Title: RTI Working Group Issues
1RTI Working Group -- Issues
Object Models
HLA/RTI
- Endorsement of RPR FOM
- RDE Reference FOM
- Multiple Inheritance Impacts in FOMs
- Non-object Model Federation Data for
Interoperability - Requestable Objects
- HLA Use Cases for RDE
- RDE Lessons-learned With HLA
- RTI Goals (Timing, BW)
- RTI Integration Tools
- Interoperability Between RTIs
- Multiple routing of attributes
Other
Fidelity
- SISO-sponsored Fidelity Group
- Fidelity Glossary
- Fidelity Characterization in SOMs
- Fidelity in OM Tools
- SIW Paper Review Requirements
2RDE Lessons-learned With HLA
- Abstract The potential RDE uses of HLA need to
be sufficiently well characterized to ascertain
the functional or performance needs on the RTI or
middleware products. - Narrative The RDE will develop, disseminate,
collect and publish summary application profiles
for active/planned HLA experiments within the
RDE community. The RDE will encourage the
preparation of DMSO-standard Federation Execution
Planning Workbook Templates for experiments in
which it is possible to do so. The RDE will
encourage preparation and presentation of
lessons-learned from such experiments. The RDE
endeavor to extract significant conclusions and
trends from the assembled reports, and
proactively communicate them to appropriate SISO,
DMSO and user organizations.
3HLA Experiments Being Monitored by RDE User Forum
4Summary Experiment Description
5HLA Use Case Definition for RDE
- Abstract The pacing or key future RDE
applications of HLA must be sufficiently well
characterized to ensure that the RTI and related
HLA services are responsive to the emerging
needs. - Narrative The RDE User Forum will review the
planned 1997/98 community experiments to assess
the degree to which they span the realm of core
simulation community applications in the
1998-2003 timeframe. - To the extent that key application areas are not
represented, the RDE will investigate the
potential for defining Use Cases appropriate for
capability planning and future testing.
6RDE Use Case -- RDE-RT-1High dynamic, real-time
user
- X vs. X Fighter Aircraft Federation based on
SIW-98S-042 and associated Federation Execution
Notebook - Two example scenarios
- Synchronous platform pos-vel state data, with
infrequent full MilitaryPlatformEntities updates.
Best effort receive order. Reliable transport of
infrequent events. Scale from 1 Vs. 1 to M vs. M.
Peak latency is key performance metric (1/60 Hz). - Synchronous platform pos-vel state data, with
infrequent full MilitaryPlatformEntities updates
1 vs. M scenario. Bursty transmission of
asynchronous event sequences (0-100/sec). Best
effort or reliable receive order. Peak latency
is key performance metric (100 msec).
7Tentative RDE Use Case Class Descriptions
- Real-time (RT) - Emphasis is on real-time or
pseudo-real-time execution. Data latency is
generally a core concern. Most current DIS
applications fall in this category. Best effort
often suffices. - Complex Models (CM) - Characterized by
engineering federations with complex (deep) model
hierarchies, attribute sets and subscribe/reflect
mappings. - Proliferated Objects (PO) - Federations with
large numbers of atomic entities (tanks, mines,
). - Hardware-in-the-loop (HW) - Require interface
with (servicing of) real military hardware
devices, from entity to component level.
Generally requires time synchronization to
external standard and reliable transport.
8Use Case Documentation Needs
Scenario Descriptions
Fed Ex Notebooks
Task Flow Descriptions
9Use Case Implementation PlanMerge Object
Management and RTI WG Perspectives
Jan.
Mar.
Jun.
Sept.
I/ITSEC Dec. 1998
Interim
SISO
Interim
SISO
- Jan Developed Initial Concept
- Mar Brief and Obtain General Concensus
- Jun Focused Planning (Participant Federates)Use
Case Development - Sept Review Final PlansComplete Experiments
- Dec Present Findings at I/ITSEC
10RTI Attribute Reflect Servicing Performance
- Abstract Timed tests at AFRL with RTI 1.0.2
suggest that RTI Ambassador servicing of the tick
function may consume as much as 80 of a users
total reflect attribute cycle time. Request that
DMSO ensure that the RTI service for releases
1.3 substantially improve responsiveness. For
Use Case RDE-RT-1A, a 10-fold improvement to only
25-30 of total event processing time is
reasonable goal. Failure to reduce the overhead
will severely impact some planned federations.
- Narrative AFRL executed an X vs. X aerial
engagement federation with simple
MilitaryPlatformEntities, receive-order event
processing, not time regulated. On average, 80
of the apps time for processing attribute
updates is consumed in RTI Ambassador processing.
11RTI Integration Tools
- Abstract The HLA applications community needs
tools or protocols which permit users to
characterize the status/impact of RTI loading,
data handling and message traffic on the
execution of their federations. - Narrative Typical RDE users are constructing
HLA-based exercises to meet specific objectives,
subject to various resource (CPU, bandidth)
constraints and performance goals (timing). The
inability to read (or tag) RTI messages on the
wire seriously hampers users in their attempts
to characterize execution of their federation and
intelligently implement strategies for optimizing
performance to meet their objectives. Execution
monitoring tools for RTI performance are
essential performance prediction tools are
desirable.
12RTI Bandwidth
- Abstract Future RTI releases need to improve the
packaging efficiency of packets to avoid
penalizing current DIS users. - Narrative In a real-time application focused on
exchange of platform entity state information
(RDE Use Case RDE-RT-1), packet efficiency with
RTI 1.0.2 was substantially lower than a
comparable DIS protocol implementation, requiring
92 bytes of RTI overhead to transport 60 bytes of
information. If not addressed future RTI
releases, this inefficiency will prevent current
DIS users from realizing the bandwidth savings
potential of HLA.
13Interoperability between dissimilar RTIs
- Abstract The RDE forum is concerned about
potential interoperability barriers introduced by
the anticipated proliferation of RTI
implementations. The RDE forum encourages the
discussion of architectural approaches that may
be utilized to further promote interoperability
amongst extended federates.