Building out cloud infrastructure on a global basis is never a going to be a small undertaking so the fact that IBM is committing $1.2 billion to deliver cloud computing services should come as not that much of a surprise. But what is interesting are the factors influencing how IBM is going about spending that investment.
According to Dennis Quan, vice president of cloud infrastructure services at IBM, customers are starting to realize that location matters in the cloud. Beyond recent concerns about where data is stored from a compliance and legal perspective, Quan says that businesses with highly distributed global operations want to work with cloud service providers (CSP) that can minimize application latency by making sure application workloads are running as close as possible to the place where they are being consumed.
As a result, IBM is extending the SoftLayer cloud network of data centers and this year will add 15 new data centers worldwide to the existing 13 global data centers IBM gained via the acquisition of SoftLayer last year, and that IBM already had operational before that.
Quan says one of the things that makes the SoftLayer cloud unique is that IBM provides customers with access to over 2,000 application programming interfaces (APIs) that give customers deep visibility and control over virtual and physical servers. IBM is also moving to deploy OpenStack on top of those cloud services as a way of providing support for a standard set of APIs that will work across multiple CSPs.
Ultimately, Quan says the goal is to be able to provide service level agreements (SLAs) in the cloud that are enforceable, versus many existing SLAs from CSPs that are worth little more than the proverbial paper they are written on. Those capabilities, says Quan, account in large part for the reason why IBM has been able to add 2,400 new cloud customers since acquiring SoftLayer.
The CSP game is clearly coming into a new phase where the cost of entry requires an unprecedented amount of distributed computing horsepower. But as interesting as that might be from a computer science perspective, the real proof in the pudding will come when customers start creating large numbers of enterprise-class applications that are expected to drive a $200 billion cloud marketplace by 2020.