At ONUG Fall 2017, Apstra was invited to present an AOS™ Proof of Concept. With just 10 minutes to present, we were invited to show off the major new developments in our 2.0 release that enable custom telemetry functions and visualizations. I could talk about this stuff for hours, but was time-challenged to select just a few of the customer use cases that would be most relevant to the leaders of the open networking revolution.
AOS has at its heart a highly scalable datastore that can ingest a tremendous amount of state information and maintain the contextual value of each element. Couple this with a flexible GraphQL schema, and you can inspect both the desired state and the current operating conditions. There is virtually no delay in this process, so asking simple or complex questions about the entire network becomes trivial.
This is an amazing development in the world of IT, but the question now becomes how do customers want to use these capabilities? Do you know of a certain condition that exists in your environment that causes application issues? How many commands does it take for you to identify this issue? All of your expert knowledge about these events can be crafted into simple bits of code that can be run within AOS continuously. Got a best practice that you want to run everywhere? That takes only a few minutes to write, and then you are guaranteed that the network will always contain this search intent. You can even build your own remediation methods or simply alert the NOC or operators that a rule has been violated.
Please take 10 minutes to watch the video and then consider this question:
If you had these powers, what would YOU do to improve the state of your business operations?