People who are charged with building and operating cloud computing for cable operators need to think practically, strategically and tactically. That, at the highest level, was the consensus of participants in a breakfast panel titled “Fully Realizing Cloud-Based Architectures” at the Cable Show in Los Angeles last week. The panel was presented by BTR and sponsored by ARRIS.
The term cloud is thrown around quite loosely by much of the industry. Engineering folks must be more precise. Such clear definitions still seem a bit elusive, however. For instance, cloud can be as simple as doing the lion’s share of the computing in a different place than where the output of that operation is used. Another definition is the creation of a repository or menu of computing and networking assets that can be assigned fluidly to those who need to use them. Matt Haines, the vice president of cloud services at Time Warner Cable(NYSE: TWC ), points to self-provisioning, full automation and metering as the hallmarks of a cloud infrastructure.
The point is that those definitions - and there are more - at the end of the day say much the same thing. However, they come at the topic from such different perspectives that they seem to be contradictory. Recognition that the definitions are fluid and partly or fully overlapping is an important realization as operators begin to plan their cloud strategies.
Another important realization - and a theme that ran through the presentation - was that operators must pay very close attention to their legacy infrastructure, services and workforce expectations as the cloud gradually emerges. The legacy/cloud dichotomy must be top of mind as applications are developed, services stood up and personnel trained.
Jeff Finkelstein, the executive director of strategic architecture for Cox Communications, noted that the development of standards is a multiyear process - he pointed out that each iteration of DOCSIS has taken about five years to complete - while services in the new environment can turn over in 18 months.
That sounds like a problem, but it isn’t - if it is managed correctly. The longer term and less fluid processes such as DOCSIS and its provisioning can be served in the legacy manner in which they are maintained and managed today. Services that demand speedier time to market and are likely to be tweaked more liberally can be served by a cloud infrastructure. “Then you have the ability to achieve the velocity of new services” that is necessary, Finkelstein said.
Indeed, there's a fairly clear delineation between what can be done in a legacy environment - in hardware, more or less - and in a software-intensive cloud structure. Brian Lanier, the vice president of operations for Comcast’s (NASDAQ:CMCSA) X1 , pointed to more static data elements, such as customer relationship, device ID and user ID, as things that don’t need to be put into the cloud. New services, applications and things that need are more likely to morph quickly are more likely at home in the cloud, he said.
It is easy to underestimate the human element as the scene shifts so radically. Finkelstein said that a group he refers to as “server huggers” - as the name implies, personnel who are more comfortable in a world in which it is possible to actually see the hardware that is performing a specific task - are far more productive if steps are taken to make them comfortable. “If they are used to configuring an edge QAM in a certain way, you can present that same view, and the magic [cloud-oriented provisioning] happens behind the scenes,” he said.
The bottom line is that the industry may be moving to the cloud from several different angles at once. “It can be that you go in baby steps, but there are some paradigm shifts that take time to sink in,” said David Grubb, ARRIS' (NASDAQ:ARRS) CTO of the cloud solutions business unit.
About the Author
