One of the greatest aspects of technology is its ability to help people leap over its original intention. When 3M engineer Arthur Fry developed the Post-It, all he really wanted was a temporary bookmark for his church hymnal. When Tim Berners-Lee envisioned what ended up being the world wide web, all he really wanted was a way for academic researchers to exchange their work. He never envisioned a wholesale revision of how we get products, information, and services.
So what’s the leap for software-defined networking? To me, this is the fun part of being a technology journalist – looking past the fundamentals and imagining the “what happens next.”
The fundamental benefits of SDN are pretty good on their own: as noted in a 2013 InformationWeek survey, they encompassed faster provision, network flexibility, and lower costs in both capital and operating expenses. But as Martin Sheen would so wonderfully and impatiently bark on The West Wing, “What’s next?”
Arthur Cole posited on Enterprise Networking Planet last month that “with any new data environment … talk of the ‘killer app’ is inevitable.” His nomination: unified communications. Using Microsoft’s Lync as an example, he said, “An SDN can be easily programmed to offer dynamic QoS policies to Lync traffic, with various levels of service for, say, video and audio communications guaranteed through SDN’s built-in visibility. Lync is also network-agnostic and operates on the premise that adequate bandwidth will always be available, one of the hallmarks of SDN.”
But reading that flashed my own light bulb on – that’s why so many carriers are interested in SDN. One of the big bottlenecks facing carriers, mobile service providers, and others when it comes to voice and data communications is bandwidth. Bandwidth is like Manhattan real estate; it’s increasingly dear because they’re not making any more of it. SDN brings a dynamism to networking that has the potential to automatically adjust data flow based on pre-configured conditions. I wouldn’t be surprised if we figure out how to make our limited bandwidth more efficient before too long.
That same kind of efficient usage for limited bandwidth may also apply to the data center. One of the reasons enterprises are moving to the cloud is flexibility when it comes to processing power. But as Ben Rossi noted last month in Information Age, “Once the server, storage, and networking are all virtualized, then a data center becomes a software defined data center (SDDC). In such cases, the workloads can be easily moved between remote data centers and network routing can be changed through software controls.”
That creates for me a couple of images, ranging from the elegance of an orchestra conductor to the quotidian traffic cop. Imagine the ability to configure workloads around the globe with a follow-the-sun philosophy (the conductor), or with an ad hoc capability for accommodating traffic flow and detours based on surprises (the cop). The point is, with that kind of agility enterprises can amortize their investment in hardware by taking advantage of what’s unused when it’s unused.
Consultant Bill Kleyman envisioned a similar view of “intelligent global connections” in a Network Computing commentary earlier this year, but he also took a step further into the realm of content delivery. This is the real “what happens next” in SDN, when we have the ability to match application needs to network capacity on a wholly dynamic level.
As Kleyman notes, “Organizations like Netflix have achieved success building intelligent edge networks that cache rich content closer to the user. Advancements in streaming technologies also allow for rich media to be transmitted live via the cloud. By implementing SDN, network operators will be able to increase network responsiveness, providing a near-flawless experience for the user.”
In a recent documentary on the network of the future put together by the Telecommunications Industry of America, Nuage Networks head of marketing Houman Modarres touched upon this exact issue. “The value chain is being driven by applications. One of the major shifts is that, until recently, the network had been the end product. … Broadband was a product. That’s changed [and] we’re moving forward. Application delivery, connecting people and workgroups to the apps and content they crave is the end game.”
How often have your network administrators had to design networks based on their ability to accommodate the applications that would run on them? And what happens when you take that constriction away. Peer into your own crystal ball and see what SND might deliver to your network agility in the future.