Cloud is about more than cost savings

I just read an insightful blog post about Cloud and the recent Sidekick outage that got me thinking about a couple things:

1. Cloud is not about just cost savings. I remember when ASP sales guys would get asked by a customer about ROI. And as the ambitious marketer enabling them, I’d quickly assemble a tool kit of datasheets with financial illustrations, business savings solution write-ups, brand-name analyst TCO whitepapers and detailed spreadsheets with 3-year cost savings analysis. The industry has now moved on to slick web and flash-based ROI tools that do everything but the direct withdrawal from the prospect’s bank account.

But what we learned was that starting with ROI was the kiss of death. If the customer only cared about cost savings then they’d never buy. It would be too easy for the tech guy to raise the specter of technical bogeymen that would always trump the potential savings. There needs to be a perceived risk-adjusted net benefit of going out-of-house.

2. Cloud service levels will typically be better than you can deliver yourself.  Gartner traditionally put typical IT availability at 98% to 99.5%. 3-9’s of availability or better was reserved for ‘high availability’ applications. These were typically the most business-critical that got the clustering software, gold-tier storage, etc. If they say the Sidekicks were out for 4 ½ hours or so, and if this is the only outage they experienced all year, then this is still a respectable 99.95% uptime. Not bad for a service that appears to be based on IM-ing about concerts and adolescent hookups. I'm seeing many service providers building out with top-quality hardware, software and expertise that can only be afforded due to their large scale.

Going to the cloud has to be for more that just cost savings: It’s for avoiding the investment for infrastructure you don’t want to buy. It’s for getting services you wouldn't otherwise want to manage. And it’s for realizing SLAs, uptime, security, etc. that are typically tough to deliver yourself.

Same Cloud Different Day

A useful story on Cloud storage shows a real customer addressing the trade-offs of cloud computing, in this case around EMC Atmos storage. Beth Israel Deaconess Medical Center in Boston is balancing cost, performance, functionality, reliability and security. Despite some new aspects there's a familiarity to this "cloud" scenario: "It's a new buzzword for an old technology. It's up to you to define what it means to you and what it means to your business," said storage architect Michael Passe. Includes a helpful paragraph with key players by category.

Another article addresses this same balancing act, but along the local/remote dimension.  More from a user perspective, it illustrates that the decision to go to the cloud can be a logical one, based on a method of dividing your data into types, along the lines of access, application, performance and security.  In other words, high performance, high priority, high availability apps and data living in Oracle databases and the SAN should stay there.  And for things like archival file data, or less-business critical apps, you should look to the cloud to realize savings.

Speaking of Security, a new study from MIT and UC-San Diego has supposedly uncovered a new class of vulnerabilities in public compute clouds. This may become a bump-in-the-road for the vCloud and Compute-on-demand market.  Or maybe a basis of differentiation.

The cloud space is shaping up to be very similar to the ASP market a decade ago. Customers demanded that providers address all the tradeoffs and shortcomings that arose.  And vendors became squeezed between the cost of meeting these requirements and the slow payback of monthly usage fees.  As time went on, there were a handful of survivors, those that got exits with large IT players, and many more who faded away.  Keep this in mind as you make your Cloud vendor selections.

Having said all this, the outlook looks good for cloud when carriers like AT&T are seeing 35% growth in their network traffic in the midst of a global downturn.  We're all still doing more online, customers and users are expecting more web-based systems and thus, deploying more of your data, applications and compute into 'the cloud' seems logical and inevitable.

Welcome to Cloud Itch

According to Merriam Webster, an Itch is "to have a restless desire or hankering for something." This blog will share updates and insights on the "Cloud Computing" space, as well as my own restless desire to provide some clarity within a really foggy market... clouded mostly by disparate vendor messaging along with some emerging technologies and standards.

A very widely cited report from Berkeley defines cloud computing as having 3 key elements: the illusion of limitless resources, the elimination of upfront commitment, and the ability to pay-per-use. I'd offer these along with a couple other common criteria that help define what Cloud means: services connected-to over a network connection, typically the Internet, and Immediacy or fast-time to-solution.

For many managed service vendors, cloud computing is a circa 1999 ASP offering with new messaging. But there are many others who are incorporating completely virtualized infrastructures, offering new classes of web-based services and in some way delivering innovation that moves the industry forward. I'll do my best to take a moment now and then and share info on this latter group as I run into them during my work.

I'm eager to start this conversation - let's make this an active and mutually beneficial forum on Cloud Computing. Thanks! Mike Harding