[R] cgi/servlets/httpd in R

Samuelson, Frank* FWS4 at CDRH.FDA.GOV
Mon May 10 15:20:55 CEST 2004


Actually, what I have is probably not what you want, and I wrote
it as an example of what I was looking for, and what is probably possible. 
My code is just a hack of the snow socket code.

To answer your questions:
1.  The communication is done by serializing R objects.  That is,
the process connecting to the R compute nodes is an R process as 
well.  Not helpful in this situation.

2. Again, not what you want.  In the cluster every node has 1 R
job, and all get connected simultaneously until 
they're done computing, then released simultaneously.

2.5  No, see 1.

3.  You may have the code if you want, but like I said, you probably don't.

These recent posts gave some good information though:
https://www.stat.math.ethz.ch/pipermail/r-help/2004-May/049239.html
https://www.stat.math.ethz.ch/pipermail/r-help/2004-May/049321.html

I think RZope/RSOAP and/or Rserve may be what 
we're looking for.   They're not on CRAN and weren't listed in the faq.

-Frank


-----Original Message-----
From: wwsprague at ucdavis.edu [mailto:wwsprague at ucdavis.edu] 
Sent: Friday, May 07, 2004 12:50 PM
To: Samuelson, Frank*
Subject: Re: [R] cgi/servlets/httpd in R


Samuelson, Frank* wrote:


> Currently I use such a method on my computing cluster:  All 40 
> compute nodes run an R process/compute server that listens at a socket for
> any 
> connection and subsequent commands from another computer. 
> When the master process disconnects, the R processes go back to listening
> at the socket.  Connecting to the R compute servers this way
> takes < 2 milliseconds rather than the typical ~2 second R startup time.

Your solution seems what I am looking for, actually.  Questions:

1.  How do you communicate with an R process?  I can only think of 
embedded code like SQL in Perl.

2.  How do you find an R node that is open?

2.5  You are haveing cgi scripts connect to the R processes, right?  Or 
are the R processes available for another purpose?

3.  How much of your code are you willing to share :)?  We are 
interested in making demographic analysis tools available online 
(life-table-ish stuff to start with, then as much as we can get grant 
money for).

W




More information about the R-help mailing list