Graphs in Operations by John E. Vincent.
From the post:
Anyone who has ever used Puppet or Git has dabbled in graphs even if they don’t know it. However my interest in graphs in operations relates to the infrastructure as a whole. James Turnbull expressed it very well last year in Mt. View when discussion orchestration. Obviously this is a topic near and dear to my heart.
Right now much of orchestration is in the embryonic stages. We define relationships manually. We register watches on znodes. We define hard links between components in a stack. X depends on Y depends on Z. We’re not really being smart about it. If someone disagrees, I would LOVE to see a tool addressing the space.
Interesting post from a sysadmin perspective on the relationships that graphs could make explicit. And being made explicit, we could attach properties to those relationships (or associations in topic map talk).
Imagine the various *nix tools monitoring a user’s activities at multiple locations on the network and that data long with the relationships being merged with other data.
First saw this at Alex Popescu’s myNoSQL.