As promised, in this post I will talk about the stakes of SDN paradigm in networking. The consequences of SDN would be similar to those of introducing abstraction layers into computing (virtualization, generic less expensive hardware, cloud computing, innovation, new pure software players...).
For service providers, the advantages are many:
- As switching hardware will be decoupled of control plane, hardware would become cheaper and interchangeable according to Moore's law. This forecasts reducing CAPEX of networks and dependence of equipment vendors.
- Software orchestration in OSS forecasts reduction in network operation (OPEX) costs, and faster service implementation thus improving Time To Market for introducing new services.
- As networking becomes software oriented, innovation will happen at software speed!
Clients will also benefit from SDN paradigm according to what was mentioned previously, in terms of costs, flexibility and innovation. Clients will consume networks on demand through a real time digital experience. Check as an illustration the amazing functionalities enabled on the portal of the WAN solution proposed by NTT :
In this evolving ecosystem, the traditional vertical business model of equipment vendors will be highly impacted. Decoupling hardware from software will let these vendors face low cost switching vendors from one side, and emerging players from software world.
As previously said, SDN does not remove complexity, but only reorganize and separate problems through modular layered abstractions. It means that some issues should be considered such as interoperability, scalability, reliability...
In Orange, SDN roadmap is resumed in 4 topics:
- Network virtualization in Orange datacenters
- IT driven networks for VPN to IaaS interconnection
- Deploy SDN PoPs in order to deliver on demand cloud-based networking services (connectivity, security, optimization...)
- Deploy switching-only CEs controlled by SDN PoPs
I am truly excited to witness the major shift in networking and be a part of it!
