Re: automatic compression over communication channel ?

From: Barry Margolin (barmar_at_alum.mit.edu)
Date: 01/29/04 

Date: Thu, 29 Jan 2004 17:12:51 GMT

In article <bvbdlp$psc72$1@ID-219297.news.uni-berlin.de>,
 "Vu Pham" <vu@sivell.com> wrote:

> I am sorry if this question is not appropriate for this forum.
>
> In my client/server app, I compress the data at one end and the other will
> decompress it. To do this, I must implement the zlib api at the
> "application" level : I know in advance how long this transaction will be,
> then I compress it, add a flag and some parameters to notify the other end
> about this compression ...For apps that currently do not have the
> compression, I need to change a lot of codes..
>
> I wonder if there is any way I can implement the compression at a lower
> level - somewhere below my application level and above the tcp level - so
> that I just need to call, say zsend() or zrecv() and these functions will do
> the compression/decompression. stuff.

AFAIK there's no standard protocol for this. You could write a library
that implements it, and then use it in any applications that need
compression.

If you think this would be generally useful for lots of protocols, you
could design an architecture analogous to SSL. Just as SSL provides a
general mechanism to run any application protocol through a secure
channel, you could design a mechanism to allow any protocol to be run
through a compressed channel (you could call it CSL, for Compressed
Socket Layer). Or maybe you could piggy-back on SSL itself. I think it
starts off by negotiating the security mechanism being used -- you could
implement a mechanism that doesn't really perform encryption, but just
compression (to the software modules it doesn't really make a
difference, they're all just different ways to encode data). 

-- 
Barry Margolin, barmar@alum.mit.edu
Arlington, MA
*** PLEASE post questions in newsgroups, not directly to me ***

Relevant Pages

  • Re: What is best protocol which gives more speed then TCP for low bandwidth networks
    ... Whatever protocol is nested inside is just data, ... However, if the O/P is going to reduce the size of the packet, this is ... than header compression, or than tinkering with the header at all. ...
    (comp.os.linux.networking)
  • Re: ppp problem with isp
    ... > the compression protocols ... Adding the pppd option noipdefault would prevent the local ... > repeated NAK thing in the middle and also that Protocol reject thing (could ... the pppd asyncmap option requests. ...
    (comp.os.linux.networking)
  • Re: PPP with PPPD over GSM link
    ... > I didnt know that the VJ compression can compress from 40 Bytes (TCP+IP ... > the header of the TCP/IP and I need to compress my data too, ... if you're insistent on designing a protocol to overcome temporary ... > If I don't use PPP, and I want to autenticate and construct frames, I ...
    (comp.protocols.ppp)
  • Re: ssh extensions
    ... variable block size, what's the block size and the key size of that? ... The PuTTY source code says it's equivalent to aes256-cbc, ... What is delayed compression? ... There's a nasty flaw in the protocol definition, ...
    (comp.security.ssh)
  • Re: What is best protocol which gives more speed then TCP for low bandwidth networks
    ... , Ken Roberts wrote: ... That's not really tunneling any more than carrying a common protocol like ... However, if the O/P is going to reduce the size of the packet, this is ... Header compression is about the only thing you can do OTHER THAN the ...
    (comp.os.linux.networking)