W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > January to March 2005

hello from Streamload

From: Jim Whitehead <ejw@cs.ucsc.edu>
Date: Thu, 13 Jan 2005 16:21:13 -0800
Message-Id: <200501140021.j0E0LLMV003514@cats-mx3.ucsc.edu>
To: "'WebDAV'" <w3c-dist-auth@w3.org>
Cc: "'Michael Balloni'" <michael@streamload.com>

Accidentally caught by the spam filter. I've added Michael to the accept2
list.

- Jim

-----Original Message-----
From: Michael Balloni [mailto:michael@streamload.com] 
Sent: Thursday, January 13, 2005 3:19 PM
To: dav-announce@lyra.org; w3c-dist-auth@w3.org
Subject: [Moderator Action] hello from Streamload



Hi, I'm Michael Balloni from Streamload (http://www.streamload.com).  We are
interested in providing WebDAV access to our online file storage.

Streamload is no ordinary file system.  Hundreds of servers manage hundreds
of terabytes of customer data, all using pre-.NET Microsoft technology.
 
At this point I'm interested in hearing about different experiences y'all
have had (and services and expertise that can assist us) with implementing
highly trafficked real world WebDAV servers with custom backends.  Python
(Twisted or not) seems like it might be a good way to write a WebDAV server,
but is the performance and stability production grade for 100's of Mbps
bandwidth and big files, many > 1 GB each?  Would the old MP3.com have been
able to use Python to dish out its song downloads, for example?  Apache's
mod_dav with its storage layer API seems up to the job, but how hard is it
to get something working really well vs. IIS/ISAPI?  Has anybody had any
luck (good or bad) doing custom WebDAV programming using IIS/ISAPI?

One design goal is that all file byte processing be performed by core (read
low memory & CPU usage) web server code...I don't want "read bytes from
file/client, write bytes to client/file" loops, I just want to tell
something where to get it from and where to stick it.  The "where to get it
from" will be a server apart from the WebDAV server, so unless WebDAV
clients support redirects for file downloads (?), proxying file bytes
between servers is another requirement.  Our WebDAV software cannot connect
directly to our databases, so communicating with other servers (via HTTP
magic URLs or XMLRPC) to perform all operations is another requirement.
Streamload has a rich metadata model that's compatible with WebDAV...we're
very excited about bringing our storage technology together with WebDAV
because it creates so much value for our customers.

Any anecdotes or tidbits of advice or gut feelings you have, please let me
know.  Oh, and please CC michael@streamload.com with your responses, and
business-related replies should be to me personally so we don't clutter up
the list.
 
Thanks,
Michael Balloni
Streamload Development
Received on Friday, 14 January 2005 00:22:02 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:44:07 GMT