Computer networks: Out with the old, in with the new

The times they are a-changin’ in the world of networking: Change some would say is completely overdue. It is a complex world occupied by LANs and WANs, switches and routers, subnets and VLANs, and much more; where change does not come easy.

First, the old: The TCP/IP protocols were developed in the 1970s. Then we had the OSI model, published in 1984. TCP/IP became the foundation of the Internet as its basic communication language: the whole concept of packets and error correction that has been underpinning all of networking so far.

And then the new: Researchers in Denmark are reported to have come up with a math based protocol that could replace TCP/IP and is said to be five to ten times faster. http://www.networkworld.com/article/2459286/why-tcp/ip-is-on-the-way-out/why-tcp/why-tcp/ip-is-on-the-way-out.html Presumably there are others elsewhere working on even more exciting possibilities. Meanwhile definitive change arrived in the form of the OpenFlow protocol, its genesis in Stanford University in 2006. Release of the first version of OpenFlow was in February 2011 by the Open Networking Foundation.

So it appears to have taken all of 30 years and more for real, meaningful change to happen: A surprisingly long period of time when you consider the rapid pace of developments in the software field and even in the server and storage areas. For example, Visual Basic first released in 1991 is considered to be legacy today.

Being ushered by the OpenFlow protocol and Software Defined Networking (SDN) is a new era where networks are centrally controlled by software. A transition from data packets to data streams and data flows. Big data is certainly the elephant in the room. With terabytes and petabytes of data hurtling down the pipe and data streams that need to be analyzed in real time, the current networking paradigm will likely be inadequate. IPv6’s stepped up pace of adoption could lead to every device having an IP address and communicating with the rest of the world. One could expect the increase in network traffic to be enormous, way beyond what today’s systems were designed to handle.

In the world of SDN with its white-box switches and routers, physical and virtual devices, software is certainly taking the upper hand. Software based controllers that set the rules for the networking gear in the layer below. Along with SDN applications for things like network provisioning, quality of service, security and so on that will communicate with the controllers. All the rules and logic embedded in code will mean that changes can be made “in an instant. “

Obviously no one is in a hurry to start replacing their critical networks anytime soon. But change seems to be inevitable. Slowly but surely, users will start to migrate toward SDN, perhaps when time comes for a major refresh. The allure of lower cost, increased flexibility and agility, and easier control should be just too hard to resist. More and more SDN players will emerge on the market with their own version of OpenFlow switches, routers, firewalls, IDS/IPS and so on. Increased competition will broaden the set of options available and further push down cost.

Required skill sets are likely to change. Some of the current areas, the command line interface skills, will gradually decline in importance. Instead in demand will be network programmers who can build software controllers and write applications atop the controllers. Colleges and universities will have to start providing this next generation of network engineers. Plus a number of the current pros will be rushing to reskill themselves to avoid being left behind.

Driven by the new set of protocols, the Internet should transform itself into a more responsive platform that is about speed, hugely increased bandwidth. Considering all the challenges facing the planet, the unifying role played by the Internet should become even more important. An Internet driven by an SDN stack should hopefully get us there…to a more integrated world that shares knowledge and information to meet global needs.

Ultimately the old hardware-defined order will yield place to a newer, software-defined order; the writing appears to be on the wall.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s