[cap-talk] IOU Protocol

[Tyler Close]

Hi Toby,

Glad to hear you've been looking at the IOU stuff. Hope you find it of use.

I have been toying with how best to implement the IOU protocol in
Joe-E, so haven't put together a public release yet. But since you're
interested now, I've put up a tarball of what I'm using at the moment.
It's at:

http://waterken.sourceforge.net/iou.tar.gz

The implementation doesn't do any eventual operations. It also doesn't
use a reified Sealer/Unsealer pair, instead using Java's package scope
members with Joe-E's Token class.

Hope that helps.

--Tyler

On Fri, Jan 9, 2009 at 2:19 AM, Toby Murray <toby.murray at comlab.ox.ac.uk> wrote:
> Hi all,
>
> I've been looking into the IOU object-capability protocol recently. In
> particular, I've read the description in
>
> "Credit Transfer for Market-Based Infrastructure" by Tyler Close, from
> the Financial Cryptography '06 Conference.
> http://www.springerlink.com/content/3r80k833638m47n3/
>
> This paper has been really useful in understanding the overall structure
> of IOU but it doesn't provide any details about how one normally
> implements it in an object-capability system. The associated HP Tech
> Report [1] that is the full-length version of the paper above provides
> no further guidance on implementation.
>
> The current E distribution includes an IOU implementation as part of the
> DonutLab [2] code in com.hp.donutLab.makeIOUIssuerAuthor [3]. This
> implementation seems to be cluttered, however, with donutLab extras
> (like lamportSlots and other such things beyond my comprehension) that I
> presume are not a necessary part of an object-capability IOU
> implementation.
>
> Is there a "reference" implementation for IOU anywhere? Is there e.g. a
> Joe-E implementation, which would presumably be simpler than the
> DonutLab one in E? I've not found one looking through the current
> Waterken server distribution.
>
> Any pointers or descriptions along the lines of "here is the general
> idea about how to implement this protocol using object-capability
> primitives X, Y and Z (e.g. Sealer-Unsealers or whatever)" would be very
> much appreciated.
>
> Cheers heaps
>
> Toby
>
>
> [1] http://www.hpl.hp.com/techreports/2006/HPL-2006-5.html
>
> [2] http://www.erights.org/smart-contracts/donut-lab/
>
> [3]
> svn://svn.synchrona.org/erights/e/trunk/src/esrc/com/hp/donutLab/makeIOUIssuerAuthor.emaker
>
>
> _______________________________________________
> cap-talk mailing list
> cap-talk at mail.eros-os.org
> http://www.eros-os.org/mailman/listinfo/cap-talk
>