Software-Defined Networks (SDN) transform hard-wired corporate networks into a programmable communications infrastructure, adaptable to fast-evolving business needs,Joy Ghosh - VP, Asia Pacific & Japan, Extreme Networks, Inc.

The concept of Software-Defined Networks (SDN) has been around for many years, but it has suddenly become a red-hot topic. Much has been claimed for SDN, and the number of competing voices has spread a certain amount of confusion about SDN, the role of OpenFlow and the Open Networking Foundation (ONF). Extreme Networks is among the leading pioneers of the SDN approach, with many years experience in the field, and is a member of the ONF. So this article sets out to answer some of the basic questions about SDN and what it can do for the enterprise – the sort of questions we have been answering for our own customers since 2004.

The current media buzz around SDN really began in October last year with the first Open Networking Summit. This was the first public industry event exclusively dedicated to SDN and it was an immediate hit. The organisers had arranged seating for an ‘optimistic’ 350 delegates, but quickly had 600 applications to attend – a sign of just how much interest was already there.

The summit was jointly organised by Stanford University and the Open Network Foundation (ONF) – an organisation to promote SDN that had just been founded with members including Deutsche Telekom, Facebook, Google, Microsoft, Verizon and Yahoo!, and expanding to 70 members within months. It is still growing, with a notable surge of interest in Europe at present.

The ONF and the mission of SDNThe inspiration for the ONF had its roots in academia, notably Stanford University, and a mission to evolve networking along lines not unlike the way that computing had evolved as a discipline.

The very earliest computing devices were logic machines hard wired to solve one specific problem, and the next step in electrical computing was to incorporate movable plugs to allow a certain amount of re-configuration of the hardware – so that one machine could be re-shaped to solve a range of calculations. But true computing began with the use of “software”: instead of making manual changes to the hardware, instructions could be fed into machine that would automatically configure its logical structure for the next task. The same machine that had been doing scientific calculations could now, in a few seconds, become a machine for doing the company accounts.

On the basis of this new, flexible infrastructure there evolved the concept of “higher level languages”. The programmer writing software no longer needed to think in terms of configuring a logic array node by node for a new task, but could use familiar words like ADD, MULTIPLY etc to initiate standard operations. From there, computing has evolved to the point where anyone using software like MS Word can simply click on one icon and call up sophisticated processes, from spell-checks to auto-formatting and more.

In these terms, computing has come a long way, while its sister discipline, networking, lags way behind. Networks have hardly evolved beyond that second stage that relied on manual re-configuration.

Let us say you have a fundamental idea that could accelerate your corporate network – how could you test it? Here is an example: one student at Stanford wanted to experiment with load balancing. Load balancers usually direct traffic to the server with the lightest load, regardless of how congested the traffic to that server. What would happen if the balancer also took the traffic into account?

In theory this is a very simple experiment, but would anyone dare try it out on a large working network? Just think of the labour, the potential disruption during set-up and consequent risk that the experiment might unbalance such a complex system and lead to unforeseen consequences – with so many users relying on the service.

In fact, this experiment was actually carried out, first on Stanford’s own network and then nationwide on the GENI test network. Professor Nick McKeown, speaking at the 2011 Open Networking Summit put it like this: "For the first time I've seen a graduate student able to take an idea and run it on a national network".

The reason that this was allowable was that Stanford was pioneering a new approach to SDN, based on a standard now named OpenFlow.

OpenFlow, and the separation of the control planeProfessor McKeown is the man behind the creation of OpenFlow – an industry-standard protocol that allows network operators to reprogram a network’s control plane from a central interface. Instead of having to go into the physical network and tweak masses of boxes, general instructions can be sent out across the entire network, or subsections of the network – just as the introduction of software made it possible to automatically configure a computer for different purposes, without having to manually alter its structure. This makes the network into “a software-defined network".

Whereas in a normal router or switch the fast packet forwarding (data path) and the high-level routing decisions (control path) happen in the same device, with OpenFlow these 2 functions are separated: the datapath still resides on the switch, while the high-level routing decisions are moved to a separate controller. The open flow switch and controller communicate via the OpenFlow protocol.

OpenFlow began in Stanford University, with version 1.1.0 released in February 2011, and was first demonstrated at InterOpNet Lab in May 2011. Then the ONF was created and now has responsibility for further development of the standard.

OpenFlow-enabled switches are already available from Extreme Networks – the first Ethernet switch vendor to integrate OpenFlow into its entire product portfolio – and other major vendors are increasingly supporting the standard. Incorporating these switches into a network allows the easy deployment of innovative routing and switching protocols, not only to optimize performance, but also to address specific issues such as network flexibility to support virtual machine mobility, high security networking and next generation IP-based mobile networks.

There is a lot more work to be done to fully exploit the benefits of SDN. The existence of a separate control plane now makes it possible to program the network from a central console, but initially this is still a relatively piecemeal process, like writing a computer program in machine language. But it does lay the foundation for a new network software discipline, working towards a high-level language that will make networks as readily programmable as a PC – allowing fundamental changes to be selectively broadcast right across the network as easily as clicking a word processor icon.

SDN, the broader pictureOpenFlow has hit the headlines, but don’t forget that the basic idea of SDN has been around much longer. My own company, Extreme, has been operating in the SDN space since 2004 when it launched XOS, a common operating system allowing scripting across its networking products to adapt to changing demands on the network such as mobility, multi-tenancy, policy-based networking and so on.

Other major switch manufacturers have also adopted SDN, but the point is that these pioneers have each developed their own proprietary approach – a good solution so long as you stick to a single supplier.

The key fact about OpenFlow is that it is a vendor-agnostic industry standard that can be added to OpenFlow-enabled Ethernet switches, routers and wireless access points from any supplier. The more vendors start supporting OpenFlow, the bigger choice of equipment there will be for your software-defined network – and that is why Extreme was among the first to put its weight behind the newly created ONF.

Although OpenFlow is just one way out of many to implement SDN, its importance is that it is the first to offer a standardized interface between the switches and the SDN controllers. Again and again in industry we see the importance of this – once standards have been established the implementation can really take off.

OpenFlow has the potential to become the “Android of networking” – an open standard that will encourage an open marketplace of new SDN applications to meet every networking need and business pressure.

Is the proliferation of BYOD (‘bring your own device’) in your enterprise giving the CIO headaches? Then there will be a choice of off-the-shelf mobility management apps for your OpenFlow controller to help handle the pressure and deliver optimal service to your users. There will be network controller apps to help with identity management, security, policy-based routing, differentiated QoS or whatever you need.

Extreme is paving the way with the hosting of a dedicated network application store – encouraging the growth of a crowd-sourced marketplace along the lines today’s Android app-store. The plan is that it will grow to offer free apps, premium and freemium apps – whatever you choose it means that tomorrow’s networks will have the flexibility to adapt to every business need without the time, cost and above all risk of rolling out manual changes across the network.

That vision of the future of enterprise networking explains why a company like Extreme – that has already been satisfying its customers evolving networking needs with state-of-the-art SDN solutions since 2004 – is putting its weight behind the ONF and a nascent standard called OpenFlow.

Today’s drivers for SDN in the enterpriseIt is hardly necessary to spell out for the enterprise IT readership the many business, economic and social pressures now impacting their already demanding roles. Enough has already been written about the complexity, management and security problems caused by the transition from yesterday’s closed, plug in network to today’s permeable wireless-enabled networks linking a shifting population of mobile users with many of their own personal devices on the network. There are the pressures of increasing virtualization, of reliance on video and other bandwidth-hungry applications, the problem of IPv6 readiness, of new threats to data security and government legislation over privacy laws and border control of data.

The point is that replacing the brick-by-brick hardware management of a network with central software control system opens up a whole new world of opportunities for the IT department. With global visibility and control you can dynamically provision the network and make it more elastic and optimize the use of resources. SDN traffic management allows network virtualization and can help support cloud services, including hybrid cloud computing. Extreme's unified hypervisor, XNV, integrates with the popular server virtualization technologies from VMware, Microsoft and Citrix and most recently, Linux kernel VM (KVM) to bring complete network-level visibility and control to virtualization. Together with OpenFlow, XNV increases customer choice and breaks vendor lock-in.

Service-aware routing can streamline delivery of video and other lengthy traffic flows according to a user’s profile, the congestion state of network, and other parameters. SDN permits scalable, granular bandwidth allocation: you can use SDN to program the right access, security, QoS, and other policies along with the pipelines for each flow to offer high-level services, or an optimal user experience, without having to configure a whole lot of boxes along the path. A user could switch between chatting on their smart phone, watching a video, and streaming music without any service degradation. Videoconferences and interactive traffic can be sent down paths engineered for optimal experience while other traffic is routed normally.

The question is not whether SDN has any relevance for enterprise networking, but rather: what is the best way to find out more?

The way forwardThis article has aimed to present SDN and OpenFlow in the simplest way, but the fact is that it is the beginning of a complex and fast-evolving topic. Much has been written already on the subject, but a lot of it is speculative or biased towards a particular company’s preferred technology and approach.

By all means read further about the subject, but every business is different and it is better to consult a company with long experience in delivering SDN solutions. In view of the key role currently being played by OpenFlow in accelerating SDN development, you need a company that supports an open standards-based approach rather than locking you into a proprietary straightjacket – so choose a member of the ONF to advise you.

SDN is a hot topic, and for very good reasons. The best strategy is to be informed and to keep an eye on developments. Even if your company is not yet in dire need of this technology, SDN will be playing an increasing role in maintaining competitiveness.

A more nimble network will help your business outpace the competition – don’t be left behind.