Where’s Bob? For those of you kind enough to notice, I’ve been off-line for several months under a happy barrage of corporate marketing work, along with some travel to Guadalajara, Mexico (interviewing a cloud consultant, left) as part of an ongoing IT outsourcing project.
Much of my work has involved deep dives into cloud computing, including a Computerworld piece outlining cloud security requirements, a white paper for a vendor about traditional ISVs moving their offerings to the SaaS channel, and another upcoming Computerworld story about how to achieve “agility” in the cloud.
Some thoughts from my time in the cloud:
- First, security remains the big bug-a-boo supposedly scaring the biggest enterprises from the cloud, but the customers I talk to say it’s not a deal killer. Yes, you need to know where your data is, yes, you need auditability and, yes, if you work in geographies with strict rules (like the European Union) you need special controls. Cloud providers understand that, and the good ones will help you do the necessary work to stay secure in the cloud.
- Second, the cloud is a mega-trend transforming everything, but it’s not instant magic for every application or service. There’s a reason why some of the biggest names in “cloud” (like Google, Amazon and Salesforce) run their own data centers rather than rely on someone like…well, like themselves. For really high performance, really tight controls over operations, and running your own proprietary hardware or software, sometimes your own data center is still the best.
- The same goes, many say, for predictable workloads running on in-house gear you’ve already amortized. Running such apps in your own data center can deliver not only better performance, but lower-costs. Another note: The per-seat licensing plans used by many SaaS vendors also tend to favor smaller customers.
- On the strategic front, customers are demanding (and vendors are delivering) more “productized” cloud services that can be rolled out in a predictable, consistent way. This often takes the form of templated servers and services (such as those produced by Puppet and Chef) that are pre-tested, pre-integrated and pre-priced. All this makes for yet more choices for customers, who have to decide where a “canned” service is adequate, and when it pays to customize it to reflect their own best practices. It also raises the specter of commoditization for some Web services (how many different Exchange server configurations does the market need, for example?)
- Speaking of which, there are now “appstores” for cloud-based business intelligence platforms, application and database servers and middleware, such as those from RightScale and its ISV partners. Another interesting example of how the Web allows anyone, anywhere (with enough smarts) to sell to a worldwide audience.
- Finally, it was interesting to hear what some enterprise-level customers want, but aren’t getting, from current cloud offerings. These include better, and bundled, support for “enterprise-level” requirements such as load balancers rather than forcing customers to find one themselves. They’re also looking for true, enterprise-level public cloud services rather than the likes of Google and Microsoft that actually (can you believe it?) suffer outages. Then there’s the need for highly verticalized SaaS applications for specific industries such as banking. As the CTO of one major bank told me, does every bank need it own application for handling online payments? While the user interface makes a big difference, it would seem the back end should be standard and, thus, SaaS-able. (New term I just coined.)
In short, there’s still a lot of opportunity for differentiation and innovation as the cloud matures. There’s also huge confusion between public cloud, private cloud, shared services, IaaS, PaaS, etc. Drop a line if you’d like some help clarifying where you fit into all this.
Like this post? Subscribe to my RSS feed and get loads more!