Next week the Packet Pushers are hosting a free one-hour livestream with Gluware to explore how Gluware aims to democratize automation; that is, get you quick wins around common tasks such as configuration changes and OS updates, while also helping you evolve toward more complex automation capabilities without having to become a master of coding.
Why might this be important to you? Because network automation is still bloody hard, especially in brownfield data centers where networks have accumulated years of technical plaque: overlays, ACLs, VLANs, security policies, firewall rules, heterogeneous equipment, unpatched network OSs, and lightly documented configurations. Every potential change has to be weighed against known (and unknown) dependencies and implemented with the same caution and anxiety as a dentist’s drill near an exposed nerve.
This is not an automation-friendly environment. At the same time, most automation schemes are themselves not particularly friendly. Consider this Network Field Day presentation from Arista Networks on building out a network CI/CD pipeline. The pipeline is built from multiple open-source projects:
- VSCode as an IDE
- Gitlab as a repository and CI engine
- Batfish for pre-deployment testing and change validation
- Ansible for playbook-based deployment & orchestration
- Nautobot for your network source of truth
My first thought on seeing all these tools in the pipeline was “How is this feasible for most enterprises?” Here are five separate open source projects to download, patch, and get running. Then you have to learn how to use all five individually before you string them together, and then figure out how to integrate them into your day-to-day workflow. Is this person taking on the pipeline allowed to take time away from their day job to learn all this, or will it be a nights-and-weekends effort?
Assuming this person has the spare time and initiative to take on such a project and master it, what happens if there’s a mistake that takes down the network? Does this engineer have the support and backing of managers and executives to screw up, or has all that time and effort simply gone into a career-ending event? And even if all those hurdles are jumped, what happens to the pipeline when the person who developed all this knowledge and expertise moves on to another company?
This isn’t a dig against Arista. Artista is making a good-faith effort to support an automation framework that doesn’t lock customers into a single vendor’s products or software, and supports open, extensible tools. I should also note that Arista can take away some of the burden of the DIY approach because its Cloudvision platform can slot in to the Ansible and Nautobot sections of the pipeline.
My point is that getting such a pipeline off the ground is untenable for most enterprises. We need a simpler, more sustainable approach. That means a framework that doesn’t require a huge personal investment of time and effort before you get value, ties into how network engineers already operate, and provides a pathway to go from quick automation wins to more complex, NetOps-style initiatives.
Arista would position Cloudvision as the solution. Juniper touts Mist. Cisco has ACI. There are also third-party network automation tools available from vendors such as Gluware, Itential, and Anuta Networks (among others).
Get Informed For Free
If network automation is a priority for you, come join the livestream, taking place on September 28th. The Packet Pushers will host six informative segments with Gluware, including technical presentations, real customer experiences, and a conversation with networking legend Terry Slattery. We’ll also have a Q&A segment to answer your questions. If you’ve got an hour to spare for network automation research, sign up here and join us on Tuesday the 28th and find out whether Gluware is the right platform to help you get your arms around automation.
