RE: [sem-grd] Agents and the Grid?

["Pouchard, Line Catherine" <pouchardlc@ornl.gov>]

 Dave and all:

I'd like to add and summarize some points I made earlier to Dave
(apologies for duplicates) without ccing the list: 

*I agree with Carole that fewer groups are better than more, and the
knowledge-based effort in the agents community provides a good starting
point for bridging the two communuties.  For a start, some FIPA
performatives contains an attribute for ontologies, for content of the
message.

*Including the agents topic in the Semantic Grid under the focus of
negotiation protocols within the Grid (i.e. semantics in the Grid) and
between the content of services should avoid the pitfall that agenst
subsume or submerge the semantic grid.  

*It may be too early for a general merging with agents and peer-to-peer.
The semantic grid idea is slowly becoming of interest to which are
usually slower to react than univerisities.   The workshop at Ccgrid
2005 may create some momentum for the Semantic Grid effort within DOE.
One important aspect is that the agency is focused on scientific
applications with big instruments and computers, and they start
realizing that they have massive data problems looming.    So for
instance, I would not change the name and emphasize negotiations rather
than agents in general.

*The Semantic Grid like the Grid may be looking for a "killer app" in
the future, as well as recognition in many domains.  To re-emphasize, my
pint is not what agents/Semantic Grid can do for the Grid or what the
Grid can do for semantics, but what a leveraging and integration of the
two can do for solving real-world problems (scientific or otherwise).

*Based on such an approach it may be easier to get a re-worked charter
with additional items accepted in GGF (although I have no experience in
this :)

Line

> -----Original Message-----
> From: owner-sem-grd@gridforum.org 
> [mailto:owner-sem-grd@gridforum.org] On Behalf Of David De Roure
> Sent: Monday, November 08, 2004 7:46 AM
> To: Kashif Iqbal
> Cc: sem-grd@gridforum.org
> Subject: RE: [sem-grd] Agents and the Grid?
> 
> Kashif
> 
> I think the really positive thing here is that, in this 
> discussion, people have been pointing out what agents can do 
> for the grid.
> 
> Sometimes when a community comes to GGF they are asking 
> primarily what the grid and GGF can do for them.  Since GGF 
> is paid for by its members to do things for the grid, this 
> latter approach is problematic :)
> 
> So I think we're in a really good position for a successful 
> engagement.
> 
> Of course, some of the motivation may well be what the grid 
> can do for agents!  We've found with Semantic Grid that the 
> Grid provides an enthusiastic community of people who want to 
> do real things, and this can be very beneficial.  
> Occasionally Grid computation is useful too!
> 
> While we're talking about bits and pieces of the puzzle...
> 
> OGSA is a key activity in GGF and so the relationship between 
> agents and OGSA is quite important.  Similarly, one can argue 
> for OWL-S and WSMO being considered in this context.  At the 
> moment I guess these fall under Semantic Grid, and I've been 
> expecting them to become explicit in our charter review - 
> it's our repeated emphais on semantics *in* the Grid as well 
> as on the Grid.
> 
> Lots of pieces :)
> 
> Thanks
> 
> -- Dave
> 
> On Mon, 8 Nov 2004, Kashif Iqbal wrote:
> 
> > David,
> > 
> > I completely agree with you.
> > 
> > Because whenever we talk about integrating agents and grid, 
> usually we 
> > end up in bridging the communication gap between the two 
> technologies. 
> > Many existing efforts in integrating agents and grid 
> experienced the 
> > same. I also did the same when we proposed "Autonomous 
> Distributed Service System".
> > 
> > The other approach taken by some of the people was to build grids 
> > using MAS (multi agent systems) but this is something which Grid 
> > community can't afford now.
> > 
> > In some of the points for charter I mentioned that certain 
> > characteristics of agents like autonomy, negotiation, interaction 
> > protocols, and communication languages can certainly help 
> grid world.
> > 
> > So, I will also advocate the point made by Dave.
> > 
> > Cheers,
> > Kashif
> > 
> > -----Original Message-----
> > From: owner-sem-grd@gridforum.org 
> [mailto:owner-sem-grd@gridforum.org] 
> > On Behalf Of David De Roure
> > Sent: Monday, November 08, 2004 11:36 AM
> > To: Jon MacLaren
> > Cc: 'Semantic Grid Research Group'
> > Subject: RE: [sem-grd] Agents and the Grid?
> > 
> > John
> > 
> > My view is that we perhaps shoudn't be talking of "agents" 
> but rather 
> > we need to tease out more exactly which bits of agents 
> benefit which 
> > bits of Grid.  For example, one aspect that seems to me to be much 
> > needed at this time is the expertise on negotiation - this need is 
> > evidenced by the WS-Agreement activity.  This could, for example, 
> > result in an activity called negotiation rather than agents, and it 
> > might sit in a different area to other aspects of agency.  Another 
> > topic might be self-management a.k.a. autonomic.
> > 
> > Some of these topics sit naturally under the Architecture 
> area, which 
> > is also where Semantic grid sits currently - and therefore 
> they look 
> > to fit in, under or alongside Semantic Grid (hence this 
> discussion).  
> > But the GGF Steering Group is meeting today to discuss area 
> > reorganisation, after which I should have a better idea 
> about where things could land.
> > 
> > Having said that, regardless of area structure, it makes sense to 
> > proceed with drafting the charter and planning a BoF.
> > 
> > I'll report back later today!
> > 
> > Thanks
> > 
> > -- Dave
> > 
> > On Mon, 8 Nov 2004, Jon MacLaren wrote:
> > 
> > > I think that the most important facing those of us interested in 
> > > Agents technology is how to bring experience from the Agents 
> > > community into the
> > GGF
> > > (something which should benefit both communities).
> > > 
> > > I agree with Carole that having them partitioned into their own 
> > > group probably doesn't make too much sense.  Unless, 
> perhaps, it was 
> > > a RG whose main goal was to help figure out how best to direct 
> > > knowledge (and effort) from the Agent people into existing GGF 
> > > groups (or into fitting Agent technologies into the Grid "big 
> > > picture", as it were).  I know of at least one WG which certainly 
> > > would've benefited from more influence from Agent 
> experts.  Is this 
> > > a valid goal for a RG?  (I guess that there will be new 
> things too, 
> > > requiring new WGs, unless we've thought of everything
> > elsewhere
> > > in GGF!)
> > > 
> > > But discussing this further is very difficult without 
> having a draft 
> > > charter, or without at least knowing roughly what is intended.
> > > 
> > > Apologies Dave, I know you were canvassing for views in the other
> > direction!
> > > 
> > > Cheers,
> > > 
> > > Jon MacLaren.
> > > 
> > > > -----Original Message-----
> > > > From: owner-sem-grd@gridforum.org 
> > > > [mailto:owner-sem-grd@gridforum.org]
> > On
> > > > Behalf Of David De Roure
> > > > Sent: 02 November 2004 16:00
> > > > To: Carole Goble
> > > > Cc: Semantic Grid Research Group
> > > > Subject: Re: [sem-grd] Agents and the Grid?
> > > > 
> > > > All
> > > > 
> > > > The agents charter is now being formulated, and there's a 
> > > > discussion
> > about
> > > > whether it should be a research group or a working group.  If a 
> > > > working group, it could be spawned by Semantic Grid.
> > > > 
> > > > There's also a suggestion of having a Semantic Grid 
> workshop at a 
> > > > later GGF themed on agents.
> > > > 
> > > > All of these suggestions seem to be in line with keeping these
> > activities
> > > > together, as Carole underlined in her recent message.
> > > > 
> > > > Also, we are due to review the Semantic grid charter, which is 
> > > > another opportunity to bring things together.
> > > > 
> > > > So I'll turn the question round - would anyone like to argue to 
> > > > keep Semantic Grid and agents as very *distinct* activities at 
> > > > research group level?
> > > > 
> > > > Thanks
> > > > 
> > > > -- Dave
> > > > 
> > > > On Tue, 2 Nov 2004, Carole Goble wrote:
> > > > 
> > > > > Dave
> > > > >
> > > > > In my recent ECAI keynote on the Semantic Grid I had 
> a section 
> > > > > on Agents. This was coming from the point if view 
> that the agent
> > community
> > > > > and the knowledge rep. community were often the same.
> > > > > If you have another activity with agents and the grid 
> separate 
> > > > > to the Semantic Grid RG, isn't that just dividing 
> your already 
> > > > > small base? I would welcome more active help and 
> effort in FEWER 
> > > > > RG andWG, not yet more. I would like to see the agents people 
> > > > > involved in the SEM-GRD, scheduling, OGSA etc 
> activities rather than in a ghetto.
> > > > >
> > > > > Carole
> > > > >
> > > > > >Some GGF business.
> > > > > >
> > > > > >There will be a discussion about GGF area 
> reorganisation at a 
> > > > > >forthcoming GGF Steering Group meeting (at SC2004).
> > > > > >
> > > > > >In fact this doesn't affect the Semantic Grid 
> Research Group much.
> > > > > >We're currently in the architecture area, we could well have 
> > > > > >been in data, and it doesn't make a huge amount of 
> difference 
> > > > > >anyway(!)
> > > > > >
> > > > > >Meanwhile I have been advocating the introduction of an 
> > > > > >activity in the area of agents and the grid, and this has 
> > > > > >attracted some interest from parts of the agents community 
> > > > > >where there are some existing grid activities.  It needs a 
> > > > > >champion and I think this may happen.
> > > > > >
> > > > > >So, through reorganisation of the areas, this may 
> will end up 
> > > > > >in an area alongside Semantic Grid (which may be a 
> good thing 
> > > > > >when it comes to bridge-building).
> > > > > >
> > > > > >Anyway, I'd welcome any comments from the community 
> on this (e.g.
> > > > > >do you think GGF should have an agents and the grid 
> activity?  
> > > > > >Do you think it should sit next to Semantic Grid?), 
> so that I 
> > > > > >can go into the discussions able to represent your views.
> > > > > >
> > > > > >Thanks
> > > > > >
> > > > > >-- Dave
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > 
> > > 
> > > 
> > 
> > 
> > 
> 
>