SDN and NFV: why adoption should be an evolution not a revolution

Telecoms.com periodically invites expert third parties to share their views on the industry’s most pressing issues. In this post, Adrian Brookes, head of networking, Avaya EU, looks at best practice for telcos looking to move to NFV and SDN.

According to analyst firm Technology Business Research, enterprise operators will turn increasingly to consolidation, Network Function Virtualisation (NFV) and Software Defined Networking (SDN) technologies throughout 2016 to improve profitability.

They anticipate that telecoms operators will continue to struggle to grow revenue this due to the saturation of traditional services, competitive pricing pressures and weakening foreign exchange rates.  According to the firm, carriers are turning to SDN and NFV solutions, because by adopting these technologies they can reduce costs and gain agility in service offerings. Operators who move now will also gain time-to-market advantage, helping them to attract enterprise customers before NFV and SDN become commonplace.

However, I think operators who move too quickly could be missing a crucial step. I believe strongly that while SDN and NFV can help to address these challenges, it is critical to fix the foundations of the network, utilising modern approaches, before attempting anything with the level of complexity associated with deploying end to end services.

The current approach of nodal configuration using legacy protocols carries significant risk due to its configuration complexity and implementation, which requires multiple nodes to be reconfigured. Therefore, enterprises will normally go through an off-production lab testing and validation first. They are likely to demand a maintenance window to give them time to apply changes, and have an opportunity to rollback, in the event of a problem occurring due to misconfiguration of one of the nodes. This could be as simple as a ‘copy and paste’ typo in a script or a wrong port number, which triggers a loop on the network!

By moving to a services based architecture and focusing on point of service provisioning as opposed to a nodal model, customers can finally gain the agility and simplicity they have been looking for. However, moving CLI scripting to SDN programming does not necessarily deliver the customer’s expectations.

To avoid this, I recommend a different approach: solving the control plane issues, while adding the ability to interface with other SDN controllers to the edge of its architecture. Simultaneously, I’d maintain a simplified end-to-end single protocol architecture based on a standard IEEE Protocol, Shortest Path Bridging (SPB 802.1aq).

From my point of view, it’s critical to focus on the bigger picture and end-to-end services, rather than just concentrating on the data centre. The orchestration abilities of SDN may help realise this holistic view by enabling more end to end services,  but it’s going to be some time before the provisioning tools are sophisticated enough to achieve this across an entire network. The problems of legacy architecture are not going to disappear overnight, even after you deploy yet another management interface to provide centralised control. The same protocols are being used, which means the old recovery times and associated risks remain.

By virtualising the enterprise and moving to a services-based architecture, other virtualised network services, defined as NFV, can easily be integrated. For example, a firewall service or a session border controller can easily be integrated with the already virtualised networking infrastructure. In addition, the risk of IP DoS attacks and hacking is greatly reduced by moving away from nodal configuration. This is because Ethernet topography is used to establish communications with IP services, different to other vendors who continue to use the hop-by-hop principle of controlling data flow.

Ultimately, operators need to think of the adoption of SDN and NFV as an evolution not a revolution. It is critical that they maintain their current assets while evolving to SDN and service providers need to carefully evaluate the solutions available before taking the leap.

Adrian Brookes is head of EU Networking and Corporate Consulting Director for Strategic Solutions at Avaya.  He is responsible for the implementation of networking projects in the EU. On a day-to-day basis he is also an evangelist for Fabric Connect, SDN, and Security technologies. Prior to joining Avaya, Brookes held a number of senior technical sales and consulting roles at Unify, Siemens and Cisco.