[9fans] look, ma, linux's discovered fossil + venti

[9fans] look, ma, linux's discovered fossil + venti

[erik quanstrom]

they have funny names for it all.
and they're missing the worm part.

http://lwn.net/Articles/305740/

- erik

Re: [9fans] look, ma, linux's discovered fossil + venti

[Brian L. Stuart]

> they have funny names for it all.
> and they're missing the worm part.
>
> http://lwn.net/Articles/305740/

There's a reason for the funny names.  If
they don't include the word 'object' enough
times, they'll lose their Tier 1 status in
the buzzword hierarchy.

BLS

Re: [9fans] look, ma, linux's discovered fossil + venti

[ron minnich]

On Thu, Nov 13, 2008 at 5:33 AM, erik quanstrom <quanstro@coraid.com> wrote:
> they have funny names for it all.
> and they're missing the worm part.
>
> http://lwn.net/Articles/305740/
>
> - erik
>
>

I found this amusing: 'Every operation on an object must be
accompanied by a "capability," a cryptographically-signed ticket which
names the object and the access rights possessed by the owner of the
capability. In the absence of a suitable capability, the drive will
deny access. ' -- oh, and the objects are '2 64-bit numbers'.

no assumptions about the network or protocol, eh?

ron

Re: [9fans] look, ma, linux's discovered fossil + venti

[Enrico Weigelt]

* erik quanstrom <quanstro@coraid.com> wrote:
> they have funny names for it all.
> and they're missing the worm part.
>
> http://lwn.net/Articles/305740/

*lol*
Food old 1541 coming back (plus SCSI + a bit crypto stuff) ?

Wait! 1541 already had filenames, ah, and there was an updated
firmware going around which supported directory hierachies ;-o
(hmm, maybe I should grab it out again and code an 9P firmware ;))

WTH didn't they just take a plain filesystem and export it
via certain remote-fs protocols, eg. 9P ? Probably too obvious ;-P


cu
--
----------------------------------------------------------------------
 Enrico Weigelt, metux IT service -- http://www.metux.de/

 cellphone: +49 174 7066481   email: info@metux.de   skype: nekrad666
----------------------------------------------------------------------
 Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme
----------------------------------------------------------------------

Re: [9fans] look, ma, linux's discovered fossil + venti

[Charles Forsyth]

[-- Attachment #1: Type: text/plain, Size: 108 bytes --]

i also don't really trust the groups doing it to produce a reasonable file
or high-level storage subsystem.

[-- Attachment #2: Type: message/rfc822, Size: 3272 bytes --]

From: Enrico Weigelt <weigelt@metux.de>
To: 9fans@9fans.net
Subject: Re: [9fans] look, ma, linux's discovered fossil + venti
Date: Fri, 14 Nov 2008 13:08:14 +0100
Message-ID: <20081114120813.GB3071@nibiru.local>

* erik quanstrom <quanstro@coraid.com> wrote:
> they have funny names for it all.
> and they're missing the worm part.
>
> http://lwn.net/Articles/305740/

*lol*
Food old 1541 coming back (plus SCSI + a bit crypto stuff) ?

Wait! 1541 already had filenames, ah, and there was an updated
firmware going around which supported directory hierachies ;-o
(hmm, maybe I should grab it out again and code an 9P firmware ;))

WTH didn't they just take a plain filesystem and export it
via certain remote-fs protocols, eg. 9P ? Probably too obvious ;-P


cu
--
----------------------------------------------------------------------
 Enrico Weigelt, metux IT service -- http://www.metux.de/

 cellphone: +49 174 7066481   email: info@metux.de   skype: nekrad666
----------------------------------------------------------------------
 Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme
----------------------------------------------------------------------

Re: [9fans] look, ma, linux's discovered fossil + venti

[ron minnich]

The OSD stuff has been going on for almost 10 years or more. Started
in 1996 or so at CMU.
Original design was to run the stack in the disk drive. This explains
a lot. I could not get them to look at 9p either.

ron

Re: [9fans] look, ma, linux's discovered fossil + venti

[erik quanstrom]

> I found this amusing: 'Every operation on an object must be
> accompanied by a "capability," a cryptographically-signed ticket which
> names the object and the access rights possessed by the owner of the
> capability. In the absence of a suitable capability, the drive will
> deny access. ' -- oh, and the objects are '2 64-bit numbers'.
>
> no assumptions about the network or protocol, eh?

i missed that part.  (i also missed where they talk about
who does the crypto and what protocols they will use.)
this will make 9p latency look downright trivial.

perhaps my brain went into self-protect mode.

- erik