project8 / dripline

Slow controls for medium scale physics experiments based on AMQP centralized messaging
http://www.project8.org/dripline
1 stars 0 forks source link

how should we deal with large payloads? #168

Open laroque opened 7 years ago

laroque commented 7 years ago

There has been some discussion lately about how to best deal with large payloads. In particular, we've been trying to trying to transfer ~hundreds of kB of data from an SQL select for storage elsewhere, though in the past we've also transferred long arrays of data from, say, the lockin or others. I think that the dripline standard itself should address this, a few options come to mind:

1) Set some cap on message payload size and define a means by which larger payloads can be split, sent as a set of multiple replies, and combined at the end. This is non-trivial effort but scales and should be a "forever" solution. 2) Dictate that dripline communication is for instructions and replies, but not larger data transfers. Qualitatively state that "large" transfers are not supported and/or discouraged, and then continue to move forward roughly as we are now. This is less precise but seems to be what is happening naturally. 3) some other idea?