Network Automation Could Save Your Life

Network Automation is receiving a much-needed boost in attention, partly due to the efforts of NANOG, the Network Automation Forum (NAF), and other organizations.

Why this boost? Network Automation is moving slower than we thought it should and slower than we need it to move.

The Cost of Networks

As networks and network technology mature, we've all seen the trend to drive prices down: the prices for services from Carriers and the cost for networks in Enterprises. This aligns with the economic trend of markets "desiring" to be more efficient. No one wants to keep paying premium prices for network services (or any service, for that matter) that are becoming increasingly commonplace and that we are becoming increasingly dependent on. The Network costs the SP, the Enterprise, and the Cloud Provider.

We've already seen forces that have the potential to drive down the cost of networks in many areas (not a comprehensive list):

• Competition between vendors over 4+ decades has helped, but acquisitions and growth that outpaces the benefits of competition leave room for more cost reduction.
• Network Functions Virtualization (NFV) was an early attempt to use lower-cost software-only routers and other network devices. Still, the industry saw real limits to scale due to the limitations of generalized computer architectures.
• Developing "merchant silicon" and disaggregated network operating systems not tied to traditional vendor networking silicon (e.g., SONiC and multiple other new NOSs) provides more competition.
• OpenFlow and P4 were interesting experiments, but they had limitations.
• Network as a Service (NaaS) in multiple forms is helping customers get away from more costly and complicated network services from Carriers.

Automation and Cost Reduction

All of the above approaches (and more) move the needle to a limited extent, but there is still room for improvement. As a result, Network Automation will become an increasingly necessary component of the cost reduction for the Network. 

Here's why.

Automation happens - consider these select examples from other industries:

• Manufacturing: the assembly line 
• Office automation: the spreadsheet
• Telecommunications: the central office switch and PBX
• Retail: the self-checkout

We already employ some Automation in Networking:

• Protocols that have developed over the years are forms of automation that have allowed the Internet to scale and do the things it does today: moving from tracking IP addresses on a spreadsheet (thank you, office automation!) to DHCP, moving from RIP to OSPF and IS-IS, from EGP to BGP, are all examples.
• Network operators and individuals have been using bash scripts, Expect, Perl, Python, and more to automate certain functions for years.
• Network equipment vendors have developed and acquired their automation tools.
• Independent Automation platforms are emerging that make automation accessible to a much broader user base that can't just "roll their own.”
• We should learn from our Cloud siblings: Cloud providers have dealt with cost and scale in part by using merchant silicon in key areas (especially data centers), writing their own NOSs in some cases, making network designs simpler so they can scale more, and using a software-first mindset to automate the Network right along with apps, compute storage, and other IT stack resources.

Embracing Change

I said that Network Automation could save your life; OK, that's an exaggeration, but it could save your career. 

Here is my thinking:

• The carriers are clearly under pressure to cut costs; there is competition, and they are just not making the money and margins that they used to.
• There will be continued pressure for them to cut costs, and Network Automation is an increasingly viable way to do this.
• Some new tools and variables will continue to increase the power of Network Automation in the near future, including but not limited to AI, an abundance of inexpensive computing, and the example of cloud networking at scale.

The elephant in the room is FEAR: many are worried that Network Automation will eliminate their position.

• We know this will happen to some extent - there will be some disruption.
• AI generates the most fear right now, and it has the potential to accelerate network automation's capabilities significantly.
• Instead of simply fearing it, we need to embrace it and learn it - and what better learning environment to be in today with access to so many tools, videos, git repositories, and other learning aids?

As you learn and master concepts in Network Automation, you'll be well prepared as organizations evolve and need people with those skills:

• This is another form of disruption - as companies implement robots, there will be a need to maintain and oversee the robots, whether those robots are vehicles in a warehouse or software robots that automate your Network.

Good Disruption from Automation

Not only will network automation help save cost, it will also brings other positive disruption:

• It makes networks more resilient - this makes customers happier (another reason for companies to embrace it), and it cuts down on the number of outages that will get you up in the middle of the night.
• This can make the lives of network operations engineers better - which cuts down your stress and increases employee retention (yet another reason for companies to embrace it)

So get ready! More Network Automation is coming. Embrace it and get ahead of it now. Many people have gone down the path of Network Automation and are willing to help. If you want to learn more, please consider participating in the NAF.