ACSys II Research Theme Object Technologies

1
ACSys II Research Theme -Object Technologies
Systems

• April 1999
• Ken Hawick
• khawick_at_cs.adelaide.edu.au

2
Theme Programs

• Theme Task
• Digital Libraries for Edu Services
• Collaborative/Telepresence scenegraph object dist
• Remote Access to Data Warehoused objects
• HPC Distributed Environment management

• Program/Business Area
• Interactive Media
• Virtual Environments
• Data Mining
• Geospatial/Earth Obs

3
OTS Targets

• Contract RD Consultancies
• 1065 over 5 years
• Education and Training
• 175k over 5 years
• Fellowships
• 2 per node over duration of ACSys II
• Licenses Fee
• 90k - one license every 2 years

4
OTS Actions

• Initial Projects and Project leaders
• Researcher assignments
• Events and Calendar
• planning forums demos meetings
• Procedures and the Exec Forum
• project review reporting program integration

5
OTS Support for VE

• Telepresence Cooperative Working
• Dist / Persistent objects for exchanging
  scenegraphs between remote immersed users
• Complex obj mgmt wide areas (latency)
• Builds on DISCWorld env (ACSys)

6
OTS Support for IM

• Active Digital Libraries
• Large Objects
• Complex metadata exchange
• OO DBs and large archives
• Educational Services Support focus
• Builds on OLDA AST (ACSys)

7
OTS Support for DM

• Remote Data Access (object transport)
• Object models across firewalls (proxies)
• Scalable compute and storage systems
• Business Logic Rules (Obj Registry Mgnt)
• Builds on Persist Java/Stores DISCWorld remote
  computation

8
Geospatial Objects

• Activities in US (standards forums - OpenGIS and
  Nat Image Mapping Assoc)
• SA Government initiative to build Spatial
  Information Integration System (SIIS)
• Potential for Federal integration (eg State
  Herbariums data)
• Key is distributed objects middleware able to
  incorporate large data stores high perf
  procesing and robust, secure distribution
• Java CORBA HPCC features DISCWorld

9
Research Directions

• 1) Distribution
• latency issues
• fault tolerance - network failures
• client/server-server architecture
• dynamic join/depart - scheduling
• scalable name spaces at app level
• standards compliance
• object based parallelism - new directions

10
Research Directions 2

• 2) Persistence
• Persistent Java language support
• Stores
• Scalability
• Object has special meaning
• Orthogonal persistence (independent of other
  programming issues)

11
Theme Operations

• Theme projects IN the progams
• Push/pull relationship
• Sign off budget model
• Commercially driven research
• OTS has own targets

• OTS projects support programs
• OTS staff on Program projects
• OTS vehicle for Tech exchange
• Rapid response to Programs needs

12
International Data Corp Report on Middleware

• Middleware - future of distributed business
  applications
• Spares business applications developers from
  network idiosyncracies
• Total (middleware) markets US5.275 billion by
  2001 (416 growth from 97)
• Internet is key factor
• Wall Street Firms agree on middleware as tool of
  future (HPCWire, 29 May 98)

13
Need for Research

• Object technologies
• Enable fast application development time
• Allow applications interoperability
• Distributed objects support Internet/WWW
  framework
• Embed legacy codes systems for continued
  support
• Emerging technology - not yet mature - RD needed

14
Objectives

• Develop promote DISCWorld middleware (software
  glue to enable distributed apps)
• Develop deploy persistent Java Tech
• Transfer Tech secondments/students
• Develop app demos prototype standards compliant
  s/ware toolkits methodologies
• State-of-the-art research areas fault tolerance,
  scalability, persistence.

15
Unique Capability

• Sizeable group - critical mass of research staff,
  students industrial collaborators
• Internationally recognised group networked into
  standards bodies and forums - OpenGIS, Java
  Grande, US Nat Image Mapping Assoc, IEEE Task
  Force on Cluster Computing invited journal
  publications on middleware
• Unique High Performance Approach - large data,
  HPC parallel processing expertises, facilities
  infrastructure - DSTO using ACSys S/Ware
• Recognised Java Object Modeling Language
  Trainers/Consultants - Sun Certified Java Trainers

16
Deliverables

• VE scenegraph distribution framework in 1 yr
• Remote Data-mining support infrastruc in 1-2 yrs
• Digital Library/Data Warehouse architecture
  proof of concept in 1 yr, robust prototype in 2
  yrs
• OpenGIS prototype within 1 yr
• Defense infrastructure demonstartors middleware
  - staged delivery with new capabilities in 1,2
  and 3 yrs
• Meteorology archive delivery within 1 yr
• 4 PhD/Yr 4 Hons Students/Yr 1 Trained Prog/Yr
• 2 Secondments/Yr

17
Commercialisation Targets

• Geospatial Object Systems Middleware
  Demonstrators (SA Govt) - design in 99,
  prototypes by end of 2000.
• Business Registry Toolkit (Aust Bureau of Stats)
  - design in 99
• Support of Programs Comm. targets

18
Outcomes

• Support geospatial architectures for SA Govt
  (internat. competitive standards compliant)
• Value to Australian Businesses - software
  capabilities in Environmental Meteorological
  Emergency Services Defense areas.
• Coherent technical research across ACSys II
• Australian participation in Application domain
  Object Standards (eg Geospatial)
• Trained graduates/postgrads to Aust workforce
• New Centre/Business Alliance in Obj Tech Sys