Which Comes First: The Role of Big Data and IoT in SDN
Without software-defined networks (SDN), the Internet of Things (IoT) stands the chance of having fewer things connected to it. At least, that’s the common view espoused by CSPs, TEMs, and ISVs today.
But how much SDN is needed to eliminate the data bottlenecks that our increasingly connected IoT world will have once homes, appliances in them, cars and virtually anything that can be connected to the Internet come online to the tune of 20 billion devices by 2020, according to research firm Gartner?
The answer: consumers, enterprises, and network operators are going to need to incorporate much more big data capabilities into SDN and NFV platforms than the mere hype that exists around these evolving 5G technologies in 2016. And those SDN platforms are going to have to tap, manage and utilize big data and analytics in ways and with tools that do not exist today.
Why? The IoT implies complexity: the complexity of connected devices and service complexity that needs to scale by network traffic type, time of day, quality of service, geographically, by application, and in so many more dimensions. All of these evolving network conditions that once used to live inside operators’ networks are exploding outside of those networks, connecting more enterprises, devices, cars, homes, people, and things. The effect is compounding.
Because of that explosion of connected devices and the data they generate, network operators will need to enhance their service-delivery infrastructure to be relevant and cost-effective. And that is where SDN – with its ability to deliver infrastructure in the guise of eminently flexible software applications rather than through closed, proprietary hardware-based systems with limited or no ability to scale and flex on-demand comes in.
Telecom equipment vendors like Cisco, who are historically entrenched as telecom operators, are already building application-centric platforms. With this groundwork in place, operators’ ability to leverage SDN and NFV capabilities needed for IoT becomes a much easier SDN-enabled future to realize. Operators can use SDN-enabled networks to shape network traffic based on time of day or based on an application’s need. For example, data from a utility company’s streetlight-mounted sensors spewing air-quality data will spike during a wildfire event as would wind data during a tornado or hurricane.
The usual challenges face telecom equipment vendors and operators in the form of security, common data APIs, standards-based development, the cost to re-engineer existing platforms or to overlay networks with SDN/NFV technology.
How will that big data be used? In network monitoring tools and network operations management. In conjunction with cloud apps. In real-time and offline. But intrinsically, all of that data will drive the applications that are at the heart of software-defined networks. So the answer to my headline question is: “In the 5G network of tomorrow, big data, SDN, and virtualization (NFV) are all at parity in terms of their importance.” They all come first!
To date, the telecom industry has been focused on connecting the network elements, not in building the SDN applications that will drive the use of end-user applications. The focus in SDN/NFV circles has been on how the new platforms will be built. Harvesting and harnessing big data from these evolving services in tomorrow’s 5G-enabled networks will let vendors focus on why we are building them in the first place.