US AWACS Modernization

1
US AWACS Modernization
The Use of Distributed, Real-Time Object
Oriented Technology The AWACS Example 13
January 1999
Blue Superiority

• Major Michael K. Milligan
• AWACS Command and Control Programs
• ESC/AWC
• Hanscom AFB, MA 01731
• DSN 478-5631, [phone] milliganm_at_hanscom.af.
  mil

2
Outline

• The Future of AWACS and the Need to Modernize
• Modernization Strategy
• Lessons Learned and Future Challenges
• Real-Time DII COE
• Summary

3
The Future of AWACS

• The Role of AWACS has Evolved since the 1970s

The Future
The Past

• European war
• National defense
• CONUS defense
• Surveillance
• support to ground
• based defenses
• Bombers, fighters

• Theater wars
• Tactical Air Defense
• No-fly zone enforcement
• Counter drug
• Theater Battle Management
• Offensive Counter Air
• Surveillance support to
• ground air battle
• Identification (red/blue/gray)
• Control of engagements
• Small fighters, Cruise missiles,
• Helicopters

Mission Role Target Class Upgrade Appro
ach
Large, Block Modifications
Small, Incremental Frequent Improvements
4
Why Change is Needed

• Current Kill Triad is Dominant, but Not Perfect
• Air picture
• Track errors (JADO/JEZ, ASCIET 95)
• Multisensor data
• Low/Slow
• Strike SEAD
• Deep Interdiction
• BVR kill ratios
• Link 16
• Interoperability (F-14s)
• Information Display
• Ground Knowledge
• Terrain
• Threat volumes
• System Complexity/Crew Demands/Training Overload
• Availability
• Aircraft Turn Around Time
• Rising OM costs

• Threat is Evolving
• Cruise missiles
• Complex ID Environment
• Blue/Gray
• Emerging threats
• Small, maneuverable fighters
• To Match AWACS Capabilities to Evolving Force
  Structure
• F-15, F-22
• F-16/HTS, JSF
• Army Helicopters
• Precision Munitions
• JTAMDO Architecture
• CEC/Aegis
• Patriot
• National Technical Means
• Reduced Manpower

5
Why AWACS Computer Modernization?
Current AWACS Baseline Architecture
Key Counter-Air Requirements

• Closed Computer Architecture (1960s technology)
• Future Upgrades costly/not technically feasible
• Non-DII COE
• Human Machine Interface Deficiencies
• Track Quality/CID Inaccuracies
• CRT is DMS / Not Supportable in the Outyears
• MTBF rates (14hrs) impact Aircraft Availability
• MC Rates Continue to Decline/OM Costs Rising

Future AWACS Capabilities
Integration of Broadcast Intel
Combat ID Correction

• Completion predicated on Step 1 architecture
• High Congressional Interest in these
  capabilities
• Sept 94 Blackhawk Report highlighted consoles
  display
• Failures and need for IFF/tracker capability
  upgrades

Existing Mission Computing Architecture Presents
Barrier to Capability Modernization
6
Computer Upgrade is the Enabler
Insertion of Needed Capabilities
Current AWACS Baseline Architecture
Advanced Capabilities Broadcast Intel
Integ Facilitates Future Upgrades
Direct Benefits Improved IFF/CID
Tracking Maps/Displays/MMI COTS Open
System DII COE Compliant Plug Play
Applications Mission Availability/Reliability
Computer Upgrade (Step 1)
Step 1 CD Opens the Architecture Allows new
capabilities/technologies to be realized -
Improves Ops Performance/Reliability
7
Need to Modernize

• Capability
• Cant Support Future Missions
• Computer Architecture is Foundation for Future
  Applications
• Affordability of Upgrades
• Decreasing Budget
• Time Need Upgrades Now!
• Cost of Ownership
• Decreasing Budget

8
Outline

• The Future of AWACS and the Need to Modernize
• Modernization Strategy
• Lessons Learned and Future Challenges
• Real-Time DII COE
• Summary

9
Modernization Strategy

• Need Better Capabilities for Warfighter
• Situational Awareness
• Offensive Counter-Air
• Tight Fiscal Schedule Constraints
• Spiral Acquisition
• Joint Technical Architecture (JTA) COTS
• Migration from Legacy (not Wholesale Replacement)
• Update Mission Computing Architecture Foundation
• Future Low Cost/High Impact Improvements
  (Enabler)
• Build to Open Standards
• Use Real-Time, Object Oriented Design Techniques
• Timing Constraints Driven by Operational
  Scenarios
• OO Design Reduce Cost, Schedule and Performance
  Risk

10
Mission Computer ModernizationSpiral Development
? . . .
NATO Mid-Term
Dec 00 ? Step 1 Step 2 Begins ?
11
The Past and the FutureHardware
AWACS Central Computer
AWACS Open Systems
Data tap
Custom Mil Spec Expensive
Standard Commercial Low Cost
12
The Past and the FutureSoftware
The Past
The Future
POSIX, CORBA,ODMG-93
No Commercial Standards

• Obsolete
• Large Integration
• Central
• Proprietary
• Fragile
• Inflexible
• Quality tested-in
• Unpredictable operation

• State-of-the-art, Object Oriented
• Plug-and-play
• Distributed
• Industry Standard
• Robust
• Flexible
• Quality designed-in
• Guaranteed performance

13
The Past and the FutureAcquisition Pain
Savings

• Large integration and test burden
• Reinvents the wheel, every time
• Custom, stovepipe, closed infrastructure
• Focus on process
• Large, long-term, block upgrades
• 10 Years EMD to Field
• Serial development

• Standard, Open infrastructure
• Small integration burden
• Integration is plug--play
• Focus on Product
• Incremental upgrades
• 3 Years EMD to Field
• Parallel development

I
n
t
e
g
r
a
t
i
o
n
C
a
p
a
b
i
l
i
t
y


14
The Past and the FutureIntegration Pain
Past
Future
Milestones show Paper progress
Milestones show Demonstrated progress

• Large Integration Cost
• Lessons Learned After End
• Increasing Integration Pain
• Quality tested-in
• Unpredictable Cost/Schedule

• Tapering Integration Cost
• Lessons Learned Early
• Decreasing Integration Pain
• Quality Designed-in
• Managed Cost/Schedule

15
The Past and the FutureRisk/Functionality
Past
Future
Risk remains high until program end
Steady risk reduction
1
PDR
3
CDR
5
Qual
7
Delivery
Little fieldable functionality until program end
Functionality increase each step of the way
16
Modernization Strategy

• US Modernization is incremental
• Step 1 provides critical capability early and
  limited GCCS/DII COE compatibility
• Step 2 completes the migration to an open
  computer architecture and provides full GCCS/DII
  COE compatibility

NATO Mid-Term
Baseline Architecture
FY 95
US Ongoing
Step 1
FY 98
US Planned
BBT Tracker Maps/Colors/Consoles IFF Fix Opens
Architecture Exec Agent for RT DII COE
Step 2
MSI, MMI Distributed Computing Full Plug
Play Automated ATO
FY 02 - 04
17
US Step 1 ModernizationTechnology Insertion

• Plug play software using commercial standards
  and products

1998-2004
VME Processor
Software Standards
Encapsulated Application Modules
Software Infrastructure
Open Features (H/W S/W)
Real-Time POSIX Operating System
Sensors
Offensive Counter Air

• IFF Nav Deficiencies fixed
• IFF jitter
• Nav accuracy

• Intuitive display of information
• Windows environment

• Improved SA with map underlays, ground knowledge
  and full color symbology

• Reliable tracking of maneuvering targets

Dissemination (HCI)
Processing (Tracker)
Displays
18
US AWACS Performance Deficiencies Target Tracker
Radar
IFF
mode/code
azimuth
range rate
azimuth
range
range
AWACS Tracker
Track Reports

• Tracking Fidelity
• (Association Correlation)
• No Fusion of IFF Radar
• Limited Number of Tracks

Nav
19
Current Tracker vs Best of Breed Tracker (BBT)

• Best of Breed Tracker
• full state kalman filter complex rules
• multiple hypothesis testing
• uses Radar and IFF
• requires much higher processing capacity

• Current tracker
• suboptimal kalman filter binary rules
• uses Radar or IFF but not both
• processing capacity limits algorithm complexity

Baseline
BBT
20
New U.S. Step 1 Tracker Improves Track
Continuity/Maneuver Response
Current Tracker
U.S. Step 1 Best of Breed Tracker
21
AWACS Performance DeficienciesWorkstation
Computer Displays
Bottleneck
8 colors/symbols
ECSP (hard-wired) Drives Situation Display
Consoles (SDCs)
3-bit word
22
Workstation Computer Displays
Bottleneck

• Full Color Capable
• Eliminates Red/Green Violation
• Improved Maps
• True Fonts/Symbols
• Window Management
• Cursor in tab area
• Entry Edit Line

DTED
Vector
Nav/Geo
23
New Displays (Colors Windows)
Current Display
Example U.S. Step 1 Display
24
New Displays (Maps)
Example U.S. Step 1 Displays
Current Display
Improved Vector
Nav-Geo Raster
Digital Terrain Elevation Data (DTED)
25
Mission Computing Deficiencies

• Target Tracker Algorithm
• Navigation/Sensor Accuracy
• Workstation Computers
• Human Computer Interface (Symbols, Colors, Switch
  Actions)

• Closed System
• Legacy Hardware
• Legacy Software
• All AWACS Unique Interfaces
• No DII COE Compatibility

CC-2E
CC-2 (IBM 360)
CAU
AOCP Tracker
HCI
SDC 1
SDC 2
ECSP
Sensors Communication
DMX
Hardwired Display Processor
SDC 3
Radar
. . .
IFF
WSC
ESM
BUBBLE MEMORY
CPS
TADIL A/C
Nav
TAPE/ HDS
JTIDS
SDC 14
Etc
PRINTER
26
The Past and the FutureHardware
AWACS Central Computer
AWACS Open Systems
Data tap
Custom Mil Spec Expensive
Standard Commercial Low Cost
27
U.S. Step 1 Hardware Architecture

• Open System Implementation
• Step 1 CY98, Step 2 CY04
• Fully Networked Client-Server Arch
• Low Cost Growth Potential
• Extensive Use of COTS
• Processors (PowerPC)
• LAN (Fibre Channel)
• Power Supplies (Brandt)
• Cabinet (Zero Corp)
• Production Savings 81.7M
• 56 JTA Compliant (Step 1)
• DII COE Compliance
• Step 1 Limited
• Step 2 Full
• JTIDS/Link 16 NOT Delivered as Part of Step 1

CC-2E (IBM 360)
SDC 1
A3
ECSP-R Workstation Electronics (PowerPC)
AOCP (Mission Program)
SDC 2
A3
SDC 3
A3
W/S
W/S
SDC 4
A3
W/S
. . .
Demo 9/95
SDC 5
A3
RDMX (DMX Emulator, PowerPC)
W/S
. . .
MSC
LAN (Fibre Channel)
W/S
SCSI
Sensors/Communications
4
Radar
JTIDS/Link 16 (PowerPC)
1553
Best of Breed Tracker (PowerPC)
Monolithic Memory
IFF
ESM
Demo 5/98
CPS (I/0)
SDS
TADIL A/C
Printer
Demo 9/96
Nav
MSC
JTIDS
Etc
28
Fibre Channel LAN

• Switched Network
• Class 1 Mode
• Supports Real-Time Data Transfer
• Evaluated ATM, GB Ethernet, FDDI

29
U.S. Step 1 Modifications
n
t
S
i
t
u
a
t
i
o
n

D
i
s
p
l
a
y

C
o
n
s
o
l
e
M
o
d

K
i
t

I
n
s
t
a
l
l
a
t
i
o
n
M
i
s
s
i
o
n

S
y
s
t
e
m

S
o
f
t
w
a
r
e
M
o
d
i
f
i
c
a
t
i
o
n
s
E
-
2
1

C
o
m
p
u
t
e
r

C
a
b
i
n
e
t
30
The Past and the FutureSoftware
The Past
The Future
POSIX, CORBA,ODMG-93
No Commercial Standards

• Obsolete
• Large Integration
• Central
• Proprietary
• Fragile
• Inflexible
• Quality tested-in
• Unpredictable operation

• State-of-the-art, Object Oriented
• Plug-and-play
• Distributed
• Industry Standard
• Robust
• Flexible
• Quality designed-in
• Guaranteed performance

31
Traditional Real-Time Software Architecture

• Modular design
• Many data timing dependencies
• Timing constraints satisfied by
• optimization (via Assembly Instr)
• Schedules application processes
• Allocates storage other resources
• Dispatches tasks
• Precise, hard-wired timing Brute Force
• Handles interrupts
• Provides services
• Manages h/w configuration
• Detects faults, recovers

Applications
standard interface
Executive
standard interface
Operating System
- In some cases Executive OS are combined
(AWACS)
32
U.S. Step 1 Software Architecture

• Reduce Cost Time to Delivery (Dev, Integ
  Test)
• Reduce Complexity of Code and Project Management
• Reduce Learning Curve for S/W Designers
• Leverage Commercial/Industry Advancement in
  Timely Manner (COTS Open Systems)
• Adhere to Industry Standards (CORBA, ODMG)
• Ensure Real-Time Capability (Determinism)
• Distributed S/W Infrastructure (DSI) ES RTDB
  IM

Application Software
Information Manager
Object Request Broker (ORB)
CORBA Compliant
Open Interface
Encapsulated
Real-Time
Developed Software
Application Software
Encap Scheduler (ES)
Info Manager (IM)
Funct 2
Funct 3
Funct 4
Funct 1
RTDB
ORB
COTS Real-Time POSIX Operating System (e.g..,
LynxOS)
COTS Hardware
Operating System
Encapsulated Scheduler
Real-Time Data Base

• Temporal Data
• Some Persistent Data
• CORBA I/F
• ODMG 93, Object-Based

• COTS
• Real-Time
• Adheres to 3Ps

33
Distributed Software Architecture

• Distributed Software
• OO DataBase
• Functionality
• Common Middleware
• Dissimilar H/W Possible
• Dissimilar OS Possible
• Demonstrated in Dec 97

34
The Three Ps of Real-Time OS

• Predictability
• Predictable response times under all load
  conditions
• Preemptive Control
• Tasks may be preempted at any time
• Applies to Applications and OS
• Priority Based Scheduling
• Tasks are queued and executed according to
  priorities

Three Ps at Application and OS Level
35
Explicit Scheduling
Preemption
Predictability
JTIDS
P1
Tracker
Priority
P2
Printing
P3
Idle
Explicit Resource Management Analytically Proves
Schedulability
36
Real-Time SchedulingExplicit Resource Management
The Past
The Future
JTIDS
Tracker
Cyclic Executive
Printing
Unmanaged Resources
Explicitly Managed Resources
Hope for success
Analyze to ensure success

• Reliability Designed in
• Analytical/Guaranteed
• Contention removed
• Collisions avoided
• CPU responds to load
• Overloads explicitly handled
• Quick, cheap, simple to change
• Predictable
• Failure modes known and handled

• Reliability Tested-in
• Ad Hoc/Artistic
• Bus Errors/Contention
• LAN Collisions/Overload
• CPU Crashes
• Memory Errors
• Expensive to change
• Unpredictable
• Failure modes unknown

37
Modern Real-Time Scheduling Techniques
Reduce cost, schedule, and risk by managing
complexity

• Eliminate testing/complexity resulting from
  temporal or resource dependencies
• Formal theorems prove determinism
• Rate Monotonic Scheduling, Earliest Deadline
  First, etc.
• But formal, explicit scheduling techniques
  require
• Priority -- tasks are blocked only by higher
  priority tasks
• Preemption -- higher priority tasks run as soon
  as they are ready
• Predictability -- task resource utilization is
  deterministic or controlled such that it is
  deterministic

With the modern scheduling techniques, we can
reduce cost, schedule, and risk by managing
complexity
38
U.S. Step 1 Schedule Major Milestones
RDMX Concept Demo Sep 95 IM ES Demo Jun
96 IM ES Del 1 Dec 96 System Demo 1 Dec
96 System Demo 2 Sep 97 TS-3 Install/Ground
Checks Apr-May 98 TS-3 Flight Tests Jun
98-Jun 99 1st Wing Delivery Jul 99
39
Outline

• The Future of AWACS and the Need to Modernize
• Modernization Strategy
• Lessons Learned and Future Challenges
• Real-Time DII COE
• Summary

40
Lessons Learned

• Unanticipated S/W and H/W Integration Problems
• Plug Play not Realistic
• Software not Readily Available
• LynxOS Ver 4 - Ver 5
• Octegra - Cetia Interface
• RDMXP, STCP
• Cross Development
• Development vs Target Environmnet
• Limited Tools (SunSPARC to PowerPC, Ada)
• Data Tap (Legacy Interface) Development more
  Difficult than Anticipated

41
Lessons Learned

• Scope of Restart/Recovery S/W Design not well
  Understood
• Distributed Client-Server Architecture
• Scope Underestimated
• Commitments not met by COTS Suppliers
• Cetia Single Board Computer Delivery Schedule
• Quality of COTS Systran Cetia Conformal
  Coating, Cetia Design
• Switch of GLMs on Ancor LAN Card
• Convection vs Conduction Cooling
• Careful Selection of COTS Vendors
• Some Experience with Mil Products Helpful

42
Lessons Learned

• LAN Qual
• More Testing/Retesting than Anticipated
• Lack of LAN Drivers
• Late Development Fibre Channel Class 1
• Mislead/Misunderstood Reference Platform
• We Thought Level 1- LynxOS developed on Cetia
  SBC with In-House Personnel
• Reality Level 2 - OS developed, then
  Certified on Cetia SBC

43
Lessons Learned

• Late Team Selection of Baseline H/W S/W
• July 96 ? Feb 97
• LynxOS vs Solaris, Sun SPARC vs PowerPC
• Software Selection drove Hardware Decision
• Real-Time Support
• Commonality among AWACS Customers
• Lack of Good Benchmark Data (New Product
  Releases)
• Continual Search for COTS Solutions
• Power Supply Custom vs Brandt
• E-23 Cabinet Custom vs Zero Corp

44
Future Challenges

• Refine Real-Time Object Databases
• Communication Links
• Object Modeling and Simulation
• Real-Time Data Availability
• Commercial Real-Time ORBs
• COTS
• Hardware
• Software

45
Outline

• The Future of AWACS and the Need to Modernize
• Modernization Strategy
• Lessons Learned and Future Challenges
• Real-Time DII COE
• Summary

46
Defense Information Infrastructure Common
Operating Environment (DII COE)
COE Based Systems
GCSS
Other
GCCS
Other
DII COE
Standard Application Programs
COE Components
Reusable Software
Operating System Services
Hardware Platform
47
Real-Time DII COE Vision


ABL AEGIS/ATHENA AWACS Crusader MCE JSTARS R/SAOC
THAAD
Provide Warfighter with Capable, Interoperable,
Cost Effective Systems

V6.0
V5.0
DII / COE - RT CORBA - RT ODMG-93
V4.0
DII / COE - RT OS
V3.0
DII / COE - CORBA - ODMG-93 - JVM
DII / COE
Flexible -Rapid development and introduction
Affordable - Reduced cost of acquisition and
ownership Common Open Foundation - Built on
open standards and COTS
48
AWACS Segments - U.S. Step 1

• Incremental Migration to DII COE
• Step 1 Bridge Legacy Code
• New Code Compliant Segments
• Step 2 All Compliant Segments
• Real-Time DII COE Extensions Required
• Excellent Example of Migration Strategy
• JTA
• DII COE

Compliant Segments
Maint
DLI
Tracker
COE
AWACS DSI
COE Core Software Utilities
RT CORBA IM, RT Scheduler, RT Database
OS (COTS Real-Time UNIX)
49
AWACS Segments - U.S. Step 2
Compliant Segments
System Config
Map Server
Maint Programs
Simulation
Intel
Maint
COE
COE
COE
COE
COE
COE
Tracker/MSI
AMCP
WSCP
Battle Mgt
Comm
Comm
JTIDS
JTIDS
Comm
COE
COE
COE
COE
COE
COE
COE
COE
COE Core Software Utilities
RT CORBA IM, RT Scheduler, RT Database
OS ( COTS Real-Time UNIX)
AWACS Applications
Standard Applications
50
RT DII COE Motivation

• DII COE does not currently support real-time
  computing
• Inclusion of RT components into DII COE is moving
  too slowly to support migrating programs
• AWACS
• R/SAOC
• THAAD
• Crusader
• ATHENA
• Due to AWACSs Experience in Real-Time
  Distributed Object Technology and Implementation
  of COTS H/W and S/W, AWACS selected as AF Lead
  Agent in RT DII COE

51
Outline

• The Future of AWACS and the Need to Modernize
• Modernization Strategy
• Lessons Learned and Future Challenges
• Real-Time DII COE
• Summary

52
Summary

• AWACS Must Modernize
• AWACS must change the way we do Business
• Spiral Development
• Open Systems, use of COTS Hardware and Software
• Object Technology is an Enabler
• Drives Down Integration Test Cost
• Provides Future Growth
• Development
• Real-Time Techniques as Basis of Design
  Methodology
• AWACS Migrating to Real-Time DII COE
• Provides Opportunity for Cross Platform Data Link
  (and other) Applications