Re: [sidr] RPKI <-> allocation consistency

[Eric Osterweil <[e-mail]>]

On Aug 28, 2012, at 12:23 AM, Stephen Kent wrote:

> Eric,
> 
> It seems likely that not all RPs will have the same view of INR allocation, e.g., due to differences
> in when RPs fetch data from repositories. This seems analogous to the transient inconsistencies that
> ISPs see today if they use IRR or other INR data sources, for similar reasons.

Maybe yes, and maybe no.  I don't necessarily want to diverge into that topic now (though a separate thread would be ok).  I'm more focused on the pragmatic issues that come up when implementing.  So far, my concerns seem to be pretty specific to this design, but I'm sort of seeking guidance here...

> 
> I think Rob Austein has referred to the repository system as being "loosely consistent," a reasonable target for a large scale, distributed database with entries maintained by a large group of folks.

I admit that my memory/cache consistency background is a little rusty (though I used to be quite dialed in on it), but I don't recall ever coming across that model before.  What are the access/ordering guarantees in ``loose consistency?''  Again, it only matters from a pragmatic perspective.  I am having trouble ensuring that I have the right idea here.

Thanks a lot!

Eric