Daniel Doyle

Come hear how ESnet has worked to put more of its portfolio of network measurement collection and analysis capabilities directly – and securely – into the hands of users. These services and APIs are built around ESnet’s Stardust measurement platform and allow users not only to view data about the network, but also to control how data is visualized and reported on. Learn about how these themes also extend into some of the open source and community efforts that ESnet is involved in, with technologies and patterns inspired from the successes in Stardust.

Being more of a philosophical approach rather than a single technology or implementation strategy, we will be touching on several topics over the course of this talk. Each topic will focus on where this need came from, our approach to delivering on the need, lessons learned along the way, and where we see it growing in the future.

– The need to collect a variety of data and metadata from different sources using different protocols, as well as the ability to react quickly to changes in user needs
– The growing need at ESnet for custom network maps targeting a range of different end users and goals, such as engineering diagrams, outreach maps, or marketing and overmap maps
– ESnet’s use of Grafana Enterprise to provide federated authentication and filtered views of network data for external users
– Community benefit in Netsage and perfSONAR from open sourced data collection pipelines and storage based on lessons learned from Stardust

Central to all of these topics is the idea of Zero Trust. As we work to make these services more multi tenant capable and available to a broader segment of users, we must be able to consistently authenticate and authorize these users to ensure that they are only able to see and make changes to appropriate data. We will touch on some of the current and future efforts in this space as well.

By providing users with increased ability to see and manipulate information about the network, we hope to optimize many workflows by removing dependencies on developer or network engineer cycles while simultaneously freeing those resources up to work on other high impact items.