-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
Henrik Nordstrom wrote:
| On sön, 2008-06-01 at 10:43 -0400, Matt Benjamin wrote:
|> The most compelling reason to use SCTP for HTTP delivery would seem to
|> be the potential to improve throughput via SCTP multi-streaming.
|
| In TCP that is done by having multiple concurrent TCP connections, which
| is pretty close to the same.
Yes.
|
| The biggest mismatch for HTTP over SCTP is that there is no per-stream
| "ABORT" or "End-of-file" conditions in SCTP. While SCTP do provide
| message boundaries it's not trivial to utilize this to map to HTTP
| messages in the general case.
|
That is unfortunate. I wasn't aware of this.
|
| It's a pity SCTP got specified with these quite innocent limitations, as
| it somewhat limits it's applicability to HTTP.
Randall Stewart and students seem to be moving it in new directions (eg,
Concurrent Multipath Transfer), so maybe we'll get past it.
|
| Regards
| Henrik
- --
Matt Benjamin
The Linux Box
206 South Fifth Ave. Suite 150
Ann Arbor, MI 48104
tel. 734-761-4689
fax. 734-769-8938
cel. 734-216-5309
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFIQz0aJiSUUSaRdSURCCNfAJ4+DZdw7EXwXr5hFg66c6CzV/a3qQCgmSDf
5KnEInr/Mfaw55NE68JTBZY=
=hLQv
-----END PGP SIGNATURE-----
Received on Mon Jun 02 2008 - 00:22:39 MDT
This archive was generated by hypermail 2.2.0 : Mon Jun 02 2008 - 12:00:03 MDT