Re: Why khttpd is a bad idea (was a pointless argument about

Alan Cox (alan@lxorguk.ukuu.org.uk)
Fri, 18 Jun 1999 20:04:09 +0100 (BST)


> > ftp, even finger,
>
> High latency, request per connection design -- not exactly a thing
> where performance improvements from async sendfile will be noticeable.

Scalability on ftp is important, async sendfile reduces switching overheads
and thread count (remember 8K task struct/thread, cache and tlb overhead..)

> And IMHO the problem with the number of threads/processed necessary for
> the implementation (that still can be managed efficiently) seems to be
> so significant now because existing HTTP servers in all their forms
> have large memory footprint thus causing unnecessary paging when the
> number of requests increases.

So you want to solve the "its big" problem with "its big but I stuffed it
in the kernel" ?

Apache doesnt use CLONE and VM sharing. That is an apache problem.

Alan

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/