RE: Advantages of persistent connections

> -----Original Message-----
> From: J.P. Martin-Flatin [mailto:martin-flatin@epfl.ch]
> Sent: Sunday, June 14, 1998 4:42 PM
> To: http-wg@cuckoo.hpl.hp.com
> Cc: martin-flatin@epfl.ch
> Subject: Advantages of persistent connections
> 
> 
> Regarding the advantages of persistent connections listed in section 
> 8.1.1, under heading "Persistent HTTP connections have a number of 
> advantages", one is missing (latency), and one is not always 
> true (memory 
> savings). I therefore suggest 2 changes.
> 
While it may not be explicitly stated, I think the real latency improvement
is due to #2.  Indirect network health improvement by use of persistent
connections will lead to many things like reduced congestion and therefore
less latency, but from the perspective of the user of a browser, for
example,
it is the avoidance of the re-establishment of connections which contributes
to the reduction of latency.

So, from the server side, the big win is the CPU, memory and system call
reduction
which allows it to scale better.  From the client side, it is the latency
reduction
by not having to do a 3way handshake for every request which makes it a
faster
user experience.

If we are to specifically highlight latency reduction in the draft, I think
it should be described in terms of the connection re-use and 3way handshake 
avoidance, not the general network health example.

> 1) Add a bullet between current 3rd and 4th bullet:
> 
>     .  Latency is reduced because network congestion is reduced, so
>        less retransmissions are needed, so the elapsed transmission
>        time is shorter.
> 
> 2) Replace:
> 
>     .  By opening and closing fewer TCP connections, CPU time 
> is saved,
>        and memory used for TCP protocol control blocks is also saved.
> 
> with:
> 
>     .  By opening and closing fewer TCP connections, CPU time is saved
>        in routers and hosts (clients, servers, proxies, 
> gateways, tunnels,
>        or caches), and memory used for TCP protocol control blocks can
>        be saved in hosts.
> 
>

Received on Sunday, 14 June 1998 17:59:50 UTC