Tom Nolle is an analyst blogging at several locations who most of the time shows good vision in his analyses. He is critical in a positive way about the current trends/hype in SDN and NFV. “The End to Networking as We Know It” is a blunt statement:
We have a long way to go before we can say that traditional routers are a dying breed, or that vendors like Cisco are at risk, but we do seem to have reached a point where technical advances in server/software technology will support real competition for routers for the first time.
I am not a supporter of this statement. Within the datacenter it might be true. But then you will still run into reliability issues using commodity hardware with general software. Redundancy is not always the solution to reliability, hence the business case for carrier grade solutions. Within the datacenter it might be overcome. But when it comes to large scale physical campus or (inter)national networks you get different demands.
There you have to deal with remote locations in sometimes small spaces which are not always accessibel. Redundancy in such networks is a bit more difficult and depends (currently) on convergence time. Then there is the issue with throughput in the core-network. Software-routers can still not cope with the amount of traffic in the core. That’s why software routers better be pushed to the edge, but not in far remote small locations.
That’s why the end of networking as we know it starts in the datacenter. With virtual servers, virtual routers and switches, connected by virtual wires. And it’s happening today.