[erlang-questions] On using Mnesia to pass data

[Ludovic Coquelle]

Hi everyone,
I would like some advice on "using mnesia or not".

The problem I try to solve: A master node run some processes on slave nodes,
and gather results of all those processes (master needs to know/admin which
process is start on which slave node; the context is a small load-testing
tool I am writing mostly to learn Erlang/OTP).

Current implementation is fairly simple/direct: Master start the processes
on slaves, and wait that the slaves send back results.

Looks like using mnesia could give me another implementation to solve the
problem: slave write data to mnesia, master read mnesia.
Is it a "good idea" to use mnesia only for data transfer? (don't care for
now that master will have to store data in some db, it's another issue)

Advantage I see to mnesia is that I don't have to manage all cases where
slave failed before the end, send half the data only, send all data once or
piece by piece...
Drawback is maybe efficiency: master and all the slaves will share all the
data, where I only need a data transfer from slave->master ... looks like
mnesia will do more than what I need: am I in such a case where "when you
have a hammer everything looks like a nail"?

Any thought?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20070307/53f2adec/attachment.htm>