The Network Automation Forum was founded recently by Chris Grundemann and Scott Robohn by challenging the network industry with this simple question – “Why haven’t we seen full adoption of network automation, yet?” The group’s inaugural conference, AutoCon, was held recently in Denver to tackle this question.
Cisco was represented by our Accedian team, a recent acquisition, in the sponsor showcase pavilion along with 20+ others, including partners, such as Itential, Netbox Labs, networktoCode, IP Fabric, Pliant, and PacketPushers. I was selected as one of seven speakers, along with 6 mini-tracks to provide content across the two days.
The conference drew almost 400 attendees, which was quite respectable for a first-time event. The hallway and mealtime conversations were enlightening and encouraging as we heard some of the challenges experienced in different end-user environments. It provided an opportunity to offer suggestions for remediating them with our rich automation and orchestration offers in concert with embedded product telemetry and instrumentation.
A challenge from John Willis
Those embracing DevOps principles would recognize the name John Willis (aka “Botchagalupe”) who was integral to its formation and continued development. John gave a keynote that shared the history of DevOps, reflected on the server/app community’s leadership, and challenged us, the network community, to likewise step-up. I appreciated having a few moments to share observations and consult with him in-person.
Several speakers, including vendor peers, end-users, and government representatives, shared their experiences – some even demoed automation solutions to an impressed audience.
Automation and observability
My talk included a behind-the-scenes view of automation and observability from the CiscoLive NOC perspective where quick set-up of thousands of wireless access points and hundreds of switches is a necessity. While much of AutoCon0 focused on automation from a configuration management and provisioning perspective, my challenge to the community was “what are you doing to automate your monitoring and network management?” Observability is enhanced by the rich telemetry and instrumentation found in Cisco’s products. If this data isn’t being extracted, valuable insights are lost. Using available commercial, open source and custom-developed solutions helps to cover a lot of bases.
Why aren’t we doing better?
Admittedly, much of the overall conference messaging was contemplative – “Why aren’t we doing better? We have Ansible, Terraform, telemetry, NSO, etc. – why isn’t the industry adoption farther along?” Some of my observations were the dependence on legacy methods using CLI. Some continue to push SNMP heavily, even though we have management protocols which have progressed well with NETCONF/YANG, gRPC and gNMI. As we become comfortable with these newer models the need to sustain legacy methods will reduce and greater efficiencies should be experienced.
Nonetheless, I was greatly encouraged by the innovation displayed, the vendors who shared their solutions and the overall community desire to move forward. If you have a passion for network automation and programmability, I encourage you to join the Network Automation Forum and consider attending a future event. NAF can be found at https://networkautomation.forum/
Additionally, our DevNet program at https://developer.cisco.com provides excellent resources to help propel your automation journey!
Related developer resources
- Learn more about Cisco Full-Stack Observability