No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Clouds, Virtualization, and SOA

The trouble with the market's preoccupation with fads and buzz is that it's easy to sweep really significant issues under the rug. In fact, that may be happening with some of the hottest topics of the day, according to enterprises I've just surveyed. As planners get more involved with new trends like cloud computing, they're learning new things about older trends like virtualization and SOA. What they're learning in part is that they should have known more about these issues all along.Nearly all the 277 enterprises I survey are assessing private cloud computing, but only about 20% have done any meaningful cloud projects. All of them report that they use virtualization somewhere, though, and all of them also report they use SOA. Where the interesting stuff comes out is in comparing responses from the 20% who have cloud projects underway with those who are still just kicking the tires.

Let's start off with a reality check. Private cloud computing is based on the notion of a distributed pool of resources. There are two basic ways to create these pools: connect data centers with virtualization support using tools that manage virtualization across multiple data centers; and use SOA techniques to host application components on any of several data centers. You use the former where applications aren't fully componentized using SOA principles, and the latter where they are. But is even this basic point understood?

80% of enterprises who have cloud projects believed that virtualization was a significant element in their cloud computing strategy, and just short of 70% said that SOA was important. Among enterprises without cloud projects underway, only 43% thought virtualization was important in cloud computing. This, despite the fact that almost every vendor who touts a cloud computing strategy has virtualization as a key element. But with SOA it's much worse: only 19% of enterprises without cloud computing projects think SOA is important.

One company who launched a private cloud project using IBM as a partner told me that they were astonished to learn, in a technical briefing from IBM, that applications that are effectively componentized using SOA are naturally distributable in the cloud without virtualization. It totally changed their perspective on private clouds and virtualization, in fact.

Why are they just finding out? Because none of the material they'd read mentioned any relationship between private clouds and SOA.

There are also major misconceptions in the "no-cloud-project" enterprises about the relationship between cloud computing and other strategic IT goals. Every one of the cloud-adopting enterprises who said SOA was important said that it facilitated software-as-a-service deployment and virtual desktop deployment better than virtualization did. Of the non-cloud enterprises, three-quarters thought virtualization did a better job of facilitating both (it doesn't), and over half saw no relationship between SOA and either SaaS or virtual desktop.

This kind of survey is discouraging because it's hard to see how technologies like virtualization, SOA, cloud computing and even SaaS can succeed when there are such sharp differences between those who are really involved in a new technology and those who are still just thinking about it. Clearly, people are learning a lot when they start a project, because their views are shifting significantly. That raises two questions: how did we get to this sad state, and how will it impact IT project success in 2010?

When I asked cloud-aware enterprises why they didn't have better information before they got their projects started, the number one answer was "vendor material just promoted the technology, it didn't explain it". Other responses were "The articles we found online didn't mention this issue", "we didn't know to search under the right terms", and "there are too many hits to wade through". I did a quick search on some of the issues that cloud-aware enterprises said were hot, and there was in fact material online. The problem is that, as other enterprise planners said, you had to know what to look for.

OK, so we've done a bad job in getting people educated about the real issues (that's hardly surprising given how many times it's happened in the last decade, after all). Won't the user get the truth when they get their projects underway? The answer is clearly "Yes", but the problem is the time. In the cloud-aware group, all but 3 companies reported that their "education" on the real issues added six months or more to the project cycle. All of them said that lack of knowledge up front made it harder for them to get their projects approved. Over half said that they were still operating at a disadvantage because of the early lack of knowledge, and everyone said that their lack of insight would cost them budget money in 2010.

Next step for me was to talk to some vendors, particularly sales people. I've done a lot of sales training so I have decent contacts to reach out to, and the responses were pretty interesting. Five of every six salespeople who promoted SOA, virtualization, or cloud computing said they didn't understand the technologies well. Half said that they just used a sales script to promote their companies' approach. A third thought they were over-promoting their products, services, and capabilities. Every one said that their management's priority was to get a quick deal done and that this desire for hasty closes was contributing to sweeping a lot of issues under the rug.

Not a single one of the 277 enterprises believed that 2010 project money was going to be easy to come by, and yet we're seeing clear indications that the knowledge of key topics that planners need to get projects approved is being doled out reluctantly and ineffectively. That could really bite us this year, and beyond, if we don't figure out how to fix the problem. Smart buyers are confident, and you're more likely to kill a deal by hiding the details than you are forcing a less-than-confident buyer to sign on the dotted line.