Subject: Re: CVSup collections for a NetBSD CVS tree <199904262101.OAA11570@lestat.nas.nasa.gov>
To: Justin T. Gibbs <email@example.com>
From: Brian D Chase <firstname.lastname@example.org>
Date: 04/26/1999 18:42:35
On Mon, 26 Apr 1999, Justin T. Gibbs wrote:
> Not every NetBSD user will be pulling bits from the central
> repository. The load would be too high. They will have to use
> mirrors. How are these mirrors updated? The most efficient way is
> with CVSup. So how will an end user that cannot use CVSup benefit?
> By having the distribution to their local mirror site be efficient
> enough to allow more updates per day. FreeBSD mirror sites update
> once an hour and we could never sustain such a load on our master
> repository if we used another tool.
Hmmm... You're boasting of the benefits of CVSup, and you're holding up
the dependency of the FreeBSD mirrors upon it as proof of its value. But I
don't think anyone has challenged the "goodness" of what CVSup offers.
It's obviously a desireable tool to have in one's box.
The gripe is that the tool is only available to a subset of users. This
may not be an issue for FreeBSD, but NetBSD covers a much wider spectrum
of supported platforms. The issues of cross-platform compatiblity are of
great importance to this audience. Many of NetBSDs supported platform
have very meager resources relative to contemporary systems. These slower
systems (operating just as final destinations for the trees) would greatly
benefit from less overhead in transfering files.
And hey, what's to say that the mirror servers couldn't benefit just as
much as the master server by having to send less information. More
optimal use of bandwidth is something that is desireable across the board.
Brian "JARAI" Chase | http://world.std.com/~bdc/ | VAXZilla LIVES!!!