Monthly Archives: June 2015

From CSP to DSP

For the 10th consecutive year, I attended the annual TM Forum event that brings together communication service providers and vendors to discuss all things operational, systems, and IT. Running a network is difficult enough, but now the market wants network operators to deliver services that go beyond connectivity to each other and the Internet. As the final pieces of the digital network are put in place using IP access networks like LTE, for operators all traffic becomes data. And that should make life easier, but it doesn’t.

For starters, IP networks behave much differently and less predictably than previous versions and require an entirely different set of metrics to be collected and an entirely different type of process to be executed to determine the cause of failures, congestion, and performance problems. Add to that the fact that now every service is an application, not just a physical connection, but software. SOFTWARE!!

A very wise man once described software as being like natural gas. You can’t see it or smell it but it can quickly blow up and kill you. There is no longer just a physical connection to worry about, but applications and data and other amorphous stuff and operators have to deliver and manage all of it.

So now we’re moving from communication service providers (CSPs) that deliver connections; to digital service providers (DSPs) that deliver connected services. IT is no longer just about CRM, billing, and the systems that manage the network (the source of unspeakable animosity); but is now responsible for actual services and revenue. But if CIOs want to take over the reins of a telco, they need to take their game to a whole new level. Business requirements matter, process matters, and software has to be better. Better built, better tested, and better performing. In short, software needs to be carrier-class.

• Reliability – Applications are delivered to millions of customers, millions of times per day and they have to work every time. If business customers are putting their operations at risk, they will want SLAs that are strict with severe penalties. Consumers will just leave.

• Scalability – The Achilles’ heal of every service provider. Delivering a service once is easy, delivering it at variable intervals, billions of times to millions of unique users is incredibly difficult to do well. Virtualization works in trials, we’ll see how it works at scale.

• Availability – Applications can’t go down – ever. This is a 24 x 7 business that doesn’t pause for software upgrades or bug fixes. In addition to backups and redundancy, applications must be sturdy and run as-is without changes or modifications that require substantial regression testing or API adjustments.

• Economy – To make money, applications have to be efficiently architected and developed to run in an optimized manner on a variety of standardized hardware. Code needs to be well designed and elegantly built so that it can be readily maintained and dynamically modified.

Many IT shops are incorporating agile development methods and while that provides some collaborative benefit; it doesn’t eliminate the need for sound system architectures, efficient designs, common data models, and thorough testing. IT is now critical infrastructure and there is now a greater responsibility to customers, operations, and the bottom line. If CSPs are to become DSPs then the IT systems, software, and data must be carrier class.