Working Group

Performance and Scale Working Group (PSWG)

Working Group Health

Getting some new energy with this WG.

There are several new papers / presentations that we are reviewing.

The group has   "moved up the stack" and started to focus on provenance in the supply chain space. We are looking to define basic performance and scalability measurements for that area.

We figuring out the best way to work with other working groups, SIGs, and projects to help draw out their thoughts, expertise, etc.

We are also looking at how to integrate Caliper output directly into the website. This is ongoing.

Attendance is starting to pick up a bit. We need to figure out the draw and get more people interested. Personally I think its a mixture of reviewing other performance work in addition to blazing new ground.


The move to Confluence has not gone smoothly. No time to go learn it.

Overall Activity in the Past Quarter

We held fairly regular meetings. The goal is to meet every week but some needed to get cancelled this summer.

The mail list and rocketchat have never been very popular with this WG.

Planned Work Products


Participant Diversity

Growing slightly A mix of academia and business

Additional Information

Reviewed By

  • Arnaud Le Hors
  • Baohua Yang
  • Binh Nguyen
  • Christopher Ferris
  • Dan Middleton
  • Hart Montgomery
  • Kelly Olson
  • Mark Wagner
  • Mic Bowman
  • Nathan George
  • Silas Davis

  • No labels

1 Comment

  1. Glad to see the increased traction. Getting concrete on a deliverable everyone can rally around will be good.

    Regarding posting caliper numbers, maybe figure out the goal first and then let the how follow. Like, what is the goal? Is it because different work group members are iterating on some collaborative performance enhancement? What's the best way to achieve that collaboration? it could be posting numbers is ancillary and the real necessity is shared test infrastructure. Probably goes without saying but the figures should align with PSWG's whitepaper and guidelines. In particular I think it's important to show the network size. i.e. optimizing distributed systems for small deployments is counterproductive.