(Series

Til
Page content

Why You Should Care About Model-Driven Telemetry from CISCO blog

Summary

  • Periodic polling -> Pushing(Streaming)
  • Model based data with YANG
  • Use standard encoding such as JSON, Google Protocol Buffers
  • Easy to manipulate, connect to analytic solutions

1-min poll is too slow

For the last 25 years, network operators have heavily relied on SNMP polling and CLI screen-scraping to extract operational data from the network. But the new and automated demands of today’s networks have pushed these mechanisms to the breaking point.

Network operators often poll data from their network on the order of every five to thirty minutes. With careful tuning and testing, the bravest can push that interval down to one minute. But with today’s speeds and scales, even that’s not low enough to capture important network events. And as we move to higher density platforms, the amount of important operational data becomes truly staggering.

streaming

Network operators poll periodically because they want the data at regular intervals.
Instead of pulling data off the network, sit back and let the network push it to you

Model-driven

Of course, it’s not enough to just push a lot of data off the device. Telemetry data must be structured in a sensible way to make it easy for monitoring tools to ingest. In other words, good telemetry data must be model-based. YANG,

image-2-for-blog-550x299

Analytic friendly

Telemetry data needs to be normalized for efficient consumption by Big Data tools. In the software world, encodings such as JSON and Google Protocol Buffers (GPB) are widely used to transfer data between software applications. These encodings have an abundance of open-source software APIs that make it easy to manipulate and analyze the data.

Model-driven telemetry is your first step in a journey that will transform how you monitor and operate networks. With the power of telemetry, you’ll discover things you never imagined and begin to ask more and better questions.

References cited in the article