August 30 - 31, 2017 | Four Seasons Hotel | Las Vegas, NV

< BACK TO TOPICSTransformation to Vendor – Agnostic Intent – Based Networking

The network is the critical platform that underpins all digital operation and transformation. However, for most companies, it is not operated as a system, but rather as a collection of devices that are manually configured to work together. In fact, 85% of network teams still use the Command Line Interface as the primary method to operate their networks. The result is a fragile digital infrastructure that is difficult and dangerous to change, incurs significant OPEX to keep working, and impedes the agility required to respond to business needs as well as incorporate best-of-breed devices without deference to a selected vendor. It is not surprising that many organizations spend an order of magnitude more on operations than on acquiring equipment in the first place.

Many can agree that network automation is the key to address these problems, but there is far less agreement on what this means and how to get there. Is it a collection of PERL scripts? Or maybe a few extensions to the server automation? Or?

In this talk, we describe how intent-based networking is the key to making the network an agile yet highly reliable platform with much lower OPEX and CAPEX. Intent raises the level of specification for the network engineer to allow him or her express what is needed, and not how to achieve it. A “network operating system” can then take that intent and instruct (i.e. configure) each device to behave so as to achieve this intent, as well as collect telemetry that continuously validates that this intent is being achieved, or else can notify the operator if the situation cannot be automatically corrected. 

The network engineer benefits by being able to express the higher-level intent explicitly, rather than having it lost in the complex error-prone task of translating this intent into lower-level CLI configurations for each involved device. The network engineer is also relieved of other low-level tasks such as generating cabling diagrams as well device configurations to achieve the intent. The network engineer is further empowered to have the entire state of the network at their fingertips, and to extend the telemetry in a matter of minutes to capture any parameters of his choosing – state that captures all the relationships and that they can query to answer any questions about the network. This unburdening frees up time to think strategically about the network, to both meet application demands as well as selecting best-of-breed network devices. By getting rid of manual waste, organizations that embrace intent-based networks can easily recoup their entire CAPEX costs; and including intangibles such as opportunity cost and ability to complete, the value creation can easily amount to orders of magnitude more.

David Cheriton
Founder and Chief Scientist at Apstra
Dr. David Cheriton is the Founder, Chief Scientist, and investor at Apstra, and a Professor of Computer Science and Electrical Engineering at Stanford University. Cheriton is a world-renowned researcher in the areas of networking and distributed systems with numerous widely cited publications and patents. He has also received the lifetime achievement award from the ACM Sigcomm professional group in 2003. David has an exceptional track record in identifying big market opportunities and building the architectures needed to address these opportunities. He has mentored dozens of entrepreneurs, founded and invested in numerous widely successful technology companies – e.g. Google, where he was an early advisor to Larry Page and Sergey Brin and among the first angel investors; VMware, where he was an advisor to Diane Greene and was an early angel investor; and Arista, where he was co-founder, Chief Scientist and an investor. Professor Cheriton was a Founder of Granite Systems, acquired by Cisco Systems in 1996, serving as a technical advisor at Cisco for 7 years, and acting as chief ASIC architect for the Catalyst 4X00 line. He was also Founder and CTO for Kealia, acquired by Sun Microsystems in 2004. Cheriton received his Ph.D. in Computer Science from the University of Waterloo in 1978 and he has been at Stanford since 1981.


Event: future:net

Where: Las Vegas, Nevada

When: August 30 - 31, 2017

Venue: Four Seasons Hotel