- From: John Franks <john@math.nwu.edu>
- Date: Thu, 29 Jan 1998 15:54:53 -0600 (CST)
- To: Adrien de Croy <adrien@qbik.com>
- Cc: http-wg@cuckoo.hpl.hp.com
On Fri, 30 Jan 1998, Adrien de Croy wrote:
>
> However, reflecting more on that issue, the chances of a client requiring
> multiple created entitities (i.e those where the server cannot know a priori
> the size) in a single connection is rather low, at least at the moment.
> Multiple normal requests per connection would still be possible, and
> unaffected by this proposal. So, overall, the performance gains by allowing
> for maintained connections in this scenario may be outweighed by the data
> overhead in chunking.
>
Connections with transfers of multiple entities of unknown size will be
very common with HTTP/1.1. They would be very common today if there
were widely deployed HTTP/1.1 clients. Remember, any document with
server side includes (e.g. a counter) can be considered of unknown size.
It is much easier to chunk than to calculate the length before any
data is sent so it can be put in a Content-length header.
Chunking has very low overhead. From the point of view of efficiency
there would be no problem for clients or servers if ALL transactions
were required to be chunked. There is not much performance difference
between
Content-length: 123456
<123456 bytes of data>
and
Content-encoding: chunked
123456
<123456 bytes of data>
0
John Franks
john@math.nwu.edu
Received on Thursday, 29 January 1998 13:56:40 UTC