrastructure for VMware" is his #2 question, so this is an important question for people.
My answer is that it's not the same for every customer (and that EMC is a lot, lot more than a storage company). On the question of protocols, I think NFS is fantastic, and so is iSCSI, FC, FCoE - each has their place (and each has a VMware super-power - something they do better than the other protocols in the VMware context, and each has things they CAN'T do in the VMware context) - for most customers, I think one of the three block protocols, PLUS NFS used in concert are the right answer. Re storage virtualization - it sure has it's place, but the value proposition is totally unrelated to VMware (you can see my comments on Alessandro's post).
I guess 'nuff said (at least from me - comments always welcome).
What I love was that IDC had no issue with the quote (it was literally from the study): "For virtual servers, networked storage solutions are more heavily weighted toward EMC storage. In previous years storage attached to virtual servers was highly captive to the server hardware purchase"
OK, back to business. In the end, every customer needs to decide what's right for them, and each vendor (us included) need to make the case of why they are the right choice for that customer. It's been a lot of effort over a small change, but remind your EMC and EMC partners if you see the pie chart they are misrepresenting the case, and if other vendors are complaining - you can point them to this open, honest dialog :-)