In botany and genetics, the benefits of hybridization are well understood. By combining the desirable genetic strains of two organisms, we end up with a variant that is better than either of its predecessors.
When we consider the two major genetic strains of clouds (public and private), what new benefits can we expect when we hybridize the two?
And -- more importantly -- what sort of IT alchemy will be required to do so?
The Two Species Of Clouds
Amazon is largely responsible for kicking off serious consideration of public clouds. Amazon was the first to offer a broadly available service that provides an application execution platform that's easily consumable by just about anyone.
Heck, even I've used it.
Historically, though, there have been drawbacks: you play by their rules. One of its core foundations is that the service is highly standardized -- it does what it does -- take it or leave it. To be fair, perhaps a better term for Amazon-like public clouds might be "standardized services" clouds.
That being said, Amazon is thought to be running a nice business serving an interesting amalgamation of IT needs that were previously unmet (or poorly met) by traditional approaches. And -- by far -- its strongest appeal has been its very flexible and granular consumption model.
You pay exactly for what you use -- no more, no less. For some use cases where it's difficult or impossible to accurate forecast demand (and there are presumably wide workload variations), that can be extremely attractive proposition. In a word: convenience.
By comparison, EMC -- along with VMware and Cisco -- were one of the first entities to promote the idea of a private cloud: a cloud-like environment targeted at the needs of traditional enterprise IT workloads.
The idea was simple: in environments of decent size, cloud benefits could be achieved, along with flexibility of consumption options: do it yourself internally, use a compatible service provider, or any combination in between.
The most appealing aspect of private cloud model was, simply put, control. Enterprise IT groups are organizationally responsible for service delivery, security, data integrity, information access, etc. They are generally willing to out-task the work, but not the responsibility.
Many enterprise IT workloads are variable, but vary over a more modest range than the workloads that end up on more public clouds. And, for those who have scrutinzed the numbers, there's a realization that delivering predictable workloads at decent scale -- well, public clouds aren't really a bargain after all.
That being said, there's obvious motion towards the middle from both sides.
Public cloud services appear to be busily investing in control mechanisms needed by enterprise IT organizations. In doing so, not only can public cloud providers attract new classes of workloads, but -- more importantly -- retain existing customers as their needs increase.
And providers of more robust private cloud services are also starting to provide "dialed down" versions of their enterprise-class offerings in an effort to attract workloads that are starting modestly, but have the potential to grow and demand more over time.
Indeed, it won't be long before the lines between "public" and "private" blur even further. At that point, labels aren't useful, and it all becomes about the specific service being provided -- and how it fits into an overall IT strategy.
Moving The Discussion To Internal And External
The next phase of the discussion gets to the core of the matter: what IT tasks are better performed with internal resources and infrastucture, and which ones are ultimately better done externally?
And, if there is some significant benefit to mixing internal and external resources in this new intermixed cloudy world, what steps need to be taken to benefit from it?
The simple answer to the first part (what's ultimately the right mix, and what's the potential) has to be more than the usual "it depends".
Indeed, EMC Consulting has teamed up with a renowned consulting organization to offer a fascinating new service: a cloud strategic and economic diagnostic.
Using a time-bound multi-stream methodology, they can come back to you with a good answer to a simple question: what's the potential economic and strategic opportunity for mixing internal and external cloud-based approaches?
Answering that question involves more than intelligent guessing and hand-waving: you need a clear view of the client's business model for the next five years, an honest assessment of the current state of in-house IT, an inventory of projected workload characteristics, as well as a good handle on current and near-term-future external service provider offerings.
I know, many of you reading this have an inherent disdain for external consulting engagements.
However, for IT leaders in larger organizations who are wrestling with big-picture "what to do?" problem, it's turning out to be not only a highly useful engagement, but ultimately a bargain in the larger scheme of things.
The headline from the introductory deck is attention grabbing in itself: the potential to save 25% - 30% in overall IT expenditures by using a combined approach, *in addition* to delivering better services *and* being more responsive to the business.
If your organization's total IT budget is seven, eight, nine or (gasp!) ten digits -- well, that's a matter of potential interest, isn't it?
Getting There Is Half The Fun?
If you're a regular reader of this blog, you'll remember that I repeatedly state that the journey is turning out to be more about organizational change and process maturity rather than simply substituting one set of technologies for another.
An interesting recent insight comes from Leadership Council For Information Advantage, a rather illustrious group of senior IT executives, organized and sponsored by EMC. Their most recent report makes some rather pragmatic recommendations for key projects and initiatives in support of a transition to a more hybridized cloud model.
The framing here is refreshing -- at least to me -- it's an information-centric perspective, which can be strongly argued is at the crux of internal vs. external cloud discussion.
Specific recommendations include:
- Own the information, even if you own nothing else. Assert your organization's right to own the information, even if you don't own the infrastructure, application or service associated with that information.
- Hope for standards, but prepare to integrate. Lay the strategic groundwork for future data integration activities today.
- Control cloud platform proliferation. Look for shared requirements in standardized business functions, such as finance, HR and CRM, and consolidate services organization-wide on a limited number of platforms.
- Make information "cloud ready." Encrypt information and reorganize data sets so they're accessible and usable across multiple platforms.
- Master solution integration. Shift the IT team's mindset from owning and operating IT systems to becoming master information service integrators.
The full report is worth a read, although it requires a simple registration.
But I Think There's More Here
I tend to think of the problem in terms of "control planes": mechanisms and disciplines that can use either internal or external resources (with associated processes) that permit the increased use of external resources as they become more attractive.
By no means a complete list, I'd start with the following:
- Information Logistics -- in any multi-site cloud-like scenario (whether internal or external) a lot of information is going to have to be moved around through networks. The idea behind information logistics is to have the required information in the right location *before* it's needed. This is turning out to require not only some interesting new technologies, but better tools to characterize information usage patterns. Some initial thoughts on this can be found here.
- Information Protection -- information must always be available when needed by the business -- even if the external cloud or service provider is having a very bad day. The art of creating a service catalog of required protection levels -- and best approaches to ensuring recovery and resliency -- will become more important in this world. Continually validating that information is protected and accessible will become important as well.
- Information Security and GRC -- OK, so this one is obvious. You don't want the wrong people seeing your information, and you want to comply with the confusing tangle of specific regulations and rules regarding how and where information can be used. Once again, continually verifying that all measures are being complied with becomes crucial in this world.
- Application Service Delivery -- as more application experiences will comprise a mix of internal and external services, managing the end-to-end delivery of that application service will become more critical.
- Application Integration -- many businesses create unique IT value through customized applications and business processes. These, in turn, frequently rest on the shoulders of existing application and business logic stacks. The potential for an excruciatingly painful form of PaaS (platform as a service) lock-in has been noted by a few; I'd tend to agree with that perspective. Unique applications represent unique intellectual property ; you should easily be able to take this with you if things change.
I feel somewhat inadequate offering up only partial answers -- there's probably more to the discussion as well. And each of these topics is probably worth an extended dialog in their own right.
Where Does That Leave Us?
If the headline proves to be true -- that by using a combination of internal and external clouds, many IT organizations can achieve a 26%-30% reduction in IT expenditures, *and* deliver better IT services, *and* be more responsive and agile -- well, that's going to be too big a prize to ignore for many.
Once the vision is established, the hard work will be in the journey.
This, in turn, will be less about enabling technology (although there will be plenty) and more about new skills, new roles and new processes that result. In many respects, I'm coming to believe that the journey will largely be paced by the maturity of these newer processes -- and the people who implement them.
Comments