In his recent blog entry Five things Cassatt taught us about cloud computing, James Urquhart emphasizes a core point that I think a lot of folks still do not understand: “cloud computing is an operations model, not a technology.” Understanding Urquhart’s message would help a lot more people get above the hype to understand how cloud computing might become a greater part of their business.
In the telecommunications industry, we have always used a visual depiction of a cloud in our network diagrams to simplify an otherwise complex topic (how we connect two physical locations using our network infrastructure). Whenever I am asked by customers about whether Sprint is going to “get into the cloud space,” I need to step back and explain what that means to us in the context of the big, amorphous cloud that we usually put in center of many of our presentation slides.
Customers typically think of cloud computing as putting their applications and data center into the cloud itself. From our perspective, what is really happening is that the customer is connecting through our Internet backbone either to a service provided by Sprint (hosted at a number of our data centers) or to a service provided by one of Sprint’s partners or customers (who have highly scalable and diverse connectivity between our cloud and their hosting centers).
Sprint’s philosophy about what services should be provided by us versus our partners and customers typically comes down to how ubiquitous the need for an application is amongst the whole of our customer base.
For example, we see applications that filter IP network traffic to mitigate threats from emails or web pages coming into your company as being a key service to provide ourselves as a value added service. This includes services like SprintSecure Message Protection and SprintSecure Web Protection.
In addition, as we look to the future of Unified Communications, we see the need to provide all customers with capabilities that combine the network intelligence embodied in our wireless and wireline voice networks with advanced functionality from presence and location-based services to enable companies to start communicating in a more contextually relevant way.
On the other hand, specialized applications that address the needs of only one corporate function or industry niche would be the kind of application that our customers tell us they want Sprint to enable with connectivity – rather than re-selling that application ourselves. This could include horizontal applications like salesforce automation or vertical applications for a specific industry like banking.
In either case, what is identical here is the concept that a company is using a different IT operations model by purchasing an application as a service. Sprint’s cloud or network-based applications typically allow customers to pay for services on a monthly per-seat license basis – while those from many of our customers take the cloud model even further into virtualization to offer fee structures based percent of server utilization, number of transactions and so on.
The key here is that either approach allows a company to forego the capex of building a data center and the opex of managing servers in a data center for any given application. In this global era where all companies need to scale their IT operations up or down more-quickly in response to fluctuations in supply and demand, cloud computing is here to stay.
For a telecom carrier, the most important aspect of its cloud computing strategy has to be to create a level playing field whereby both carrier-provided applications and partner- or customer-provided applications perform reliably with equal degrees of high availability and low latency. Sprint’s approach to cloud computing is just that – one that provides our customers to the broadest array of options whether those are provided by our company or by one of our customers.