Strymon is a platform for online modelling of enterprise datacenter behavior

Our Research Get Started

Strymon (pronounced “stream-on”) is a system for carrying out online analyses and simulations of large datacenters. Strymon leverages existing logging and monitoring pipelines of modern production datacenters to ingest cross-layer events in a streaming fashion and predict possible effects of such events in what-if scenarios by simulating the hypothetical datacenter state alongside the real one. Strymon’s ultimate goal is to drive research and innovation in datacenter management. It advances the current practices in that the hypothetical datacenter state is fully exposed to Strymon users in a timely manner (as well-defined data streams), enabling the composition of deep analytics and cross-layer simulations in real time and within the same execution engine.

Latest news:

DS2 accepted at OSDI'18

July 24, 2018

Our paper “Three steps is all you need: fast, accurate, automatic scaling decisions for distributed streaming dataflows” by V. Kalavri, J. Liagouris, M. Hoffmann, D. Dimitrova, M. Forshaw, and T. Roscoe has been accepted for presentation at OSDI’18.

Strymon 0.2 Released

March 8, 2018

We are pleased to announce a new release of Strymon Core. This release comes with a redesigned publish-subscribe subsystem, overhauled documentation and testing, and other various bug fixes and improvements.

🎁 Strymon Holiday Season Release 🎁

December 22, 2017

We are very happy to announce our last release within the Strymon project for this year: The open-source release of SnailTrail and our work on online sessionization and structural reconstruction.