Python web development, really

Jacob Smullyan smulloni at bracknell.smullyan.org
Fri Jan 24 18:38:16 CET 2003


In article <7x3cni1plo.fsf at ruckus.brouhaha.com>, Paul Rubin wrote:
> I don't understand why more
> multi-process servers don't use shared memory instead of disk files,
> or even (gak) SQL databases and worse.  Sometimes I wonder whether MPI
> (the pseudo-shared memory used in scientific clusters like Beowulf)
> might be an appropriate abstraction for programs like web servers.
> There are very efficient MPI implementations that let processes on the
> same machine communicate through shared memory, and can also let
> processes on separate machines communicate through a LAN.

A cache like SkunkWeb's needs to be able to hold a large amount of
data, and accessing it has to be faster than recalculating it (which
in the real world often involves going to a SQL database for it).  A
naive implementation with shared memory and sql databases is unlikely
to meet those requirements; however, a naive implementation with disk
files works quite well.  When a naive implementation works well
enough, it isn't unreasonable to continue using it, and innovate
elsewhere.

jacob smullyan





More information about the Python-list mailing list