[kepler-dev] different versions of SSH actors
altintas at sdsc.edu
Mon Jun 18 16:37:11 PDT 2007
I agree that we need to think about this. It is becoming for some
users when they need an ssh actor. Norbert's one seems to be the
most elaborate one.
Dan and I will look into how the two actors could be consolidated
based on your suggestions when I am back from vacation.
On Jun 18, 2007, at 10:23 AM, Bertram Ludaescher wrote:
> I noticed there are different versions of Kepler SSH actors out
> there. If I'm not mistaken, the two most recent versions are
> authored/co-authored by Norbert and Efrat, respectively.
> Would it make sense to try and consolidate and unify these?
> Maybe the authors could send a quick summary of what each of these can
> and can't do. I recall that Norbert's version can maintain a single
> SSH session throughout multiple actors (btw, without showing the
> between them..) while Efrat's version can re-route stdin and stdout
> thanks, cheers
> PS Additional info follows: One way to think about this is that both
> have currently different "standard uses". Here is my cartoon for that:
> "OLD" SSH (Ilkay, then extended by Efrat?):
> - can redirect Kepler output to stdin of remote command, i.e.,
> "ties" the remote process into the Kepler dataflow:
> R RX1 RX2
> DATA /\ /\
> K --A---PA1----B--C---PA2--> D -->
> "ALTERNATE" SSH (Norbert):
> - can reuse an open SSH session
> R --- RX1 ----------> RX2 ----> ...
> CTRL /\ /\
> K--A---PA1-----B--C----PA2---> D -->
> Here R stands for the remote site, K stands for Kepler site.
> In the first setting, data is routed through Kepler;
> in the second setting, control is routed through Kepler.
> Kepler-dev mailing list
> Kepler-dev at ecoinformatics.org
Lab Director, Scientific Workflow Automation Technologies (SWAT)
San Diego Supercomputer Center(SDSC), UCSD
9500 Gilman Drive, MC: 0505 La Jolla, CA 92093-0505
phone: (858) 822-5453 fax: (858) 534-8303
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Kepler-dev