Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Viva slides_secured objective programming
1. Secured Objective Programming
Support to Intention Driven
Autonomic
Cloud Computing
Yasir A. Karam
of Liverpool John Moores University for
the degree of
Doctor of Philosophy
8-1-2014
2. Contents
• Introduction to Goal – Actor Model
– Goal – Actor Model
– Goal - Actor Society
– Call Dispatching in Dynamic Actor Model
• Research Problems & Resolutions
– Annotating i* Goal Modelling Method over XML Intentions
– Neptune Architecture with Automated Planning Support
– Axioms Used as Predicate Propositions
• Examples of Problem Domain Description
– Adding Formal Logical Representation of XACML over Neptune using
PAA + CA-SPA
– Domain Description Language for XACML Problem
• Publications
– Results from published work
4. Research Problems & Resolutions in Snapshot
Modeling Secured Interoperable Architecture based on Capability
Security Model and SOA
Modelled XACML using PAA, CA-SPA
Modeling problem domain for (a)
Used STRIP style problem domain description language PDDL to write
declaratives for goal propositions.
Used Temporal Logic Planning as automated reasoning engine to solve
PDDL problems
Search algorithm that fit graph model representation of the problem.
Several problem types were anticipated such as search heuristic
problem
Search algorithms used is breadth-first.
All the above support was added and implemented inside Neptune
Architecture
Problem I
Problem I
5. • Adding Goal Modeling support to Intention Model so that Goal
expressions that fit best known goal modeling methods like i*, GRE,
Tropos.
– Goal modeling artifacts were annotated over Intention XML style
language
• Writing LTL specification to support Goal reasoning model using
existing Neptune Scripting Language, PAA and CA-SPA
– This task was solved through representation of LTL specifications for
goal modeling using STRIP style problem domain description
language., CA-SPA and PAA
• Modeling of qualitative and quantitative capacities of goal
modeling such that ++, -- are goal satisfiability axioms (soft-goals) to
be used to test Assurance of how much is been achieved.
– A stochastic aggregative model like Basian networks and MDP
(Markov Decision Process model) was approached and analyzed (this
task was under progression stage of 35% )
– A subsequent is anticipation for optimal points from non-determinism
problem
Problem II
Problem II
Sub-
Problem
Sub-
Problem
6. • Re-designed Neptune (all the language ) over
distributed concurrent logic programming
paradigm, new enhanced features such as
asynchronous call dispatching and others.
• The use of “shared memory” is more effectively
to comprehend in carrying immune state
characteristics.
• This helped us to add new language support to
model Team Object Model and team guards (this
is completed feature) in which is published in
author’s paper.
Problems
cont.
Problems
cont.
The below problem was tackled optionally as
an engineering effort needed to enhance
legacies in Neptune Architecture
7. • Re-designed Neptune (all the language ) over
distributed constrained logic programming
paradigm, new enhanced features such as
asynchronous call dispatching and others.
• The use of “shared memory” is more effectively
comprehend in carrying immunized state
characteristics.
• This helped us to add new language support to
model Team Object Model and team guards (this
is completed feature) in which is published in one
of authors papers.
Problem III
Problem III The below problem was tackled optionally as
an engineering effort needed to enhance
legacies in Neptune Architecture
9. Research Hypothesis
• Identification, representation, expressing and classifying new type of
atoms that support Goal Oriented Requirements,
• With the aid of existed structured atoms of Intention Model, what we did
is testing socio-communal interaction between multiple intention models
and how to use distributed objectives/goals to identify recognition and
resolution areas .
• Defining, design and implement Reasoning Model for Neptune
architecture, which adds capabilities of writing aided constructs of
Strategies, Plans, Schedules Tasks and also Objective oriented attributes
like objectives type, rules, situations and fluent’s
• Model of performance based attributes like KPI objects, cost,
performance targets, weights this is based on Capability Driven Actor
Model (CDAM).
• Define and model elements for Capability Actor Architecture like
modeling of No cost will be paid unless farthest goal is evaluated
• Value based dispatching through introspective delegation (exchange of
accountability)
• Planned invocation through stages between Early and Late binding
19. Contribution to Knowledge
• Adding constructs to existing “Intention Model” that helps actors to
specify “preference” or priority to their intentions, this will be
compiled and linked with right composition level NBLO’s (High
NBLO’s) that will in turn be used to constrain decomposing big
computational coarse grained problems into smaller ones.
• Adding support for concurrent transaction modelling to Neptune
model using management of “shared memory” - Adding support for
autonomic social behavior to runtime actors though dynamic
adaptation of goal oriented requirements from intention model.
• Using PAA style of modelling advices, we provided support for
writing dynamic objective model to “Linear Temporal Logic” and
use metricized constructed objects with the aid of CA-SPA in
providing support to model “Propositional Logic”
• The support for LTL makes it then easy to solve problems of
situational predicates used to achieve goal modalities “achieve”,
“avoid”, “maintain” and “avoid&maintain”.
20. Annotating i* Goal Modelling Method over XML Intentions
Actors annotation over Intention
Goal SD Constructs
Intention Description
22. Cont.
• On the other hand we used CA-SPA policies to design formalities for
satisfiability properties.
• Following formalisms above, we provided support to write STRIP style “Domain
Specifications” and “Problem Domain Specifications”. The way writing
specifications used is similar to PDDL language “Planning Domain Description
Language”
• In order to ensure better “Satisfiability Propagation” or “Value Proposition”
between socially networked actors and objects, we provided code design
support metric softgoal objects that used to measure quality of provisioning for
problem main objects in order to achieve new hard goal states, this is by using
PAA, Accounting and Auditing with CA-SPA, to count the number of violations
to formal constrains.
• We provided new design support to design patterns through situating and
fulfilling of object role based requirements to new injected “concepts” at
design and compile time. This is proved with the aid of “dynamic
polymorphisms” and “object teams architecture”
• Automation support for the above features is added to assist in performing
feasible design decision dynamically, this through using multiple solvers
depending of problem type.