Here is my understanding of what we are seeing in testing. (Peter, please
correct me if I'm wrong).
IPP is a request-response protocol that sits on top of HTTP (a
request-response protocol) that sits on top of TCP (a connection protocol).
TCP provides the end-to-end message integrity including packet retries for
lost or mis-ordered transmissions.
HTTP 1.0 uses a single TCP connection for each request-response
transaction. In many cases, this is because the only way to signal the
end-of-data is by closing the connection. (HTTP 1.0 is described in the
Informational RFC 1945)
HTTP 1.1 provides persistent connections (connections that stay intact for
multiple request-response transactions for the same client-server). The
technique for signalling end-of-data for objects whose length may not be
known at the beginning of the transmission is chunked encoding. In HTTP
1.1, you may still signal end-of-data by closing the connection, but this
is not recommended, since you will then lose the benefits of the persistent
connection. (HTTP 1.1 is currently described in the Standards Track RFC
2068, but there will be a new RFC shortly reflecting implementation
experience)
At 12:31 PM 4/16/98 , Carl-Uno Manros wrote:
>Randy et al.,
>
>Here is my take on the use of chunking for IPP:
>
>1) Each HTTP 1.1 implementation is mandated to support chunking.
Yes, but all transactions do not need to use chunked encoding.
>
>2) All POSTs are initiated by the client, which I assume also means that
>the client determines whether to use chunking for a particular HTTP request.
>
All transactions may use chunked encoding in either direction. The
requesting client determines whether to use it for POST or PUT requests,
and the responding server determines whether to use it for responses.
>3) The only situation where chunking is absolutely necessary for IPP, is
>the case where the client does not know the document length when it starts
>sending the document.
Strictly speaking, this is not a requirement, in that the client could
close the connection to signal end-of-data.
>
>4) With the exception of the situation in 3) the client can always decide
>not to use chunking, but send even long documents in one request. The
>downside of doing that is that if something goes wrong on the lower
>protocol layers, the client has to start over and send the whole document
>again from the beginning. The main advantage of using chunking is that if
>you get an error on the lower layers, you only have to resend the latest
>chunk, not the whole document.
This is not correct. Lower layer transmission errors will be handled at
the TCP layer regardless of whether chunked encoding is used.
>
>5) Again, with the exception of the case in 3), you can in practise
>actually use HTTP 1.0, which does not support chunking, for most of your
>IPP transfers.
>
Our current implementation of IPP transaction support requires that HTTP
1.1 support be turned on. In practice, Peter overrode this constraint and
is successfully running IPP over HTTP 1.0 regardless of whether job size is
known at the beginning.
Bob
----------------------------------------
Bob Van Andel
Allegro Software Development Corporation
43 Waite Road
Boxborough, MA 01719
(978) 266-1375
(978) 266-2839 fax
Information on the RomPager embedded web server toolkit is at
<http://www.allegrosoft.com/>