Autonomic Web Processes

1
Autonomic Web Processes

• Presenter Amit Sheth
• METEOR-S project, LSDIS Lab
• Computer Science, University of Georgia
• Presentation of the Vision Paper (Invited)
• Kunal Verma and Amit Sheth. Autonomic Web
  Processes. In Proceedings of the Third
  International Conference on Service-oriented
  Computing (ICSOC 2005), - Vision Paper (invited),
  LNCS 3826, Springer Verlag, 2005, pp. 1-11.

2
Introduction

• Growing need for creating more adaptive/dynamic
  process frameworks
• IBMs vision of autonomic computing lays
  foundation of adaptive/self managing systems
• Our vision seeks to elevate Autonomic Web
  Processes from the infrastructure to the process
  level

http//www.research.ibm.com/autonomic/
3
Autonomic Nervous System

• Responsible for maintaining constant internal
  environment of human body by controlling
  involuntary functions like
• digestion, respiration, perspiration, and
  metabolism
• Divided into two subsystems
• Sympathetic and parasympathetic

http//www.nda.ox.ac.uk/wfsa/html/u05/u05_010.htm
4
Autonomic Nervous System

• Sympathetic
• providing responses and energy needed to cope
  with stressful situations such as fear or
  extremes of physical activity
• Increases blood pressure, heart rate, and the
  blood supply to the skeletal muscles at the
  expense of the gastrointestinal tract, kidneys,
  and skin
• Parasympathetic
• Brings normalcy in between stressful periods
• which lowers the heart rate and blood pressure,
  diverts blood back to the skin and the
  gastrointestinal tract

5
An Example
http//www.sirinet.net/jgjohnso/nervous.html
6
Autonomic Computing

• Autonomic Computing is an initiative started by
  IBM in 2001
• Aims to make systems that simulate the autonomic
  nervous system by having the ability to be more
  self managing
• Objective to let user specify high level policies
  and then the system should be able to manage
  itself

7
Autonomic Computing - properties

• Infrastructural Components with Self-CHOP
  properties
• Self Configuring
• Self Healing
• Self Optimizing
• Self Protecting
• Examples
• Self Adaptive Middleware
• Self Healing Databases
• Autonomic Server Monitoring

8
Autonomic Web Processes (AWPs)

• Natural Evolution of Autonomic Computing from
  infrastructure to Web process level
• Web processes are Web services based workflows
• Require Web process frameworks that have the
  following properties
• Support Self-CHOP properties
• Policy based interaction with other components
• Based on open standards (WS technologies)
• Based on the synergy between a number of broad
  fields
• Autonomic Computing, Web Services, Service
  Oriented Architectures, Operations Research,
  Control Theory, Semantic Web, Dynamic and
  Adaptive Web Processes/Workflows

9
Use Case

• Supply Chain of computer manufacturer
• Self Configuring Can the process be configured
  based on constraints and policies
• Self Healing Can the process recover from
  physical and logical failures
• Self Optimizing Can the process reconfigure
  itself in case of changes in environment.

10
Architecture
11
Self Configuring

• Depending on the scope, configuration may include
• Creation of process (manual/semi-automatic/plannin
  g)
• Discovery of partners (internal/external
  registry)
• Negotiation (manual/automated)
• Constraint Analysis (quantitative/logical/hybrid)
• Require representation of
• Functional semantics for discovery
• Non-functional semantics for constraint analysis
  constraints, policies, SLAs

12
Self Configuring
Configured Process
Constraint based Configuration
13
Self Healing

• Process must be able to recover from
• Failures of physical components like services,
  processes, network
• Logical failures like violation of SLA
  constraints/Agreements
• Delay in delivery, partial fulfillment of order
• Require representation of execution semantics
• Physical and Logical Exceptions and recovery paths

14
Self Healing Creating Execution Graph of a SM
Actions
Events
Flags
15
Self Healing
Execution Graph- Generated from Operations,
Events and Flags 5 Flags, thus 25 32 possible
states (only 8 reachable states)
S1- Ordered True (All other flags false) S4 -
Ordered True and Received false S5-Ordered
True and Delayed false ---Transition due to
action - - Exogenous events (example
probabilities of occurrence of the events
conditioned on the states)

• One proposed approach Use Markov Decision
  Processes to choose optimal actions

K. Verma, P. Doshi, K. Gomadam, J. Miller, A.
Sheth, Optimal Adaptation in Autonomic Web
Processes with Inter-Service Dependencies,  
LSDIS Lab, Technical Report, November 2005
16
Self Optimizing

• Process must be able to reconfigure itself with
  changes in environment
• Fluctuations in currency exchange rates of
  overseas suppliers
• New discounts or cheaper suppliers available
• Must choose between long term and short term
  benefits
• This requires both functional and non-functional
  semantics

17
Self Optimizing
Change in Currency Rate beyond threshold
Sympathethic Policy Reconfigure process for
immediate gain May including canceling order from
previous Supplier
Parasympathethic Policy Consider long term
supplier relationship
18
Model

• Functional and Data Semantics
• Service (WSDL-S)1
• Non-Functional Semantics
• Policies (Semantically Annotated Policy)2
• Business Level Policies, Process Level Policies,
  Instance Level Policies Individual Component
  Level Policy
• Agreements (SWAPS) 3
• Execution Semantics
• State based representation of exceptions/failures
  
• Process (BPEL Semantic Templates) 4
• Ontologies
• Domain Specific Ontologies,
• Domain Independent/Upper Ontologies

AWP Property/Type of Semantics Self Configuring Self Healing Self Optimizing
Data
Functional
Non-Functional
Execution
1 Web Service Semantics WSDL-S, W3C Member
Submission., http//www.w3.org/Submission/WSDL-S/
2 K. Verma, R. Akkiraju, R. Goodwin, Semantic
matching of Web service policies, SDWP, 2005 3
N. Oldham, K. Verma, A. Sheth, Semantic
WS-Agreement Partner Selection http//lsdis.cs.uga
.edu/projects/meteor-s/swaps/ 4 K.
Sivashanmugam, J. Miller, A. Sheth, and K. Verma,
Framework for Semantic Web Process Composition,
IJEC, 2004
19
AWPs vs. Autonomic Computing
20
Conclusions

• The Vision
• AWPs seek to create next generation of Web
  process technology
• Current Work
• Initial work at UGA on using MDPs for adaptation
• IBM work on WSDM for autonomic Web services
• Paolo Traverso et al. - Autonomic Composition of
  Business Processes
• The Future
• We invite researchers from SOA, Web services, AI,
  multi-agents, operations research, control theory
  to contribute to this vision
• Dagstuhl-Seminar Autonomic Web Services and
  Processes (possibly in August 2006) Contact
  Paolo Traverso, Amit Sheth