How Do SDN and Wireless LANs Fit Together?
SDN (Software Defined Networking) has been discussed for years. In this series, Craig Mathias looks at how SDN fits into the Wi-Fi landscape.
SDN is well on its way to becoming the dominant architectural and management philosophy for the implementation and operation of organizational networks. Modern networking is all about policy, efficiency, and making sure that the network is optimized for the mission of the organization, including cost-effectiveness and security.
With wireless LANs (WLANs) now an essential element in organizational networks everywhere, how do (or, since it’s early in the history of SDN, should) SDN and WLANs fit or at least work together?
This is a topic of more than passing interest. Wireless LANs are now really the only access that matters. Many, if not most mobile devices, have no way to plug into a wired network, and most of their value would be lost if they were not Wi-Fi capable. Given that Wi-Fi has been able to keep pace with demand for availability, throughput, manageability, and, most importantly, capacity, it stands to reason that Wi-Fi’s value further increases if we can integrate it into SDN-based architectures and implementations.
For example, wireless LANs for years have been dealing with (and, really, leading in innovation) traffic-flow issues, most notably traffic prioritization. This is important since a given radio channel is accessed serially, so reliability and efficiency are paramount. Extending the policy definition, implementation, and management capabilities of SDN into wireless LANs therefore makes a lot of sense.
But this is where the questions get particularly interesting. A controller-based WLAN architecture need only extend the SDN implementation as far as the controller; SDN firmware and functionality are unaffected.
It can be argued, however, that implementing SDN capabilities directly in APs makes much more sense – end-to-end is the overarching requirement in any traffic- or policy-management capability today, as it would not make sense to have only part of the link between client and server so controlled. While we may even see elements of SDN in client firmware in the future, extending SDN functionality to the AP is likely on the near-term horizon.
It’s important to note here, by the way, that most SDN implementations have a logical function known as the SDN controller. This is not the same as a wireless-LAN controller or the control-plane functionality of all wireless-LAN systems, despite the similarity in nomenclature.
Just as controller-less WLAN architectures implement all of the functions of a controller-based solution without the requirement for additional (and redundant-configured) hardware, SDN implementations can also be highly-distributed into APs or even virtualized – more on that next time.
For now, the place to start is with policy and management strategy. SDN is going to have a major impact on (and benefits for) the evolution of not just the control plane, but also for the management plane of WLAN systems as well. This demands an unprecedented degree of interoperability in an industry that’s already famous for such, but SDN nirvana is going to take a while.
This is good news, as planning timeframes are uncompressed at least for the moment. But do spend a few minutes with your vendor to learn about their plans for SDN at your earliest convenience. Not all WLAN architectures will see the benefits of SDN – just having a WLAN controller, for example, doesn’t imply the presence or benefits of SDN!
We’re expecting to see a lot of diversity and creativity in the SDN space going forward, which is our topic for the next and final installment of this series.
But it’s important to remember the pivotal and strategic role that wireless LANs today play in all networks everywhere – and how SDN is likely to influence the evolution of your network not just in the core, but from edge to (wireless) edge.
All Posts In This Series: