Mon 20 May 2024

  • Mon: Met with Anil
    • We skimmed the Carbon-aware name resolution paper. Anil said it was a narrow contribution, we’re making too many points in one paper, and that it needs to be framed in a research style with a hypothesis and experimentation attempting to disprove the hypothesis.
    • He liked the mailserver redundancy example (powering up the backup when the primary goes down), which he likened to self-scaling applications managing the power state of machines.
    • He said that the carbon-aware load balancing (Patrick’s section) seems more general than specific to emissions.
    • Anil also said carbon isn’t the best terminology. I agree – we were following HotCarbon’s wording.
    • We talked about doing a simulation to disprove the hypothesis.
    • And how there’s a pareto front of latency/availability vs power efficiency.
    • As a next step Anil requested a table of protocols and their properties regarding local and global operation, their data model, key management, etc, for consideration of how they can be made carbon-aware (and spatially aware?).
  • Tue: Preparing for supervising concepts of PL and supervising concepts of PL.
  • Wed: Tarides Opam Nix system dependancy management.
  • Thur: Created table of protocols for consideration in carbon-aware name resolution and spatial networking.
  • Fri: Met with Anil and Patick, and ended up talking about Nix a lot. I said I’d send over my configuration as a starting place for Anil to re-try NixOS.

I read RFC2352 which descibes a scheme for creating domain names based on legal names, namespaced by country. This is similar to the geocivic heirarchy described in the spatial name system paper.