Feb 17, 2009

Cloud formations

Lots is being said about Cloud Computing these days. How it can cut down overhead, make costs more predictable, and is attractive to many customers. All of this is true, whether it's a sign of the economic times, the overwhelming noise of hosting vendor marketing and sales drives, or back-to-basics common sense of network-based computing that is now more accessible due to lowered barriers to entry. Whatever it is, Cloud Computing, SaaS, hosting, you name it is a popular topic.

Like Web 2.0 - who's demise was touted on TechCrunch this last week - Cloud Computing is another term for doing something that we've been doing a lot of all along. New technologies like Ajax and XML have made easier to provide improved user experiences in the browser. But in reality Cloud Computing - or at least Cloud Computing concepts - isn't so new (anyone remember ASPs*?), just that it's reaching into areas that were previously considered but never taken seriously enough by enterprises. IBM has been doing hosted data and application services for enterprises for many years using IBM equipment in data centers all over the world.

The difference now is that the hoster to hostee relationship is changing. No longer is it strictly one-to-one where the customer feels that their service is unique (even if its on shared servers) but literally the hoster is going big time with a commercial twist when selling services to customers. Again, nothing really new here, as a long time customer of a hosted, shared service provider I've enjoyed a love/hate relationship with the vendor mostly out of necessity since I'm a small business and needed something I could rely on other than my own self-managed solution.

Essentially we're seeing a convergence of the commercial and enterprise hosted services. When the search providers (i.e., Google) or on-line shopping pioneers (i.e., Amazon) are first in line to offer their vast data centers up for hosting productivity applications and data storage then something is going on. Microsoft, IBM, Cisco, and Zimbra legitimize these ventures signifying that this is a market that is likely to attract significant business dollars.

Needless to say, it boils down to costs these days. There are two important factors that I've found customers need to be mindful of when considering Cloud Computing; user and service segmentation. Knowing what services users need - depending on their jobs and the business needs - will help save more money than just going blanket hosted solutions for everyone. Hosted cost models are different than in-house costing. Pricing is usually based on a per user per month/year price for a specific service or service bundle. It's up to the customer to know before buying which hosted services they really can use otherwise they may end up paying more in hidden costs or overpaying for services they don't use. Economies of scale do not work with hosted pricing models like they do when the service is deployed in-house, it's a buyer beware situation. And when we're talking about mission critical applications or platforms of service (such as data centers or communications) a well defined strategy is key to making the best decision.

Which takes me to the second important factor, deciding which services to host and which not to. I recently completed an e-mail strategy analysis for a large financial firm. The assumption going into the study was that hosted e-mail was cheaper. What we found out, through a series of vendor interviews and cost analyses, was that although hosted mailbox costs may be cheaper, when we looked into entire messaging services framework - archiving, DR/HA, compliance, faxing, network security, staffing, IM, fax, and mobile - hosting costs would be more expensive than staying in-house and cleaning up the existing e-mail messes. The study was unique to this company but the story illustrates how important it is to know the technical and business requirements before going into Cloud Computing strategy.

The results may be different in 2 or 3 years when hosting providers get better at manging service options and competition drives down costs, but today, there are many show stoppers that are likely to be costly in the long run for businesses. By looking at the user needs and the required services it was much easier for us to compare costs between in-house and hosted services. We also found out was that some services could go to the cloud sooner rather than later, such as perimeter security or archiving.

I am certain that things will continue to heat up competitively as cost is a huge concern these days and promises of 80- 90% savings has got to turn CIO's heads. I also believe the companies can get some decent savings through economies of scale with in-house services if they clean up their messes and look to "greener" data center designs. Still, Cloud Computing services are likely to continue to grow in popularity especially at the "edge" of the organization where predictable costs are needed. It's not cut and dry and Cloud Computing providers will need to be nimble enough to support the range of requirements needed to support small, medium, and large business alike.


*No, not Active Server Pages but Application Service Providers.

No comments: