> From jon@zeeff.com Fri Aug 20 15:09:48 1999
> Delivery-Date: Fri, 20 Aug 1999 15:09:48 -0700
> Return-Path: jon@zeeff.com
> Subject: Re: Taxonomy draft - web replication problems
> X-ELM-OSV: (Our standard violations) no-mime=1; no-hdr-encoding=1
> To: touch@ISI.EDU (Joe Touch)
> Date: Fri, 20 Aug 1999 18:09:07 -0400 (EDT)
> Cc: wrec@cs.utk.edu, mbeck@cs.utk.edu
> From: jon@zeeff.com (Jon Zeeff)
>
>
> A common real-world difference is that one can execute cgi programs on
> a mirror and not on a cache. Other distinctions are more easily
> blurred.
That depends - there can be caches that save CGI output,
and, when seeing a matching request, reply the response.
And mirror usually refers to the static data - not necessarily
the dynamic code, e.g., the the CGI code. I don't know if that
has a separate term or not, but it's possible to have a 'passive'
mirror of the data without the CGI.
It might be useful to leave these kinds of distinctions
out of the definitions, unless they're necessary to
distinguish them.
Joe
>
> > It seems like there are a few dimensions in which caching and replication
> > differ:
> >
> > Caching Replication
> >
> > what is copied server output server source
> > (as Micah observed, though
> > I would apply it to all replication,
> > not just mirroring)
> >
> > who initiates it client server
> > (unless push-
> > cached, which
> > blurs the line)
>
This archive was generated by hypermail 2b29 : Thu Nov 18 2004 - 11:21:26 MST