As is so often the case, it turned out that my travel plans didn't align, so I tried to catch up afterwards.
It's probably good that I missed it, as I certainly would have been seen as a disruptive influence :)
Make no mistake: I know and like the gang at Wikibon. We've all been around the industry for quite a while.
But even a great bunch of guys can have a bad day once in a while.
Brian Gracely took a stab at elevating the industry discussion around private clouds at the end of last year. It's a quick read here.
TL;DR version: a "true" private cloud is presumably sold and supported by a single vendor: hardware, operating system and hypervisor. I think the goal here is to differentiate against the familiar practice of rolling-your-own from different vendors.
Not to quibble, but I think Brian is attempting to describe a better way of going about obtaining a private cloud: e.g. get it all from one vendor, and saving yourself a bunch of time and money.
I have met more that a few IT shops who've taken their favorite hypervisor, implemented it on their favorite hardware, and inserted their favorite operating system.
They say they are happy with the results, and who are we to impugn on the veracity of their approach?
That being said, the infrastructure biz is clearly moving towards more pre-integration up and down the stack: both on-premises and in the public cloud -- that much we can agree on.
But What About Pragmatic Hybrid Use Cases?
I would encourage Brian and the rest of the gang at Wikibon to look beyond the here-and-now, and extrapolate out just a bit.
My assertion (indeed, one of the big reasons I came to Oracle) is that we won't have the luxury of considering public and private clouds separately for much longer. So many current and future enterprise workloads are candidates for both: working together -- and not separately.
Anyone who's tried to lash public and private together in any meaningful way has come away with the deep realization that it's a lot harder than it looks on the vendors' powerpoint.
Unless the two ends are designed and architected to work together, you're signing up for a living hell of adapters, gateways, dongles, bridges, converters, script code, endless support calls, etc.
Maybe you can invest the time to get something working initially, but of course everything has to be maintained in perpetuity.
And that ain't what cloud is about, is it?
What Really Got Me Going
OK, I was a bit perturbed in Brian's choice of terms, and also thought he might introduce a bit more to the private cloud discussion and its inevitable collision course with incompatible public cloud, but what really got me going is that they published a 2015 market share chart, and -- errr -- completely forgot about Oracle.
Just to be clear: the following Oracle Engineered System products would qualify for inclusion using Brian's definition: Exadata, SuperCluster, Exalogic, Exalytics, Database Appliance, Private Cloud Appliance and Big Data Appliance.
I hope I didn't forget any ...
All share the attribute that hardware, operating system and hypervisor are sold and supported by the same vendor, have a self-service consumption model, etc. I'd go even further to state that Oracle isn't simply assembling other people's products (e.g. VCE), we engineer everything in the stack -- and continue on through the database and applications.
And, oh yes, Oracle provides a compatible public cloud consumption model for most of these private cloud offerings, unlike any other vendor in their roundup.
You might not be thinking public cloud today, but -- trust me -- that train is coming fast. If it isn't part of your thinking today, it should be.
To Their Credit
When one of my co-workers brought the omission to Dave Vellante's attention, he did the right thing: admitted they made a serious error, and attempted to correct their white paper. Still more work to do in the accuracy department, but at least Wikibon has Oracle on their leaderboard.
Unfortunate, really, for both us and them.
Wikibon wasn't the first analyst to make a big mistake like this, and they undoubtedly will not be the last.
Lesson: it never hurts to double-check your work prior to publishing :)
In Other News
Yes, I haven't been updating this blog as much as usual.
However, I've been busy writing for our corporate platform --there's now a half-dozen Forbes posts, and two videos that came out OK :)
Like this post? Why not subscribe via email?