Archived Content

The following content is from an older version of this website, and may not display correctly.

Within the last month I’ve had the opportunity to speak about SDN at two IEEE-related events. The first was the Smart Tech Metro Area Workshop, followed by a Consultants' Network of Silicon Valley (CNSV) gathering. I was very pleasantly surprised at the level of interaction and the visible interest and enthusiasm about SDN. Not surprisingly, there was also a healthy level of skepticism, ultimately manifesting itself in the well-known question: “Is SDN a solution in search of a problem?”

This made me think about the message I had just delivered. To me the value of SDN seems obvious. But as one of the attendees pointed out, we sometimes get so close to the inner-workings and hype of technological advancements that we tend to talk in terms of “what and how” and assume the “why” is clearly understood – and agreed upon – by the consumers of our efforts.

Amidst the flurry of proposals, standards, open source development activities and new products, when someone asks the question “What business needs are being addressed by SDN?”, how should we answer? I offer up the following thoughts and invite the SDxCentral community to join in the discussion.

Timely deployment of new IT services to support growth of and variations in business demands.

The historical architectural, operational and organizational models of networking often prevent IT organizations from “spooling up” services on demand.  Abstracting the underlying resources from higher-level definition of services, combined with policy-driven automation will allow the network to respond to service needs in much the same way server and storage virtualization deliver “on-demand” resources.

Rapid injection of existing and new IT services into operational environments to support new business services.

There are millions of lines of “service functionality” code embedded in network devices, which makes it hard to deploy new or modify existing services. Extracting embedded services from network devices and placing them in an open, agile development environment will enable faster development of enhanced functionality, increase reusability and improve deployment processes and timeliness.

Secure and controlled segregation of customers and business functions.

For cloud and service providers this is non-negotiable. The same is true for enterprises. The digital supply chain is intertwining systems, services and relationships. The rapid pace of mergers and acquisitions is outpacing our ability to safely and securely integrate business functions. Along with existing capabilities, emerging protocols and standards combined with holistic topological views will allow us to more easily “slice” the network into secure chunks that can be allocated to individual customers and functions.

Improved asset utilization

Network virtualization allows us to leverage common networking resources for multiple purposes and customers. While the network model may not be able to leverage commodity hardware to the same degree as servers, the push towards programmable merchant silicon does allow for much more flexibility in determining how network resources are used. Like server virtualization, network virtualization will more effectively enable dynamic workload management, allowing the reallocation of resources and placement of workloads based on business needs. In the end, it’s about less cost for the same resources or more resources for the same cost.

The above descriptions are gross oversimplifications of what will really have to occur behind the scenes. Hopefully they provide business context within which we can describe how SDN will help businesses be successful. But SDN alone is not the only solution. The system will have to come to equilibrium with respect to which standards, protocols, technologies and processes ultimately go into the mix - carefully balanced against investments in existing infrastructure models and solutions. And that will take time, energy and a lot of teeth gnashing. In the meantime when someone asks “so what?” maybe we can give them the elevator pitch that changes the “so what?” into an “ah ha” moment.