Thanks Phil. For now, others can proceed with their homework so we're good. I'll try to "help with the wording" as we incorporate the other contributions.

These are shared milestones/deliverables for SNL as well.

--Lee

On Aug 1, 2012, at 7:55 AM, Phil Carns wrote:

Thanks Lee.  Notes on your two comments below:

* COMPONENT CONVERGENCE: on a system object storage model to simplify  
  implementation of higher level client abstractions
     ::Lee:: A little more explanation? Is this a deliverable we both would
have, like the previous two.

You may need to help with the wording on this one.   I was trying to capture our effort to build a shared object storage API for Sirocco and Triton.  Not the local storage interface, but the client side interface with collections/objects/forks/records and so forth.  If I understand correctly your team is pretty far along on this.  On our side, we have a reference implementation that uses local storage to help exercise the API and semantics, but  we haven't implemented it atop Triton yet.

* COMPONENT CONVERGENCE: shared mechanisms for performing RPC operations in the
  storage system
    ::Lee:: Confused again. More explanation?

The wording needs some help here as well.  I don't know that Sirocco and Triton can share a complete RPC mechanism, but it seems like we can share some aspects of it at least, like unexpected messages queues, encoding, etc.  This is something we are hoping to look at in the next version of libssi.

-Phil

On 07/31/2012 04:39 PM, Ward, Lee wrote:
With Sandia modifications. Many of Phil's "CONVERGENCE" items are flagged as being due by both SNL and ANL.

--Lee


On Jul 31, 2012, at 12:52 PM, Phil Carns wrote:

> Hi all,
>
> Here is a list of milestones for our work.  We reserve the right to revise it a little as the week goes on, but it should be mostly on track :)
>
> I flagged some bullets as "COMPONENT CONVERGENCE" or "ARCHITECTURE CONVERGENCE".  The former refers to points where we will integrate low level components (network, local storage, etc.) that we've reached a consensus on.  The latter refers to major storage architectural downselect decisions.  Those architectural topics were sprinkled throughout the original timeline, but here we've pushed them all back to years 3-5.  Lee, you should probably look at those in particular and see if they make sense in relation to your schedule.  They might need to move around some.
>
> Remember everyone that we have to turn in this milestone information to our sponsors no later than Friday.
>
> thanks,
> -Phil
> <asg-milestones-anl.txt>_______________________________________________
> ASG mailing list
> ASG@lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/asg