Cloud Computing: From Blue Sky Thinking to a Real Solution
Charles Arnold, Managing Director, EquaTerra IT practice (North America)
Stan Lepeak, Managing Director, EquaTerra Research (Global)
One of the biggest buzzwords in IT today is Cloud Computing. Nearly every IT infrastructure team is at least discussing it, if not actually planning for what it will mean to them. Not surprisingly, it has made its way into IT outsourcing conversations as well. But some serious questions remain. How is it going to affect outsourcing in the near and longer terms? How do I make the leap from current state to a future, cloud-based state? Just what exactly does it mean to me? What is a realistic cloud agenda for 2010?
There are two major barriers to using the cloud. The first is that applications or functions running in the cloud have to be architected to be platform agnostic. The cloud relies on purely standardized compute power. But getting every application team to agree on a single platform standard is a near impossibility. That is why even time-tested IT initiatives like virtualization are still major undertakings that unfold over years. It is simply not practical to leap straight into the cloud, expecting in one fell swoop to move giant swaths of infrastructure. Rather it is a migration, and a long-term one at that. It is a journey, not an event. That means that the first steps to the cloud will be baby steps, moving a single application or small group of applications into the cloud almost as a proof of concept.
The second barrier is that the nature of the cloud is that you do not know exactly where your data resides. While the security of the cloud is evolving rapidly, security teams are still uncomfortable with the cloud in general. This is why so many people speak first of “private clouds” where you create a cloud-like computing engine but contain it entirely within your four walls. The advantage is that the security concerns are alleviated, but the large-scale leveragability that is the fundamental allure of the cloud is significantly hampered. Private clouds are really not substantially different than traditional infrastructure outsourcing mixed with virtualization – useful but not revolutionary.
Given these issues and concerns one area that is becoming more viable for mainstream buyers is storage, particularly long-term archive type storage. While live data that is continuously accessed by critical applications requires a more dedicated SAN (storage area network) -type storage array, archival storage tends to more readily moved elsewhere. As cloud-based storage gets cheaper, it becomes a real alternative to off-site tape-based solutions. This is just a first step towards a truly shared, cloud-based infrastructure but can serve as a proof of concept that can build momentum towards a truly cloud-enabled future.
So what do these barriers mean for buyers? First, buyers should craft infrastructure agreements not just with the long-term vision in mind, but with specific provisions for some of these initial steps clearly articulated. Second, they should include robust change control provisions to allow their pricing and service levels to adjust when fundamental shifts in the nature of their infrastructure management occur. Almost all major managed service providers are developing or have already come to market with cloud-based service messages. While some of these offerings remain immature, the direction is clearly set. These changes are coming and will likely become a reality within the five-year term of the new deals being done today.
Leave a Reply