httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mark D. Anderson" <>
Subject Re: Apache 2.0 ideas
Date Mon, 16 Nov 1998 07:29:15 GMT
>> the network. However, you lost me in your transition from "1Gb" of data
>> to "1Gb/s" of bandwidth. The bandwidth to the system could consist
>If you have more than 1Gb of data which is "active", i.e. required, in any
>given second, then you're dealing with at least 1Gb/s of outgoing
>bandwidth.  Yeah you could need more, I'm just saying 1Gbyte is the
>minimum -- any less bandwidth and you'd be reusing some data, i.e. you
>wouldn't "require" all the data.  This is an extreme case, and as you
>point out, far from what a real site would see... which is why I was
>implying "why don't you think the disk can fit in RAM?"

I think we all agree with the conclusion (most web sites can fit in
an affordable amount of RAM). But this argument still doesn't work for me.
Suppose i have a movie on my 22Gb DVD, and i want to play that to
someone over the net. I'm not going to get 22Gb of RAM, and even
if i did, it wouldn't really do much for me (hence my reference to
Gray's "5 minute rule").

>> (2) Regarding implementation approach, what I was trying to ask in
>> my stumbling fashion was for a comparison between: (a) mapping the
>Oh, if you want to play fun games, then you don't want a real filesystem.
>What you want is a disk layout that very closely resembles TCP/IP packets
>including precomputed checksums and the work.  Then a minor amount of
>twiddling at xmit time gets you the packets you want for the wire.

Right, and then we can start sending the packets proactively, and
achieve negative latency :).

But seriously, isn't the design of a web server's caching algorithm
confronted with the same set of issues as a virtual memory system -- 
because in fact it is doing practically the same thing?
To me, that means two things: (a) in an OS with a well-designed working
set algorithm (this excludes NT), it should work fine to just use the
virtual memory system directly, and (b) even if that isn't done, surely
that is a well-studied area and applicable here -- and surely it should
be possible to simplify the tuning/configuration to just telling apache
how much RAM it has to play with, and it'll do the right thing, given
enough history with past requests.

On a somewhat unrelated note -- and apologize for slipping this question in here --
is there an apache module for hooking the caching algorithm, so that i
can fill and invalidate the cache for POST requests using my own information?


View raw message