1. Collaboration in Grid Systems Activities where collaboration important Functional Structure of a Collaborative System Collaborative Visualization Grid Event and Message Service

2. Grid Collaboration Opportunities zCollaborative Portals: (Portal is a Grid Computing or Training or ? Environment) for Computing or Education zEducation Grid: Distance Education is implemented as collaborative Service (between teacher and student) on top of a education/training portal zComputational Grid: Typical examples yCollaborative Visualization yCollaborative Consulting (help desk) yCollaborative selection of parameters (possibly dynamically) for job submission (Collaborative computational steering)

3. Collaborative Grid Collab Server Data base Peer to Peer “Illusion” Among collaborating clients Collaborating Grid with Redundant Server Network (Industry leader WebEx has 120 in 6 centers) for persistence and routing (optimal software multicast) Pu blish Sub scribe

4. What do Collaboration Systems have ? zAudio Video Conferencing (HearMe at low end; Access Grid at high end) zText chat; white board; instant messenger zSecure access; group e-mail; surveys, quizzes z Shared documents or objects including yShared display (framebuffers) and yShared Export (version of shared object in some standard form like HTML SVG PDF JPEG) yShared Event (original object)

5. Publish/Subscribe Collaboration z The web is full of objects – web pages sitting on web servers – and these support asynchronous collaboration yYou post a web page and I later look at it in my own time z Replacing web document by a “CGI script” or servlet (web interface to program, database etc.) gives asynchronous general multi-tier object sharing z This is Publish/Subscribe mechanism yIf add some mechanism (channel or automatic email or word of mouth) to tell viewing client when new information is posted yWe build on JMS (Java Message Service) as Industry standard for publish/subscribe systems z Synchronous Collaboration provides “real-time” notification and automatic update of changed objects yJMS seems fast enough to do real-time as well asynchronously

6. Collaborative Computing z Shared Display for consultants (remote control for debugging). Basic help desk mode z Visualization done by shared display or shared export (Java3D SVG GIF) yShared SVG for Flash, Adobe, PowerPoint, Gateway Whiteboard and ? yAnnotations (pointers) can be shared for scalable export formats as can separately zoom and still position overlays z Job Submission done by shared Java Server Pages in Gateway and shared web pages for HTML based interfaces

7. Collaborative Visualization z Consider a computer program (object above) and then its output and input wend their way through multiple filters (tiers) until they are finally rendered on some sort of device: CAVE through PDA. z One can share “object” at any stage in pipeline Object Filter Map Transform Broker Event Adapter Input Output Object Filter Map Transform Broker Event Adapter Output Input A B Shared Display C W3C Custom Master User B Collaborators A and C Shared export

8. Research Issues: Collaborative Grids z Optimal routing and possibly dynamic server arrangement based on aggregation of topic based subscriptions z Event and Message service (an event is just a time stamped message) integrating clients and servers yXML specification of events zCompare standard JMS with proposed GMS yGMS XML not text based properties yGMS supports synchronous and asynchronous collaboration yGMS has agent based routing yHHMS or GMS Micro-edition for Hand held clients yGMS uses meta-objects managed MyXoS – Web OS based on interpreted XML (RDF statements) zAll Information Service issues in access of shared resources

9. JMS (Java Message Service) Structure Convert Events to JMS JavaScript JavaC++ ….. Publish PDA without JMS PDA with JMS Global (distributed) Event Receptor (Queue) HHMS Subscribe HHMS (Hand Held Message Service) Optimized for Performance. Assume complex (XML) processing done on “personal server”

10. Bandwidth of Commercial JMS

11. Commercial JMS Latency

12. Comments on JMS z Message latency varies from 3 millisec for small to 300 millisecond for 1 megabyte messages (OK as dealing with human perception time) z JMS has rather limited matching criterion for topics – replace by XML based objects in GMS z Do not use on hand helds: these managed by personal server and linked by HHMS or GMSME which is very light weight zPersistence by subscribing a database system to all topics zNatural distributed model (replication or distribution of topics) for performance and fault tolerance

13. Proposed GMS Model for Messages zAgent based messaging with server (broker) side intelligence and archiving Publishers Subscribers Server Message Input Queue Subscriber Profiles Agent combines Message Properties Subscriber Requests to decide on message recipient Database GMS: Grid Message Service can be built on commercial JMS Subscribes to all events to get persistence Central Knowledge Set via GMS All events in XML

14. MyXoS MetaObject Event-based Computing Paradigm zAll entities are MetaObjects defined in universal GXOS XML Syntax – Meta means GXOS doesn’t really want to manage Object, just information required to find, render and share it yRendering includes Palm devices as well as PC’s yEntities are people, computers, data sources, information from e-mail, scientific visualization to digitized Bible z All actions including object changes are events – all events are GXOS objects ye-mail, paging etc. are all GXOS events uniformly routed/archived etc. z All action is instantiated or recorded in messages which are events themselves z All events are archived allowing actions to be undone and to be reviewed z There is an MetaObject Shell MyXoS with basic Services (copy, create, collaborate etc.) expressed in RDF yDiscovery etc. via Jini/Napster mechanisms