There are always fascinating discussions going on in EMC's executive briefing centers.
Sure, there's plenty of traditional bread-and-butter updates: virtualization, security, cloud, backup, storage et. al.
But, more frequently these days, we get customers and partners who put some interesting challenges in front of us, and encourage us to be our very creative best.
And, from those stimulating brainstorming sessions, we frequently come up with repeatable patterns to new opportunities that are customers are encountering.
Such is the case with the concept of an analytics lab.
And I think we'll be seeing more of these before long.
The Situation
The business has now recognized that the next wave of competitive differentiation will be driven by advanced analytics. And it won't be the domain of a few experts, it needs to be a capability that just about every aspect of the business can use if needed.
They've already come to the inevitable conclusion that the traditional IT approach to data warehousing and analytics won't suffice.
For starters, the environment needs to be fluid and flexible, self-service and adaptable to all sorts of data types from many traditional -- and untraditional -- sources. Not the sort of rigid and inflexible environment that's so common in our industry.
The ideal delivery model is more about empowerment and enablement -- giving business people the tools and skills to get their jobs done in entirely new ways -- and less about dumping more work on a narrow band of data specialists.
The ability to scale takes on a whole new dimension -- any single traditional data set is large enough, but the ability to handle the sum of all potential interesting data sets (and do so economically!) adds a few meaningful zeros to the capacity equation.
The business rationale is quite different as well: sure, there are a few initial identified projects and sponsors in hand, but it's more about business enablement and creating a core competency around business analytics as a way of doing business.
What It's Not About
It's also important to be clear what this sort of analytics lab isn't about. For example, it's not about re-platforming and attempting to integrate all the legacy data sources into one mongo behemoth.
Although that might be interesting, that's not the mission. There will always be lots of different data sources in this new world: some old, some new. That's where the value comes from.
It's also not about a classic linear IT closed-end project with well-defined requirements, attributes and ROI. What is learned from the analytics lab will undoubtedly drive all sorts of other interesting IT projects across the enterprise.
And that's a good thing to some.
Enter The Concept Of An Analytics Lab
The idea is rather simple.
Start with a fully-virtualized pre-integrated cloud-like infrastructure (think Vblock, or perhaps the Greenplum DCA). Add a next-generation data analytics database that's inherently flexible and adaptable to the new way of working (that would be the Greenplum database itself here at EMC).
Use self-service consumption portals to make access to information, storage and compute about as simple as booking a conference room (that might be vCloud Director plus Greenplum Chorus offering in EMC's world).
Wrap the offering with three types of professional services. Back-end services to get the environment up and running in a minimum of time -- and with a minimum amount of drama. Add in front-end consultative services to engage with a growing cadre of business users who need to be, well, empowered and enabled.
And, the secret sauce, Greenplum's own data scientists who can show you how to get incredible insight from the data you probably already own :)
All are available from EMC Consulting today, by the way.
Are You Thinking Small? Don't …
I know what some of you are thinking -- what's so hard in lashing up some extra servers and storage, loading up some software and data, and giving users a go at it?
Sure, you could do that … sort of.
But our customer experience paints a very different picture.
First, we're frequently talking about several hundred terabytes (or sometimes a few petabytes) to simply create an interesting enough 'sandbox" for the business. That, of course, requires more than a few commodity servers to power through.
Of course, unless you get the business user engagement model right, all you'll be left with is a bunch of under-utilized tin sitting on the floor.
And, it just might be the case that there's some sensitive data being studied, so a quick sanity check on security and GRC concerns might be in order :)
There's more, but I think you've made the point -- meaningful business impacts usually require meaningful investments.
Bold Discussions
You might be sitting there in disbelief and horror that a responsible business leader would commit serious financial and people resources to what might be perceived as a "field of dreams" project, e.g. build it and they will come.
But these conversations have gone from isolated to somewhat more familiar ever since we acquired Greenplum and started getting serious exposure to people who are deadly serious about analytics.
I guess that's what happens when people start to appreciate what happens when cloud meets big data :)
A big tip-of-the-hat to Bill Schmarzo of EMC Consulting who clued me in on these discussions. Thanks, Bill!
Comments