Simulation and Evaluation of Adaptive Mobile LargeScale Network Systems SEAMLSS

1
Simulation and Evaluation of Adaptive Mobile
Large-Scale Network Systems(SEAMLSS)

• Gary Warren, Vince Diehl, Laurance P. Longtin,
  Aaron Steigerwald, Chris Kostas, Shawn Chou,
  Andre Ghetie, Chris Heydemann
• SAIC
• Adam Sulesky
• PTI

2
Outline

• Objective and Approach
• Overview
• SEAMLSS Scenario Development
• Some Key FY99 Activities
• Conclusion

3
Objective

• Develop and populate a simulation environment
  that allows medium- to high-fidelity evaluation
  of DARPA Global Mobile (GloMo) communication
  devices and technologies in realistic military
  scenarios.

4
Approach

• Build flexible, extensible simulation environment
  based largely on COTS tools
• Populate environment with simulation components
  provided by GloMo technologists
• Develop library of scenario-independent military
  communication behaviors
• Use environment to evaluate technologies

5
SEAMLSS Overview
Scenario Compiler
SDF/HLA
Complete SEAMLSS Scenario
Input Interface
Simulation
Protocol Modeling Language
PDEF
HLA or SNMP
Analysis Interface
Evaluation
6
Modular Scenario Development

• From the start, SEAMLSS built using modular
  scenario definition
• Benefits
• Accurate Specialists develop high-fidelity
  models in their domain, whether operational
  threads, communications devices, network plan, or
  scenarios.
• Fast Allows all these components to be defined
  once, then reused in a mix-and-match fashion

7
SEAMLSS Scenario AssemblyStep 1 Base Scenario

• Select base scenario
• Includes entity motion, entity type, command
  hierarchy, entity status and triggers
• Scenarios sources
• Create in ModSAF
• Use Surrogate Data Engine (DARPA Small Unit Ops
  program)
• Created by hand

8
SEAMLSS Scenario AssemblyStep 2 Attach
Communication Devices

• Select communication device models
• Select OPNET, PARSEC, NS2 or CPT device models
• Create communication types
• Assign one or more devices to a scenario
  independent platform to get, for example, Soldier
  with SINCGARS or Squad member with HMT

• Map base scenario entity to SEAMLSS communication
  type

9
SEAMLSS Scenario AssemblyStep 3 Develop
Communications Threads

• A thread is sequence of messages that perform a
  task. A thread may contain one or more links,
  branching, spawn additional threads, etc.
• Develop one or more threads for each SEAMLSS
  Communication Type (e.g. Squad member with HMT).
• Specify thread schedule deterministically,
  stochastically, or based on triggers in the base
  scenario (e.g. ModSAF).

10
SEAMLSS Scenario AssemblyStep 4 Define
Communications Groups
Sample network planning based on legacy
communications

• Generate default communications groups
• Base scenario command hierarchy used to generate
  default communications groups
• User can tailor as needed
• Run-time process automatically converts to IP
  address lists and attaches to threads

CO Net
CO
PLT
PLT
PLT
PLT Net
SQD
SQD
SQD
TEAM
TEAM
TEAM
SQD Net
11
SEAMLSS Scenario AssemblyStep 5 Scenario
Compilation

• Scenario compiler builds OPNET executable.
  Includes
• Entities and associated devices
• OPNET device models
• Threads
• Internal stochastic or deterministic thread
  triggers

12
Scenario Execution

• Once scenario assembled or existing scenario
  selected, automated tools start and monitor all
  components, on various machines, automatically
• Simulation progress and limited analysis feedback
  provided to user

Network Analysis
SAF
OPNET
PARSEC
Thread Manager
Prop Model
SNMP Gateway
HLA/RTI
13
Analysis

• Several analysis tools and concepts under
  development.
• Emphasis is on thread-level metrics (Did the task
  complete?), not packet-level performance

• SNMP Gateway delivers SNMP MIB messages for
  display and analysis in standard network analysis
  tools (e.g. HP OpenView).
• Enables direct comparison with real networks or
  hardware-in-the-loop emulations.
• Quantify packet, thread and battery performance
• During FY99, additional analysis tools will be
  added to federation.

Network analysis
SNMP Gateway
HLA/RTI
14
FY99 ActivitiesTech Transfer Opportunities

• CECOM
• GloMo Testbed - VV via replication of simulated
  scenario in real world
• Next Generation Performance Model - Include
  SINCGARS and EPLRS models and situation aware
  communications behaviors into SEAMLSS
• J6I NETWARS
• SEAMLSS thread databases, communication
  behaviors, simulation interfaces
• SUO
• Shared scenarios
• Deployment requested as GFE by integrator
• CRDA with NIST
• VVA of environment, models, simulation and
  metrics with NIST Information Technology Group
• SMART-Nets
• Manage real or simulated networks with same tools
  via SNMP

15
FY99 ActivitiesThreads

• Additional threads
• Test, development and validation threads to
  support simple experiments
• Video threads
• Sensor threads
• Threads based on situation awareness from base
  simulation (e.g. ModSAF)
• Thread processing enhancements
• Improvements to triggering of threads
• Probabilistic thread branching
• Dynamic selection of communications device
• Refinement for reliable/best-effort data
  messages.
• Support for retries

16
FY99 ActivitiesSEAMLSS-Lite Delivery

• Created and delivered via web site at
  www.seamlss.com
• SEAMLSS-Lite allows user to test and validate, at
  own facility, device model integration
• Replace template with OPNET model in SEAMLSS-Lite
• Validate
• Add SNMP calls and put into SEAMLSS
• Validate against SEAMLSS-Lite results
• SEAMLSS-Lite is a subset of the entire SEAMLSS
  environment that
• Does not use HLA
• Does not produce SNMP messages
• Has only randomized mobility

17
FY99 ActivitiesCPT Model Wrapper

• Objective
• To make use of large number of Rooftop CPT
  models, build OPNET wrappers to enable inclusion
  of high-fidelity CPT models in SEAMLSS
• Approach
• Create a wrapper that makes GloMo-relevant CPT
  models callable from OPNET.
• Risks
• Time synchronization between OPNET and CPT
• May require new OPNET modules for interfacing
• May encounter compatibility issue between the
  existing OPNET models and CPT models.

18
FY99 ActivitiesLink-Level Propagation Model
Frequency, Antenna heights
Link performance (Data rate, errors, throughput,
etc.) to OPNET simulation
P.L.
Path loss model (new Bellcore model)
Pr Successful xmsn
P.L., variability
SNR
Determine node coordinates (x,y,z)i ,
(x,y,z)j for link
Fully automated for SEAM-LSS, using all
available environment data
Other link parameters (Antenna gains, power
levels, etc.)
Perform for all candidate bands and data
rates, based on the technologys adaptation
algorithm
19
FY99 ActivitiesEnvironment Population

• Scenarios
• 4 DARPA Small Unit Operations scenarios
• Additional test scenarios
• Additional operational scenarios
• Threads
• Additional voice, sensor, video threads
• Thread editor useable by the Operational User
• Models
• Integration of GloMo device models (PARSEC,
  OPNET, CPT, NS2)
• NETWARS library
• CECOM library

20
FY99 ActivitiesSummary
Input Interface

• Enhance import and user editing

• New GUI
• Enable Operational User to create threads and
  execute simulation

SDF/HLA
Scenario Compiler
Complete SEAMLSS Scenario
Simulation

• Support for retries
• Application interface
• Multiple devices per platform

Protocol Modeling Language

• Parallel OPNET
• HLA 1.3
• New RF Prop Model
• Implement HLA thread manager
• OPNET-PARSEC co-simulation

• CPT model wrapper
• GloMo model pop.
• GloMoSim APIs
• API Upgrades

PDEF
HLA or SNMP
Analysis Interface

• Sensor, video, sit. awareness threads
• Triggers
• Branching
• Support for retries

Evaluation

• Leverage SMART-NETS activities

21
Conclusion

• SEAMLSS is useful today for medium- to
  high-fidelity simulation and evaluation of DARPA
  GloMo technologies in realistic military
  scenarios.
• Key features
• COTS simulation models (OPNET, PARSEC)
• HLA backbone
• Modular, flexible scenario definition
• Fast, flexible propagation model