Re: [Fwd: Urgent as Framing Hint?]

["Randall R. Stewart" <randall@stewart.chicago.il.us>]

Matt:

Very interesting paper. There are many things that are
very close to what we did in SCTP :) (I know you did
not want to hear that but its true). I am not sure how
this could be applied to TCP unless you just always
broke your iSCSI data into very small pieces. But you
still come back to the problem that you must make
sure that a chunk begins on a TCP Packet boundary.. which
is what I don't think we can do...

Or am I missing something ??


R


Matt Wakeley wrote:
> 
> 
> 
>   ------------------------------------------------------------------------
> 
> Subject: Re: Urgent as Framing Hint?
> Date: Fri, 1 Dec 2000 09:53:41 -0500
> From: Lyman Chapin <lyman@bbn.com>
> To: "Jonathan M. Smith" <jms@central.cis.upenn.edu>
> CC: Matt Wakeley <matt_wakeley@agilent.com>,
>      Douglas Otis <dotis@sanlight.net>, end2end-interest@ISI.EDU
> References: <200012011339.eB1Dd3W24152@central.cis.upenn.edu>
> 
> At 8:39 AM -0500 12/1/00, Jonathan M. Smith wrote:
> >Hey folks. Just to provide a little "blast from the past", David Feldmeier,
> >then at Bellcore, proposed an idea called "chunks" for TP++, a protocol
> >design effort to address 1Gbps+ networking, that did exactly this. There
> >was a SIGCOMM paper around 1993 or so - I don't have the ref handy as I'm
> >typing this. There was a lot of nice thinking in the TP++ effort, and
> >at least at a high level, much seems to apply to iSCSI.
> >                                                               -JMS
> 
> Feldmeier, David C., "A data labelling technique for high-performance
> protocol processing and its consequences," Proceedings of the ACM
> SIGCOMM Conference on Communications Architectures, Protocols, and
> Applications, September 13-17, 1993, pages 170-181.
> 
> It's at http://www.acm.org/pubs/contents/proceedings/comm/166237/.
> 
> Abstract:
> 
> "Reordering and reassembly of data before processing can reduce
> communication system performance as seen by the application. We
> examine a method of explicitly labelling blocks of data with
> sufficient information to allow process-ing of disordered data. Our
> labelling syntax for data blocks, which we call chunks, is cleaner
> and more general than that of other protocols. We show how chunks can
> be used for efficient fragmentation/reassembly and compare chunks
> with other fragmentation systems. End-to-end error detection is
> complex for chunks or other systems that allow both fragmentation and
> processing of disordered data. We show that it is possible to design
> an end-to-end error detection system that does not compromise chunk
> processing performance. Chunks can take advantage of processing
> techniques such as Integrated Layer Processing and can be used to
> implement concepts such as Application Layer Framing [CLAR 90]."
> 
> - Lyman

-- 
Randall R. Stewart
randall@stewart.chicago.il.us or rrs@cisco.com
815-342-5222 (cell) 815-477-2127 (work)