[kepler-dev] Re: VNC for 'detached execution' of Kepler?
ludaesch at sdsc.edu
Thu Aug 5 10:19:01 PDT 2004
oops, here are some links:
previously this stuff was here:
>>>>> "BL" == Bertram Ludaescher <ludaesch at sdsc.edu> writes:
BL> Dear all:
BL> Steve Parker just mentioned the Virtual Network Computing Package.
BL> This could provide a (simple?) solution for some of our Kepler/HPC
BL> activities, i.e., where one starts a long running job and where the
BL> control instance/workflow engine (Kepler/Ptolemy) needs to continue to
BL> run as well for a long time.
BL> So here is an example:
BL> Someone wants to start a long running workflow/computational
BL> experiment (say via Nimrod, Condor, etc) from the laptop.
BL> Obviously, after disconnecting the workflow needs to continue to
BL> execute, and one should be able to reconnect to the running instance.
BL> By starting Kepler remotely on a "Kepler/Ptolemy engine server" (not
BL> necessarily the compute nodes, mind you!) and routing the display etc
BL> through VNC, it seems the above can be done (disconnect and reconnect
BL> So this is a bit like Citrix++ but for "real systems" ;-)
BL> Just a thought..
More information about the Kepler-dev