I had to hurry up to write this, because EMA’s data collection has been getting faster and faster. Just a few days after May 21, I may have the first wave of solid data to know if I’m right or wrong. (My next column, I promise).
With its just-launched “Ecosystem Cloud” research, EMA is trying to understand how different sized IT organizations are managing applications and other services with options for internal cloud, external cloud, hybrid (internal/external), and traditional infrastructure. (I know that cloud vendors like to call these “workloads,” but I keep reminding them that workloads are only part of the story.)
Getting meaningful statistics on how this is falling out by company size, vertical, etc., should be both interesting and useful. After all, aren’t you curious what your peers are doing here? And by implication, why? Not that it’s necessarily a footrace. We’ll also look service-delivery success rates — and be able to correlate those to levels and types of cloud adoption.
The first bet I’m making is that it’s all over the board. If any meaningful percentage of our respondents comes in as all internal, all external, or even all cloud, frankly I’ll be surprised. I know, that’s a pretty safe bet — especially if you’re looking at larger enterprises. So I’ll place a few more bets here:
- Slightly more than 60% will view cloud as “already adopted and an essential part of our business.” A little over a year ago, the number was 48%.
- Messaging, CRM and desktop productivity will continue to lead in “cloud” adoption overall.
- There will be a growing move to pick up some applications in all internal cloud environments, but very few will be depending yet on all internal cloud for critical business application services, whether for running their businesses (e.g. SAP), or critical, externally facing business applications — even among those businesses where Internet retail is primary and dominant.
- Internal cloud for application development by contrast will already begin to look “mature” in many environments.
- External SaaS for applications is nothing new, and the growth here will admittedly be hard to assess in part for this reason. SalesForce was around, of course, long before Cloud meant something other than Wide Area Networks. Exchange is another popular SaaS offering. So I expect nearly 100% of respondents to have some deployment of external cloud (new or not). But I also expect the number to remain low (under 30%) when asked if cloud adoption is “primarily external”. By way of reference, a little over a year ago (fifteen months to be precise) the number was 24% for “primarily external”.
- On a separate point, the degree to which many IT organizations see their businesses depending on some mix of internal and external cloud services I also expect to see modest growth. For instance, if 45% claimed a balanced mix of internal/external cloud 15 months ago in “Optimizing Cloud for Service Delivery”, that number should now exceed 50%.
- I also believe that we will see a growing number of IT organizations looking to more creatively optimize resources for individual application systems across a mix of internal and external cloud. This is a new and, I believe, a growing trend (versus Exchange all on external SaaS and Development all on internal PaaS). After all, cloud is supposed to be all about flexibility, including flexibility in close to real-time to accommodate optimization for cost and business impact. This is, in my view, the endgame of cloud (although I don’t view cloud as an endgame in itself). As this will be brand new data, I won’t make very specific predictions here, except to expect the beginning of meaningful adoption patterns where, say, external application Web hosting is combined with internal IaaS-based storage or DB access for a meaningful mix of application services.
We’re going to be looking at other trends as well. For instance we’ll be getting more data on ITIL adoption priorities (see my recent Q&A with EMA Director Torsten Volk – EMA Talks about ITIL and Cloud).
Torsten and I are still wrestling with certain patterns that seem contradictory. For instance, in multiple research efforts (including recent research) my data shows CIOs have a meaningful preference for ITIL over the general IT organization. In Torsten’s, to our mutual surprise, it was just the opposite.
I’m going to make a bet here that once again, my data will show CIO’s leading in ITIL, not because I’m asking it (I remain anonymous) but because of context. Torsten’s work was a deep drill down focusing on Internal Cloud. This research looks at service management in the context of cloud and mixed cloud/non-cloud environments. While a perfectly complete internal cloud deployment may diminish at least some ITIL requirements for process definitions as they become automated, the messy notion of services across a mosaic of internal, external and traditional environments may argue for just the opposite. We’ll see what the data says here very soon.
Another area we’ll be examining is the role of service modeling and CMDBs in cloud. So far the data has been ambiguous. If I had to sum it up, recent research shows that while service modeling is gradually rising in importance in assimilating cloud, CMDB adoption specifically in support of cloud remains a low priority.
On the other hand, those IT organizations with effective CMDB deployments are also significantly more likely to be successful in managing services across cloud than those without.
Here my hope is that we see more about why these trends are true. Is service modeling becoming the new spine for IT as I believe it must? Or are we faced with continued reluctance to deal with the complexities of modeling given the temptations for speed and (over) simplification that still surrounds a lot of cloud hype. I’m going to bet modestly for the growing importance of modeling. After all, unless you know better, why not bet for what you hope for?
In the end, this research is something of a moratorium on to what degree cloud is fundamentally changing the game for service management. Do IT organizations see cloud as a separate universe altogether, or do they view it as advancing many of the requirements for managing services across increasingly dynamic infrastructures and ecosystem interdependencies that already had deeper roots in the past? Once again I’ll place a bet, and this time I’ll just let you guess where I personally stand.
Dennis Drogseth's follow-up blog: Ecosystem Cloud: The Results Are In!