RTI

1
RTIC Forum

• Keith Briggs, Chair
• Duncan Clark, Vice Chair
• Jeff Fischer, Secretary

Total attendance 70-90 Papers presented 20 4
with RDE Discussion Sessions 2 RTI
Interoperability Study Group Meeting
2
Goals for this Workshop

• Provide a forum for discussion of critical HLA
  issues associated with
• RTI Performance and Performance Prediction
• Ownership Transfer
• Evolving RTI Architectures
• Data Distribution Management
• Component and Embedded HLA Applications
• Determinism
• RTI Interoperability issues

3
Key Points

• Understanding the RTI
• RTI performance benchmarking standards
• Black Box Behavior
• IF specification improvements
• Ownership management changes
• Additional time managed services
• Deterministic RTI behavior
• Controlled Arbitration / Direction services
• RTI Flexibility Interoperability

4
Understanding the RTI
5
Performance Benchmarks

• Explanation A common / standard approach to
  performance measurement and benchmarking is
  needed.
• Three levels of benchmark may be needed
• Application specific benchmarks
• Standard benchmarks for all RTIs
• Performance models where possible
• Forum voted to develop a SISO standard benchmark
  suite.
• 75 SISO standard, 5 de-facto standard, 0 not issue

6
Black Box Behaviour

• The RTIC forum has asked for access to the
  Government sponsored RTI source code at every
  SIW.
• We still dont have it
• We still want it
• Lacking source code an internal detailed and
  accurate process description of the RTI is
  required.
• It is needed to perform a reasonable federation
  design process.
• Forum suggests that a standard documentation
  package be produced for all RTIs including this
  information

7
IF Spec Changes
8
Ownership Management

• Explanation The existing ownership management
  services are inadequate for effective use. Four
  improvements were identified as being required
• Transfer of simulation objects to include public
  private state data and serialized methods.
• A means to provide best and final state data
  with ownership transfer.
• Time managed ownership transfer services.
• Directed transfers
• Recommendation Utilize the tiger team
  reflector to establish a consensus comment for
  submission in response to balloting.

9
Time Managed Services

• Not all services can be time managed within the
  HLA specification.
• Use of these services in a time managed
  environment can produce indeterminate and
  undesirable results
• A subreflector to the RTIC reflector is
  suggested to continue discussion.

10
Deterministic RTI Behavior

• Certain RTI services do not result in
  deterministic federate - RTI contracts when
  invoked.
• Subscription, Modify Region are examples.
• Return services indicating contract completion
  are required.
• Implementation and policy issues remain with
  proposed solution.
• This will be discussed on the time management
  reflector.

11
Controlled Arbitration / Direction Services

• Explanation There is a desire to allow
  federates increased control over the targeting
  and arbitration algorithms contained within the
  RTI.
• A capability to direct a service at a particular
  federate can provide increased utility and
  performance.
• Methods to identify and select the target
  federate need to be researched.
• Discussion and papers on this topic will be
  solicited.

12
RTI Flexibility

• Explanation An explicit capability to control
  the set of services for run-time use within the
  RTI is desired
• There is a concern that within some RTIs the
  addition of advanced capabilities affects
  performance and memory requirements whether used
  or not.
• The forum would like a method to indicate that a
  service capability is not a required service.
• As a result, the RTI could rationally release
  these resources.

13
Recommended Reading

• Recommended papers and presentations
• Implementation of the next generation RTI
  99S-SIW-118
• Javelin 99S-SIW-158
• Recommended presentations
• Benchmarking the RTI for Simulated Radio
  Applications. 99S-SIW-046