NFV is no doubt ground breaking and a disruptive force. But is it really a new model and a new concept? Is there something new and recent technology breakthrough that triggered this market? Does this suffer from broader adoption due to lack of maturity of solutions or questions about feasibility of these solutions? The answer is a no to all of them.
Managed services concept was tried in late 90s and early 2000s and Service Providers were interested in these solutions as a means to reduce OPEX and expand market reach. The value proposition then was primarily based on OPEX reduction and not on CAPEX as the equipment costs were still prohibitively high. Shasta Networks, Cosine Communications and Corona Networks were few companies that focused in this area.
Fast forward to cloud era, where the IaaS and PaaS solutions are mature and economical and COTS performances have shot up tremendously that CAPEX reduction also became a possibility. So for Service Providers who have been looking for a long time to reduce CAPEX the timing has become right to come together and drive the market requirements for NFV.
NFV management solutions can and will inherit a lot of components that are well proven in cloud management solutions space to provide scale and performance. These management solutions have been deployed on a large scale on traditional IaaS and PaaS infrastructures. NFVI and IaaS are closely related with the added notion of network virtualization layer in NFVI. Amazon has built one of the largest and now mature cloud based services with home grown solutions during the early stages of cloud market. It will be relatively easier act to follow with NFV for Service Providers.
But NFV does require a different twist and customization to traditional cloud management solutions and platform or packet processing functions. Among others differences in management solutions, one key distinction lies in packet processing components called VNFs that require scale. But again OS bypass solutions such as DPDK are providing tremendous performance boost to the VNFs. Interestingly even OS bypass solutions have been existence since 2003. Companies such as Precision I/O and other bigger names such as SUN microsystem implemented OS bypass solutions to accelerate IO throughput and reduce latency for Grid computing applications as early as 2003. So many of the NFV key technology components are nothing but evolutionary in nature and some of them are mature and were tried before and deployed in other markets for a while now.
Which leaves one to wonder if there a room for skepticism about building a decently mature NFV solution? Again the answer is no. Perhaps the adoption rate could be faster. NFV adoption will be impacted from inertia of established methods and practices more than anything else. But this goes with any new solutions replacing old solutions. Certainly R&D is very active with new POCs and related projects. Investment and funding to enable a new eco system of solution providers perhaps could accelerate the deployments. But sooner or later and by “law of maximizing returns” NFV will take roots. It remains to be seen how quickly NFV will mature in comparison to other cloud solutions.