Service Oriented Architecture

1
SERVICE ORIENTED ARCHITECTURE
George F. Hurlburt Change Index Inc.
301-994-3305
2
SERVICE ORIENTED ARCHITECTURE
OVERVIEW
WHAT SOA IS NOT WHAT SOA MUST BE AND
WHY ACHIEVEMENT OF SOA
3
SERVICE ORIENTED ARCHITECTURE
- NOT part of Department of Defense Architectural
Framework (DoDAF)
Architectur
Model
Weak Affinity
Strong Affinity
SOA
DoDAF
CONTEXT DoDAF
4
SERVICE ORIENTED ARCHITECTURE
- NOT part of Department of Defense Architectural
Framework (DoDAF)
MATRIX
TV-1
Architectur
Model
SV-1
SV-6
OV-6
OV-5
DoDAF
SOA
No Significant Affinity
(True with all AV, OV, SV and TV)
CONTEXT OV-3
5
SERVICE ORIENTED ARCHITECTURE
- NOT the same as Web Services
The Enterprise
RUNTIME MANAGEMENT
ENTERPRISE SERVICE BUS
DESIGN-TIME GOVERNANCE
business process management
application front-ends
complex event processing
SECURITY GATEWAY
web services description language (WSDL)
rich internet applications
representative state transfer (REST)
(W3C) Web Services -
The Internet
6
SERVICE ORIENTED ARCHITECTURE

• NOT more Middleware

Scattered Capability Suites
Tools Added Overhead Software
Hardware (AFE) Cost to scale Based on
XML (Custom) Database (Rigid) API
Mapping
7
SERVICE ORIENTED ARCHITECTURE

• NOT a Top-Down thing awash in a sea of standards

MAINSTREAM
MANDATED PRACTICES
ADA HLA XML
SOAP
BPEL4People
REST
WSDL
BPEL
HTTPS
UDDI
JMS
WS-
HTTP
JSON
WEB WEB 2.0 SOA
CRITERION OF PRACTICAL VALUE
8
SERVICE ORIENTED ARCHITECTURE

• NOT a static phenomena

9
SERVICE ORIENTED ARCHITECTURE

• NOT easy to do in theater

XML
XML
XML
Central PKI Authentication Access

Verification Encryption
Key
SERVICE Discovery

reach back to CONUS
XML
10
SERVICE ORIENTED ARCHITECTURE

• NOT easy to do in theater

parsed
parsed
XML
parsed
XML
XML
Central PKI Authentication Access

Verification Encryption
Key
SERVICE Discovery

reach back to CONUS
XML
11
SERVICE ORIENTED ARCHITECTURE

• NOT easy to do in theater

parsed
OVERHEAD
parsed
XML
parsed
XML
XML
Central PKI Authentication Access

Verification Encryption
Key
SERVICE Discovery

reach back to CONUS
XML
12
SERVICE ORIENTED ARCHITECTURE
SOA IS NOT
SOA IS Connected! Dynamic! Complex! Adaptive!
Evolutionary!
13
SERVICE ORIENTED ARCHITECTURE
CONNECTED!
How many degrees of separation (links) are you
from The President? The Dalai Lama?
The US Oyster Festival Web Site? A
core service?
14
SERVICE ORIENTED ARCHITECTURE
CONNECTED!
How many degrees of separation (links) are you
from The President? Average 4-6
Links The Dalai Lama? Average 6
Links The US Oyster Festival Web Site?
Average 13 Links A core service? Its a
small world after all!
15
SERVICE ORIENTED ARCHITECTURE
DYNAMIC!
APPLICATION
OUTPUT (e.g. Temperature)

INPUT (e.g. Location)
API
16
SERVICE ORIENTED ARCHITECTURE
DYNAMIC!
APPLICATION
OUTPUT (e.g. Temperature)

INPUT (e.g. Location)
Washington Washington D.C. Washington,
DC District of Columbia DC
API VARIATION 1. Semantic I/O Variation

17
SERVICE ORIENTED ARCHITECTURE
DYNAMIC!
APPLICATION FUNCTION Gettemperature
OUTPUT (e.g. Temperature)
Degrees F
APPLICATION FUNCTION Determinehumature
INPUT (e.g. temperature)
Degrees C
INPUT (e.g. Humidity)
OUTPUT (e.g. Humature)
INPUT (e.g. Location)
API VARIATION 1. Semantic I/O Variation
2. Parameter I/O Variation
18
SERVICE ORIENTED ARCHITECTURE
DYNAMIC!
OUTPUT (e.g. Temperature)
INPUT (e.g. temperature)
INPUT (e.g. Wind Speed)
OUTPUT (e.g. Wind Chill)
INPUT (e.g. Location)
API VARIATION 1. Semantic I/O Variation
2. Parameter I/O Variation 3.
Functional Metadata
19
SERVICE ORIENTED ARCHITECTURE
Process definition 1
DYNAMIC!
Process definition 2
APPLICATION 1
APPLICATION 2
How is the overlap defined?
APPLICATION 3
API VARIATION (n-Dimensional) 1. Semantic
I/O Variation 2. Parameter I/O Variation
3. Functional Semantics 4. Business
Process Metadata
20
SERVICE ORIENTED ARCHITECTURE
COMPLEX!
APPLICATION 1
APPLICATION 2
COMBINED SERVICES
Service B
Service A
Service C
REQUIRED SERVICES
APPLICATION 3
HIERARCHIAL SERVICES
Pre-Req Service D
Sub- Service C
21
SERVICE ORIENTED ARCHITECTURE
ADAPTIVE!
UNREALISTIC TAXONOMIC VIEWPOINT
FUNCTIONAL HIERARCHY
SERVICES HIERARCHY
22
SERVICE ORIENTED ARCHITECTURE
ADAPTIVE!
ADAPTIVE ONTOLOGICAL
RELATIONSHIPS
FUNCTIONAL HIERARCHY
SERVICES HIERARCHY
23
SERVICE ORIENTED ARCHITECTURE
ADAPTIVE!
UNLIMITED OVERLAPPING
CONTEXTUAL RELATIONSHIPS
24
SERVICE ORIENTED ARCHITECTURE
EVOLUTIONARY!
EVOLUTIONARY ONTOLOGICAL
RELATIONSHIPS
FUNCTIONAL HIERARCHY
SERVICES HIERARCHY
25
SERVICE ORIENTED ARCHITECTURE
EVOLUTIONARY!
A NECESSARY PRECURSER TO THE CLOUD
26
SERVICE ORIENTED ARCHITECTURE
SOA IS NOT

SOA IS Connected! Dynamic! Complex! Adaptive!
Evolutionary!
SCALE FREE!
27
SERVICE ORIENTED ARCHITECTURE
SCALE FREE NET
Characteristics Contextual Viewpoint
(Essential) Force Multiplication
Self Organizing (can recover from losses)
Adaptive Complex Evolutionary
Diverse (Trans-disciplinary) Vulnerable
to coordinated attack Examples The
Internet Human Social Structures
The National Power Grid A Bee
Hive/Termite Mound Any corporate
INTRANET The Human Brain
Semantics
circa 1999
28
SERVICE ORIENTED ARCHITECTURE
SCALE FREE NET
Semantic Example Hub Connectivity Hub
Nodes contain a large number of links Most
nodes contain only a few links Any node is
a few hops from any other node Implicit Layered
Semantic Relationships Document Semantic
Network Network of Documents (Domain)
Network Node Locations Many to many
relationships exist Pattern of usage
Domain/Context specific Evolves with new
documents Encompasses Idiosyncratic
usage
29
SERVICE ORIENTED ARCHITECTURE
30
SERVICE ORIENTED ARCHITECTURE
SOA IS Connected! Small World and Scale
Free Dynamic! n-Dimensional Semantics Complex!
Multiple Logical/Lexical Relationships Adaptiv
e! Living Ontology/Context Evolutionary! Not
static over time Efficient Achievement of
SOA Attack SOA at the Semantic Level Change
Index Inc. Solution
31
SERVICE ORIENTED ARCHITECTURE
SCALE FREE NETS
Characteristics Contextual Viewpoint
(Essential) Force Multiplication
Self Organizing (recovers from partial)
Adaptive Complex Evolutionary
Diverse (Trans-disciplinary) Vulnerable
to coordinated attack Examples The
Internet Human Social Structures
The National Power Grid A Bee
Hive/Termite Mound Any corporate
INTRANET The Human Brain
Semantics
32
SERVICE ORIENTED ARCHITECTURE
Ingest NGRAMS Master Nodes Association Nodes
Context Term Processes Flow
Connections
AFFINITY ENGINE
.txt .doc .xls .ppt .htm .html .xml .rdf etc
DOMAIN-CENTRIC SMALL WORLD SEMANTIC NETWORK
WordNet (Optional)
33
SERVICE ORIENTED ARCHITECTURE
CONTEXT Context Word Context
Phrase Classifier GA Training
AFFINITY ENGINE
CONTEXTUALLY MODIFIED SEMANTIC NETWORK
DOMAIN-CENTRIC SMALL WORLD SEMANTIC NETWORK
34
SERVICE ORIENTED ARCHITECTURE
MODIFIERS Breadth Depth
Weight Associations
CONTEXTUAL MODIFIERS
Breadth
DOMAIN-CENTRIC SMALL WORLD SEMANTIC NETWORK
Depth
35
SERVICE ORIENTED ARCHITECTURE
CONVERGENCE Combine
Union Attach
DOMAIN CONVERGENCE
Combined (All Domains)
Union
DOMAIN-CENTRIC SMALL WORLD SEMANTIC NETWORKS
Attached (2 Domains)
36
SERVICE ORIENTED ARCHITECTURE
Data Representation XML/Schema
RDF/Taxonomy OWL/Ontology Mood Indicators
Attitude Mood
BACKEND APPLIANCES
CONTEXTUALLY MODIFIED SEMANTIC NETWORK
37
SERVICE ORIENTED ARCHITECTURE
Genetic Algorithms Multiple Agents Evolutionary
Tools
DOMAIN OPTIMIZATION
38
SERVICE ORIENTED ARCHITECTURE
SMALL World
Context and Modification
Backend
GA
Small World
Contextual Training
Converge
PROCESS DYNAMIC
39
SERVICE ORIENTED ARCHITECTURE
ATTACK SOA AT THE SEMANTIC LEVEL
Semantics are dynamic small world phenomena
Central Hubs connect many lesser nodes
Contextual affinity binds specific sets
Self generating schema/taxonomy/ontology
possible SOA is an applied semantics problem
API definition of I/O and function
Contextual views are necessary at all levels
Service/functional relationships are fuzzy at
best Semantics accommodates Dynamic SOA
Linguistic nuance (fuzziness) no longer an
issue The full range of Services are
available for discovery Novel new
service use can be exposed Context
frames ID Management
40
SERVICE ORIENTED ARCHITECTURE
CHANGE INDEX INC RESEARCH SOLUTION
Ingest all data salient to the SOA Network as
available Applications and
Specifications Use Cases
Pertinent documents and briefings Build small
world semantic network from chaotic
data Influence service harvesting with context
Direct insertion of limited context words
Powerful Classifiers for large concepts
Binary choice for Genetic Algorithm
training Produce optimized working service
ontology supporting Dynamic service
environment (eco-system) Dynamic service
discovery (service harvesting) Dynamic
service management Dynamic optimal
service evolution