ZFS

[Vincent-58]

Did you read that about ZFS ?

http://www.appleinsider.com/articles/09/10/23/apple_shuts_down_zfs_open_source_project.html

Vincent

[Lubomir Sedlacik]

On Sat, Nov 07, 2009 at 02:01:37PM +0100, vincent wrote:
> Did you read that about ZFS ?
> http://www.appleinsider.com/articles/09/10/23/apple_shuts_down_zfs_open_source_project.html

Yes, we also read this:
 http://mail.opensolaris.org/pipermail/zfs-discuss/2009-October/033125.html


regards,

--
-- Lubomir Sedlacik <salo@{NetBSD,Xtrmntr,silcnet}.org>   --

[Vincent-58]

Le 7 nov. 2009 à 18:20, Lubomir Sedlacik a écrit :

> On Sat, Nov 07, 2009 at 02:01:37PM +0100, vincent wrote:
>> Did you read that about ZFS ?
>> http://www.appleinsider.com/articles/09/10/23/apple_shuts_down_zfs_open_source_project.html
>
> Yes, we also read this:
> http://mail.opensolaris.org/pipermail/zfs-discuss/2009-October/033125.html

Agreed. What I have retained is that there is some legal issue tied to  
some obscure and private patent war. Is there a possibility NetBSD  
port of ZFS be a "colateral" victim of this battle?

[matthew sporleder]

On Sun, Nov 8, 2009 at 5:17 AM, vincent <10.50@...> wrote:

> Le 7 nov. 2009 à 18:20, Lubomir Sedlacik a écrit :
>
>> On Sat, Nov 07, 2009 at 02:01:37PM +0100, vincent wrote:
>>>
>>> Did you read that about ZFS ?
>>>
>>> http://www.appleinsider.com/articles/09/10/23/apple_shuts_down_zfs_open_source_project.html
>>
>> Yes, we also read this:
>> http://mail.opensolaris.org/pipermail/zfs-discuss/2009-October/033125.html
>
> Agreed. What I have retained is that there is some legal issue tied to some
> obscure and private patent war. Is there a possibility NetBSD port of ZFS be
> a "colateral" victim of this battle?
>

My guess is that if you were Wasabi Systems shipping NetBSD/ZFS-based
products, then you would be a probable target for a cease and desist
order pending a licensing agreement with NetApp (of course, that's
certainly not a specific problem anymore).  However, I think the worst
case for NetBSD would be a requirement to remove any sun-sourced,
infringing code.

[Rhialto]

On Sun 08 Nov 2009 at 08:29:14 -0500, matthew sporleder wrote:
> My guess is that if you were Wasabi Systems shipping NetBSD/ZFS-based
> products, then you would be a probable target for a cease and desist
> order pending a licensing agreement with NetApp (of course, that's
> certainly not a specific problem anymore).  However, I think the worst
> case for NetBSD would be a requirement to remove any sun-sourced,
> infringing code.

Given this:
http://mail.opensolaris.org/pipermail/zfs-discuss/2009-October/033134.html
it sounds like there is no such thing. Besides, software patents are not
allowed anyway.

-Olaf.
--
___ Olaf 'Rhialto' Seibert    -- You author it, and I'll reader it.
\X/ rhialto/at/xs4all.nl      -- Cetero censeo "authored" delendum esse.

[Rhialto]

On Sun 08 Nov 2009 at 14:35:21 +0100, Rhialto wrote:
> http://mail.opensolaris.org/pipermail/zfs-discuss/2009-October/033134.html

You can find the mentioned work by googling for "diplomarbeit joerg
schilling"; the main reference is at
http://cdrecord.berlios.de/private/index.html . It is in German.
I haven't read it yet but given that this thesis is about a file system
on WORM devices, and as far as I understand, the "new thing" by NetApp
is all about not overwriting old data, this sounds quite similar on its
face.

-Olaf.
--
___ Olaf 'Rhialto' Seibert    -- You author it, and I'll reader it.
\X/ rhialto/at/xs4all.nl      -- Cetero censeo "authored" delendum esse.

[Vincent-58]

Le 8 nov. 2009 à 14:35, Rhialto a écrit :

> On Sun 08 Nov 2009 at 08:29:14 -0500, matthew sporleder wrote:
>> My guess is that if you were Wasabi Systems shipping NetBSD/ZFS-based
>> products, then you would be a probable target for a cease and desist
>> order pending a licensing agreement with NetApp (of course, that's
>> certainly not a specific problem anymore).  However, I think the  
>> worst
>> case for NetBSD would be a requirement to remove any sun-sourced,
>> infringing code.
>
> Given this:
> http://mail.opensolaris.org/pipermail/zfs-discuss/2009-October/033134.html
> it sounds like there is no such thing. Besides, software patents are  
> not
> allowed anyway.

Well, in Europe may be not, but in the US, read the history of BSD,  
and the birth of BSDLite. GIF also is a good example.

Okay, it can be no more than a voyage back in time to the  
aforementioned episode – so, is there any estimation as to the  
percentage of Sun code in today's NetBSD ZFS port?

Tot ziens :)
Vincent

[Greg A. Woods; Planix, Inc.]

At Sun, 8 Nov 2009 17:51:27 +0100, vincent <10.50@...> wrote:
Subject: Re: ZFS

>
> Le 8 nov. 2009 à 14:35, Rhialto a écrit :
>
> > On Sun 08 Nov 2009 at 08:29:14 -0500, matthew sporleder wrote:
> >> My guess is that if you were Wasabi Systems shipping NetBSD/ZFS-based
> >> products, then you would be a probable target for a cease and desist
> >> order pending a licensing agreement with NetApp (of course, that's
> >> certainly not a specific problem anymore).  However, I think the  
> >> worst
> >> case for NetBSD would be a requirement to remove any sun-sourced,
> >> infringing code.
> >
> > Given this:
> > http://mail.opensolaris.org/pipermail/zfs-discuss/2009-October/033134.html
> > it sounds like there is no such thing. Besides, software patents are  
> > not
> > allowed anyway.
>
> Well, in Europe may be not, but in the US, read the history of BSD,  
> and the birth of BSDLite. GIF also is a good example.

Hopefully not long in the US either, though I'm just blue-sky dreaming

I haven't read the whole article, and I dare not claim to understand
anything about US law, but the leading paragraphs are interesting:

http://tech.yahoo.com/news/ap/20091107/ap_on_hi_te/us_tec_supreme_court_software_patents

   With the technology industry looking on, the Supreme Court on Monday
   will explore what types of inventions should be eligible for a patent
   in a pivotal case that could undermine such legal protections for
   software.

   A ruling that sides with the Patent Office could bar patents on
   processes and methods of doing business, such as online shopping
   techniques, medical diagnostic tests and procedures for executing
   trades on Wall Street. And it might even undercut patents on
   software.


--
                                                Greg A. Woods
                                                Planix, Inc.

<woods@...>       [phone]        http://www.planix.com/

[Ty Sarna-2]

On Nov 8, 2009, at 11:51 AM, vincent wrote:

> Le 8 nov. 2009 à 14:35, Rhialto a écrit :
>> Given this:
>> http://mail.opensolaris.org/pipermail/zfs-discuss/2009-October/033134.html
>> it sounds like there is no such thing. Besides, software patents  
>> are not
>> allowed anyway.
>
> Well, in Europe may be not, but in the US, read the history of BSD,  
> and the birth of BSDLite. GIF also is a good example.
>
> Okay, it can be no more than a voyage back in time to the  
> aforementioned episode – so, is there any estimation as to the  
> percentage of Sun code in today's NetBSD ZFS port?

What does that have to do with anything? It could be 100% NetBSD code  
and that wouldn't change anything as far as patents.

[Vincent-58]

Le 8 nov. 2009 à 18:33, Ty Sarna a écrit :

>> Okay, it can be no more than a voyage back in time to the  
>> aforementioned episode – so, is there any estimation as to the  
>> percentage of Sun code in today's NetBSD ZFS port?
>
> What does that have to do with anything? It could be 100% NetBSD  
> code and that wouldn't change anything as far as patents.

I am not sure about this. I think you can't patent a concept, but you  
could patent an implementation. For example, in the electronics field,  
you maybe can patent a type of high-quality oscillator, but you cannot  
claim any property on a different implementation achieving the same  
goal. If you can code from scratch a FS with the same functionalities  
as ZFS, I don't figure out how you could be liable to patent suing.

Vincent

[Adam Hamsik-2]

On Nov,Sunday 8 2009, at 5:51 PM, vincent wrote:

> Le 8 nov. 2009 à 14:35, Rhialto a écrit :
>
>> On Sun 08 Nov 2009 at 08:29:14 -0500, matthew sporleder wrote:
>>> My guess is that if you were Wasabi Systems shipping NetBSD/ZFS-
>>> based
>>> products, then you would be a probable target for a cease and desist
>>> order pending a licensing agreement with NetApp (of course, that's
>>> certainly not a specific problem anymore).  However, I think the  
>>> worst
>>> case for NetBSD would be a requirement to remove any sun-sourced,
>>> infringing code.
>>
>> Given this:
>> http://mail.opensolaris.org/pipermail/zfs-discuss/2009-October/033134.html
>> it sounds like there is no such thing. Besides, software patents  
>> are not
>> allowed anyway.
>
> Well, in Europe may be not, but in the US, read the history of BSD,  
> and the birth of BSDLite. GIF also is a good example.
>
> Okay, it can be no more than a voyage back in time to the  
> aforementioned episode – so, is there any estimation as to the  
> percentage of Sun code in today's NetBSD ZFS port?

We want to change as least code as possible in ZFS. Because it makes  
later updates much easier for us.

Regards

Adam.

[Quentin Garnier]

On Sun, Nov 08, 2009 at 06:42:27PM +0100, vincent wrote:

> Le 8 nov. 2009 à 18:33, Ty Sarna a écrit :
>
> >>Okay, it can be no more than a voyage back in time to the
> >>aforementioned episode ? so, is there any estimation as to the
> >>percentage of Sun code in today's NetBSD ZFS port?
> >
> >What does that have to do with anything? It could be 100% NetBSD
> >code and that wouldn't change anything as far as patents.
>
> I am not sure about this. I think you can't patent a concept, but
> you could patent an implementation. For example, in the electronics
> field, you maybe can patent a type of high-quality oscillator, but
> you cannot claim any property on a different implementation

The software field is different.  Implementations are protected by
copyright.  Why do you think people tend to oppose software patents?

--
Quentin Garnier - cube@... - cube@...
"See the look on my face from staying too long in one place
[...] every time the morning breaks I know I'm closer to falling"
KT Tunstall, Saving My Face, Drastic Fantastic, 2007.

[Steven Bellovin]

On Nov 8, 2009, at 12:42 PM, vincent wrote:

> Le 8 nov. 2009 à 18:33, Ty Sarna a écrit :
>
>>> Okay, it can be no more than a voyage back in time to the  
>>> aforementioned episode – so, is there any estimation as to the  
>>> percentage of Sun code in today's NetBSD ZFS port?
>>
>> What does that have to do with anything? It could be 100% NetBSD  
>> code and that wouldn't change anything as far as patents.
>
> I am not sure about this. I think you can't patent a concept, but  
> you could patent an implementation. For example, in the electronics  
> field, you maybe can patent a type of high-quality oscillator, but  
> you cannot claim any property on a different implementation  
> achieving the same goal. If you can code from scratch a FS with the  
> same functionalities as ZFS, I don't figure out how you could be  
> liable to patent suing.

That's not the way patents work.  What you've described is more like  
copyright, where the particular expression of the concept is  
protected.  With patents, it's the concept itself -- but details matter.

Let's consider the case of the famous (and now-expired) RSA patent: http://www.google.com/patents/about?id=daEsAAAAEBAJ&dq=4,405,829 
  .  Fundamentally -- and I believe this is true throughout the world,  
though what can be patented differs in different jurisdictions -- a  
patent consists of two parts, a description of the invention and  
"claims".  The description is just that: it describes how the  
invention works.  A patent is basically a deal between the government  
and an inventor: the inventor describes the idea in sufficient detail  
that others can replicate it, learn from it, etc.; in return, the  
government grants a limited-term monopoly on the *idea*.  Precisely  
what is protected is described by the claims.  You infringe a patent  
if you build something that contains all elements of any one claim.  
Thus, if you look at claim 1 of the RSA patent, to infringe you'd need  
a system with a communications channel, an encoder that worked  
according to that description, and a decoder that worked according to  
its description.  If you built a system that had no communications  
channel, you wouldn't infringe claim 1.  However, you might infringe  
claim 37, which doesn't speak of a channel.

There's a lot more I won't go into, like the "doctine of  
equivalents".  The net, though, is to the extent that the patent is  
valid in jurisdictions of interest, you can't evade it by changing  
code; you probably have to change functionality.

                --Steve Bellovin, http://www.cs.columbia.edu/~smb

[Vincent-58]

Le 8 nov. 2009 à 22:08, Steven Bellovin a écrit :

> There's a lot more I won't go into, like the "doctine of  
> equivalents".  The net, though, is to the extent that the patent is  
> valid in jurisdictions of interest, you can't evade it by changing  
> code; you probably have to change functionality.

Ok. You are probably right, my knowledge of patent limits itself to  
the French root of the word (in French, patente means licen[cs]e) ! :)

I was just wondering if, given the current environment, it was really  
worth carrying on work on ZFS rather than waiting for the whole affair  
to subside.

Cheers
Vincent

[Matthew Mondor]

On Mon, 9 Nov 2009 09:50:52 +0100
vincent <10.50@...> wrote:

> I was just wondering if, given the current environment, it was really  
> worth carrying on work on ZFS rather than waiting for the whole affair  
> to subside.

If ZFS could be built as an optional module, would it mitigate the
issue?  I.e. delegating risks to every entity deciding to use it, but
default distributions missing the functionality?

Thanks,
--
Matt

[matthew sporleder]

On Mon, Nov 9, 2009 at 3:50 AM, vincent <10.50@...> wrote:

> Le 8 nov. 2009 à 22:08, Steven Bellovin a écrit :
>
>> There's a lot more I won't go into, like the "doctine of equivalents".
>>  The net, though, is to the extent that the patent is valid in jurisdictions
>> of interest, you can't evade it by changing code; you probably have to
>> change functionality.
>
> Ok. You are probably right, my knowledge of patent limits itself to the
> French root of the word (in French, patente means licen[cs]e) ! :)
>
> I was just wondering if, given the current environment, it was really worth
> carrying on work on ZFS rather than waiting for the whole affair to subside.

Don't fall victim to these innovation-stifling techniques until you're
ordered to do so.  I think netbsd is gaining from the zfs porting
effort regardless of the eventual outcome.

[Steven Bellovin]

On Nov 9, 2009, at 4:55 AM, Matthew Mondor wrote:

> On Mon, 9 Nov 2009 09:50:52 +0100
> vincent <10.50@...> wrote:
>
>> I was just wondering if, given the current environment, it was really
>> worth carrying on work on ZFS rather than waiting for the whole  
>> affair
>> to subside.
>
> If ZFS could be built as an optional module, would it mitigate the
> issue?  I.e. delegating risks to every entity deciding to use it, but
> default distributions missing the functionality?
>

Possibly, possibly not.  Patents govern the right to make, sell,  
import, or use the covered invention, though generally the patent  
owners go after the creator of the infringing device.  If it isn't  
patented in, say, Europe, and all development were done there, no  
offenses would be committed except by someone who imported the code  
into (for example) the US.

Also note that for "willful infringement" -- you knew something was  
patented but went ahead anyway -- the damages are tripled.

                --Steve Bellovin, http://www.cs.columbia.edu/~smb