Software Defined Networking (SDN) continues to move forwards; the next step
First published: October 2013
During an Analyst Briefing at the IBM Development Laboratory in Bucharest, a series of IBM executives took invited technology analysts through its thinking about the evolution of SDN and what will likely come in the future.
During the course of the day IBM staff talked much sense concerning general developments taking place and likely to occur in the coming years in datacentres around the globe. Unsurprisingly, given the nature of the work undertaken at the site, the importance of networking was illustrated and its likely integral role going forwards was discussed in detail. What especially stood out was IBM’s commitment to deliver improved automation of network management and administrative processes, to improve and even optimise network resource usage, and especially its importance to the optimisation of IT resources in general. The end SDN goal is clearly to help IT staff deliver better business services via greater dynamic management of networks/configurations/security with more attention thus being payable to service quality issues than has been the case in the past.
The discussions also took some time to consider the potential requirement for closer integration between the different elements of the IT management infrastructure as well as of IT management teams and skills. In the SDN of IBM’s future vision, ’orchestration’ of resources across platforms, potentially applying both inside and outside the organisation, will be the objective – though this “hybrid” capability will probably not be widely used at first.
As with all areas that impact multiple IT infrastructure technologies, vendors and management tools, standards are important, but as with all standards, organisations which want to move first will look to vendors to add quickly the required functionality on top of those standards, in effect asking SDN providers to take responsibility for maturing the technologies until the standards mature.
That said, SDN management tools have a long way to go. One example demonstrates this. In the discussions it was clear that IBM had not considered modelling tools either to facilitate management visualisation of a software defined network or to model ‘what-if network scenarios’. Networking is already complex. SDNs will add more complexity which can be counteracted by including tools which not only present what exists but which can discover what is changing. Making SDNs tangible will be a significant step towards acceptance (the converse is also true – too abstract and they will be left by the roadside).
IBM is not the only vendor working on SDN. Indeed enterprises will be encouraged that other vendors such as Brocade, HP, Cisco and Juniper, not to mention many smaller companies, are working together as well as competitively to move SDNs forward.