Presented to:
Presented by:
NDIA Systems Engineering Conference
Jim Moreland NSWCDD Chief Engineer
GWU PhD Candidate
1
24 October 2012
Experimental Research and Future Direction
On Evaluating SOA Challenges In A Real-Time,
Deterministic Combat System Environment
James D. Moreland, Jr. Shahram Sarkani, Ph.D Thomas Mazzuchi, D.Sc.
NSWCDD Chief Engineer
PhD Candidate
George Washington University George Washington University George Washington University
Department of Engineering Department of Engineering Department of Engineering
Management and Systems Management and Systems Management and Systems
Engineering Engineering Engineering
Washington, D.C. 20052 Washington, D.C. 20052 Washington, D.C. 20052
Introduction
The Marketplace and Economics are Driving C4I and C2 Infrastructure to be Increasingly Common – Central processing units
– Memory architectures optimized for multi-threaded operations
– High-performance network switches and routers
– Hardware-enabled time synch and distribution technology
– Hardware-based prognostic failure management instrumentation
– (Dynamic) Resource Management for load-invariant performance
– System security and surety technology
– Runtime dynamic state validation
AIS VTS
SPS-73
RLGN
NAVSSI
AIAS
IBS DSVL
Fathometer
NAV
Nav and HM&E
Systems
AAG EMALS
DMRT
C
A
T
C
C
D
AI
R
IFLOLS
LSODS
ILS ILARTS
MWS
TACAN JPALS
ACLS
MOVLAS LRLS ADMACS
Blk III
Aviation
SHIP
Universal Gateway
Ship
Network
CEC
IRST CV-TSC
TISS
DBR
CIWS
IFF
SSTD
SSDS
Tactical
Systems
SGS/AC
EWS
GMLS
RAM
CST
SSES / SCI COMMS MU
OS
MI
DS
CDL
-S
NM
T
Gertr
ude
MC
CP
JTT
-M
SMQ
-11
RCS Turnkey
CDLMS TV-
DTS
G
B
S
T
V
S
D
M
R
HF
RG
AR
C
III
HSF
B
A
R
C
21
0
AD
NS
External Comms Systems
Functional Enclave
Functional Enclave
Functional Enclave
Fun
ctio
nal E
ncla
ve
Fun
ctio
nal E
ncla
ve
JMPS
CANES Coalition
DMS
Proxy
GCCS-M
UASS
DCGS-N
NITES
TCS JWARN
TBMCS
DCRS
IA
DMS
Proxy
MLAN
NITES
TCS
TMIP-M
NTCSS
BFEM
IA
CANES
(Unclassified)
C4I Systems
GCCS-M
SCI
DCGS-N
SCI
SSEE
INC
TC2S
JSF ALIS
CANES
(Genser)
Radiant Mercury
CANES
(SCI)
Point-to-Point interfaces, unique solutions, and weak cross-domain integration
Networked interfaces, common/interoperable solutions, significant cross-domain integration
Research Topic Context - S&T to Achieve Desired Shipboard Environment: Universal Gateway
MIDS
MOS
URC-141NMT SMQ-11
CDL-S
AN/USQ-167
RCS
Turnkey
RFRF RFRFHSFB
SSR-1
RFHFRG
URC-131
RFDMR
SRC-61
RFMUOS
Terminal
RF
ARC-210RF
JTT-M
USQ-151
RFSSES/
SCI COMMS
RFGBS
USR-10
RFTV-DTS
OE-556U
RF
MCCP
BFTT
USQ-
T46
CDLMS
UYQ-86
TVS
USQ-155
ARC III
USQ-16227TV
6TV
14TV
DBR
EWS
CEC
TISS **
CV-TSC
SSQ-34C
RAM
MK31
CIWS MK15
MOD21&22
GMLS
MK29
RF
RF
RLGN
WSN-7
AIAS
Fathometer
UQN-4A
DSVL
WQN-2
IBS
(SCC)
SPS-73
Lite
RF
RF
IFLOLS
ACLS
SPN-46
LSODS
ILS
SPN-41
ILARTS
MWS
CATCC/
DAIR
TPX-42A
LRLS
MOVLAS
SEA
BASED
JPALS
AAG EMALS
VTS
TACAN
URN-25
RF
RFRF
SATCC
Announcing
MC
PPLAN
JSLSCAD
JBPDS
IPDS
MK26
MD0
JCAD
PDR-65A
RADIAC
CANES
(Unclassified)
NTCSS
BK2
TMIP-M
BFEM
IA
AIS
(URN-31)
MCMS
ADNS
USQ-144
TCS
USQ-171
GCCS-M
Genser
NITES
UMK-4(V)
(Genser)
DCGS-N
(Genser)
JMPS
IA
TBMCS
JWARN
NAVSSI
Bk 4.X
GCCS-M
SCI
Radiant
Mercury
SSEE INC
(X)
DCRS
UASS
UMQ-12
CANES
Coalition
MFR
RADIAC
SVDS
SXQ-10B
CANES
(Genser)
TC2S
RF
Gertrude
WQC-2A
RF
GPS
DMS
Proxy
DMS
Proxy
Non-Warfare SystemsInternal WS Interfaces
External WS Interfaces
RF
DAS
IRST **
SSTD
SLQ-25A
CANES
(SCI)
NITES
UMK-4(V)
(Unclass)
JSF ALIS
AIMS
MK XII
UPX-29
RF
TCS
USQ-171MLAN
DCGS-N
SCI
DMRT **
Space & Weight**
RF
ADMACS
BK III
NN
Network
Distribution Bus
SSDS
MK2
RF
RF
RF
TCS
USQ-171
PEO IWS
NAVAIR
NAVSEA
PEO C4I
Today Future
Determinism, Latency, Jitter and Hard Real-Time
“Scientific Determinism”
– All events have a cause and effect and the precise combination of events at a particular time engender a particular outcome.
Fundamental Weapons System Design
– Maintain positive control of weapon.
Latency
– Latency refers to the age of information. System latency is an inherent performance characteristic of any modern computer system.
– Known and predictable latencies can be negatively expanded in a system as the result of application layer faults, hardware malfunction,
network transport layer collisions, and a host of other system response issues. These latencies tend to result in nonlinear behavior of the
system.
Jitter
– Jitter is the ability of the system to repeatedly perform a function to a specified schedule. Many key combat system functions rely upon
predictable periodicity.
Hard Real-Time
Real-time Performance is a Spectrum that can be Met by a Variety of Different Standards, Products, and Techniques. Changes to the Technical Approach in one Area has Consequences on the Others
Low-Latency, Low-Jitter Example
Critical To The Use Of COTS Technology Is To Ensure That Latency Requirements Are Met At Each Step In A Deterministic
Manner While Still Meeting Overall System Reaction Time Requirements Well Below Fault Recovery Time
a/s
mode
detect
mt206
trk msg
update
mt206
a/s
tent.
engage
mt230
mt206sent
tSTART t2 t4 t5 t6 t7 t9 t10 t12 t13 t14 t15 t16 t17 t18t11 t19 tTOTAL
A/S
conf.
engage
mt230
mt306sent
recv/procmt366a/s cand.chk
recvmt306
mt206sent
recv/procmt206
mt362sent
recv/procmt362
recv/procmt365(a/s rvw)
mt365sent mt230 sent
(confirm)
C&D WCSSPYWCSC&DSPY
1. Initiates a track & sends
A/S Detect Notify to CWS.
2. Checks bearing & does
range vs. range rate check
for interceptability.
3. If track passes, Tent. A/S
Engage Order is sent to
WCS. An ID Msg is sent to
SPY and an IFF Mode 4
interrogation is issued.
1. Receives Second Track Msg and sends
a New Track Msg (mt362) to the ID
Function which sends a New Firm Track
Msg (mt365) to the CWS module.
2. A full interceptability check against the
A/S doctrine parameters is performed.
3. If the track passes, a second A/S
Confirm Engage Order is sent to WCS.
t3
recv/procmt206
mt230 sent(tentative)
t8
WS-10650B para.
3.2.1.9.2.4.1
tent.trk
report
A/S
detect
msg
tent.
engage
order
(mt230)
track
update
report
A/S
reso-
lution
A/S
notifi-
cation
engage
rqst
update
(mt240)
engage
order
CONF
(mt230)
tent.
engage
rqst
(mt240)
AN/UYK-43
LEGACY AWS
COTS BASED
AWS
Prototype
a/s
mode
detect
mt206
trk msg
update
mt206
a/s
tent.
engage
mt230
mt206sent
tSTART t2 t4 t5 t6 t7 t9 t10 t12 t13 t14 t15 t16 t17 t18t11 t19 tTOTAL
A/S
conf.
engage
mt230
mt306sent
recv/procmt366a/s cand.chk
recvmt306
mt206sent
recv/procmt206
mt362sent
recv/procmt362
recv/procmt365(a/s rvw)
mt365sent mt230 sent
(confirm)
C&D WCSSPYWCSC&DSPY
1. Initiates a track & sends
A/S Detect Notify to CWS.
2. Checks bearing & does
range vs. range rate check
for interceptability.
3. If track passes, Tent. A/S
Engage Order is sent to
WCS. An ID Msg is sent to
SPY and an IFF Mode 4
interrogation is issued.
1. Receives Second Track Msg and sends
a New Track Msg (mt362) to the ID
Function which sends a New Firm Track
Msg (mt365) to the CWS module.
2. A full interceptability check against the
A/S doctrine parameters is performed.
3. If the track passes, a second A/S
Confirm Engage Order is sent to WCS.
t3
recv/procmt206
mt230 sent(tentative)
t8
WS-10650B para.
3.2.1.9.2.4.1
tent.trk
report
A/S
detect
msg
tent.
engage
order
(mt230)
track
update
report
A/S
reso-
lution
A/S
notifi-
cation
engage
rqst
update
(mt240)
engage
order
CONF
(mt230)
tent.
engage
rqst
(mt240)
a/s
mode
detect
mt206
trk msg
update
mt206
a/s
tent.
engage
mt230
mt206sent
tSTART t2 t4 t5 t6 t7 t9 t10 t12 t13 t14 t15 t16 t17 t18t11 t19 tTOTAL
A/S
conf.
engage
mt230
mt306sent
recv/procmt366a/s cand.chk
recvmt306
mt206sent
recv/procmt206
mt362sent
recv/procmt362
recv/procmt365(a/s rvw)
mt365sent mt230 sent
(confirm)
C&D WCSSPYWCSC&DSPY
1. Initiates a track & sends
A/S Detect Notify to CWS.
2. Checks bearing & does
range vs. range rate check
for interceptability.
3. If track passes, Tent. A/S
Engage Order is sent to
WCS. An ID Msg is sent to
SPY and an IFF Mode 4
interrogation is issued.
1. Receives Second Track Msg and sends
a New Track Msg (mt362) to the ID
Function which sends a New Firm Track
Msg (mt365) to the CWS module.
2. A full interceptability check against the
A/S doctrine parameters is performed.
3. If the track passes, a second A/S
Confirm Engage Order is sent to WCS.
t3
recv/procmt206
mt230 sent(tentative)
t8
WS-10650B para.
3.2.1.9.2.4.1
tent.trk
report
A/S
detect
msg
tent.
engage
order
(mt230)
track
update
report
A/S
reso-
lution
A/S
notifi-
cation
engage
rqst
update
(mt240)
engage
order
CONF
(mt230)
tent.
engage
rqst
(mt240)
AN/UYK-43
LEGACY AWS
COTS BASED
AWS
Prototype
Implementing Open Architecture: Surface Navy OA Technical Model
Common Computing Environment:
• Standards-based Interfaces to network
• Commercial Mainstream Products and Technologies
Componentized Objective Architecture:
• Common Reusable Components
• Ship Specific Components
• Data Model
• Extensible to the Future
Infrastructure: • Common Services and
APIs
• Flexibility to Support Forward-Fit and Back-Fit
Hardware
Operating System
Middleware Decouple Equipment From Computer Programs
Display
Track Mgmt Command
& Control
Sensor
Mgmt
Weapon
Mgmt
Vehicle
Control
Upgrade Computer Programs and Equipment Independently and on Different
Refresh Intervals
Loose Coupling: a principle that minimizes dependencies and only
requires limited awareness of other services
Encapsulation: services are properly captured and packaged across
multiple implementations
Abstraction: logic and data hidden from the outside world
Composability: collections of services can be coordinated and
assembled to form more complex and capable services
Discoverability: services are designed to be outwardly descriptive so
that they can be found and assessed via discovery mechanisms
Service-Oriented Architecture (SOA) Principles
Services + Loosely Coupling Agility
Discovery, N
egotiation
Composite
Choreography Protocols
Atomic
Reliable
Messaging Security Transactions
Components
Quality
of
Service
Interface +
Bindings Policy Description
XML Non-XML Messaging
Transports Transport
SOA Stack
SOA Web Services for Combat System
Small-Scale Experiment
Real Time Non-Real Time
• Data based assessments of:
– Non-real time open-source web services middleware with real-time Java
constructs
– Determinism and hard real-time performance of a web service application
– Potential combat systems benefits from modifications at the middleware layer
and application layer for the use of SOA web services
Experiment Design: High-Level Diagram of Entity
Processing Web Service
Scenario:
• 100 entities sent from the Message Producer Service to the Message Consumer Service on a 1 second interval.
• Message Producer Service sends 5 entities per buffer.
(e.g., 20 buffers sent per second = 100 entities)
• Test duration is 1 minute.
• Nodes time synched to 10’s of microseconds with NTP.
• Messages must be received, parsed, and processed within 100 milliseconds – equating to a processing requirement
approaching 50 Kbytes per second.
Experiment can measure a time latency of 2 milliseconds and state that it is higher than a time latency of 1 millisecond.
Internal Processing Latency for Profiles
Experimental Results
10x Performance Gain Using Real-Time Techniques With Web Services
Profiles 1 and 5 Performance Comparison
Profile 1’s end-to-end processing latency, as depicted in the left image, showed a
greater level of determinism than Profile 5’s end-to-end processing latency on the
right.
Profile 1 has a tightly bound maximum values than Profile 5.
Profile 5 has a far greater number of outliers and some documented jitter in the
message processing.
SOA Challenges for Combat Systems
Semantic and metadata management
Transformation and routing
Governance across all systems
Discovery and service management
Information consumption, processing, and delivery
Connectivity and adapter management
Evolution From Traditional Point-To-Point Into Net-Enabled Capabilities
Experimentation Focus Areas
Real-Time Services – Validation in a dynamic environment – Security – Automation
Adaptive/Dynamic Computing Technology – Static vs dynamic resource management – Verification and validation for adaptive and/or dynamic operation – Closed-loop control with:
• Dynamic assignment at run time • Dynamic reassignment for load / failure modes
Precision Time – Synchronization – Distribution – New warfighting capabilities (distributed electronic warfare)
Limited Technology Experiment
Executive Summary Results
Universal Gateway prototype evaluated met Technology Readiness Level (TRL) 6, based on successful performance in a representative environment. – Targeted for use in a Product Line Architecture-based combat system
environment.
Two-way and One-way readiness message exchange across combat system and command and control domains: – Data latency of combat system to command and control system track messages
was minimal at the largest message load with typical end-to-end gateway latencies under 20 milliseconds.
– Data throughput across universal gateway was not limited by universal gateway components. Universal gateway successfully processed the maximum data throughput that the combat system sent/received.
– Mediation of combat system track and readiness message types and track position formatting was handled successfully within universal gateway with consistently small latencies of less than 10 milliseconds.
– Security tagging was handled successfully with universal gateway.
– Universal gateway rule engines successfully provided the capability to dynamically tailor the track and readiness message data.
Summary
Pace of C2/C4ISR convergence quickening
Fire control R&D community must lead the way both intellectually and with robust experimentation to affordably buy down risk – C2 experimentation plan
– Joint experimentation venues
Successful COTS insertion results from applying systems engineering to the control loop and mitigating problems via critical experimentation – Real-time infrastructure experimentation
Great Opportunities For Cross-Domain Collaboration