Watching software defined networking take hold
There’s no doubt that networking architectures are moving towards an SDN reality, and that it will eventually completely revamp how we think of a network today. However, challenges still remain at the data plane, control plane, and even application environment that will need to be solved.
According to the Open Networking Foundation (ONF), SDN is defined as the physical separation of the network control plane from the forwarding plane, and where a control plane controls several devices.
The ONF further defines SDN as:
-
Directly programmable
-
Agile
-
Centrally managed
-
Programmatically configured
-
Open standards-based and vendor-neutral
Top down, SDN is at the top an application layer, with a control plane in the middle and a data-forwarding plane at the bottom. SDN requires a way for the control plane and data plane to communicate. OpenFlow is an open communications protocol standard involving interaction between control and forwarding planes, but it is potentially not the only solution for SDN.
Today, building networks, configuring, or reconfiguring them involves manually dealing with each switch used, [a process] which will, under SDN be automated. Server administrators, cutting out a complete level of certified network engineers, will remotely accomplish network management.
Why SDN is so important, is that it addresses today’s traffic patterns based on public and private clouds. It adds flexibility to management and access on demand. It also addresses the future needs of big data. In comparison, today’s highly complex networks cannot scale sufficiently to support the explosion of traffic. They are also vendor-dependent, further challenging standards, interfaces, and the very speed of innovation that will be necessary to keep up.
What gives credence to all things steaming rapidly towards SDN? Big players such as Cisco and VMWare have already acquired such early SDN players as Insieme Networks and Nicera. Analysts are scrambling to put a price tag on SDN. IDC says it will grow to an impressive $3.7 billion from 2013’s $360 million level, and do it in just three years—seems that adds to credibility as well.
SDN, however, is just starting to ramp with full-scale deployments not likely until late in 2015, according to a recent Credit Swisse report. There are still many challenges on each of the layers that must be solved. In the meantime, watching the evolution of SDN, along with the even-more rapid evolution of the cloud is in order, as is the question of what the impact will be on not just silicon, but on the underlying software development itself?
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
