Score Team Presentation

Transcript

1. From Score to Coast From Cooperative, Distributed, and Process-Aware Web Information Systems to Web Scale Trustworthy Collaborative Information Systems
2. Ecoo : the origin • Pour développer du logiciel – Partager des données – Coordonner des activités • Domaines de recherche – Edition Collaborative – Gestion des processus métiers – Transactions de longue durée – Services
3. Support for Building Pyramids • Large scale collaboration is required to solve big problems and challenges • Collaboration of people and services • Requires confidence (trust+control)
4. Where do we stand IT Layer Business Layer Design time RunTime Infrastructure Layer Users
5. An original tension • Implicit Coordination – Data sharing – Group Awareness • Explicit coordination – Process modeling – Control of the execution A common view : Enforcing properties on data sharing and coordination
6. Results • New algorithms for optimistic replication • Collaborative systems • Inter-organisation and cooperative workflow systems • Models for safe service compositions • Software/Platform – Bonita, Libresource, Qualipso Factory, DSMW, Etc.
7. Collaborative Data Management • How to support people working together, distributed in time, space and organisations • Collaborate without a central authority IT Layer Optimistic replication Business Layer Collaborative Editing
8. Optimistic Replication • Investigate and evaluate new algorithms – Operational Transformation – WooT (Without Operational Transformation) – CRDT • Support for decentralized « Real Time » Collaborative Systems • Understand their behaviour
9. Experiment algorithms • How do they behave – With real users, – With a lot of users – With different kind of scenarios – With different kind of deployment Evaluating CRDTs for Real-time Document EditingAhmed-Nacer Mehdi; Ignat Claudia- Lavinia ; Oster Gérald ; Roh Hyun-Gul; Urso Pascal inria-00629503ACM. 11th ACM Symposium on Document Engineering, Sep 2011
10. First Results • Executing the same input with different algorithms • But we need more input Evaluating CRDTs for Real-time Document EditingAhmed-Nacer Mehdi; Ignat Claudia- Lavinia ; Oster Gérald ; Roh Hyun-Gul; Urso Pascal inria-00629503ACM. 11th ACM Symposium on Document Engineering, Sep 2011
11. Awareness & Trust • Concurrent editing generate conflicts • Important issues – How to detect conflicts – How to solve conflicts – How to avoid conflicts – How to respect privacy – How to enforce privacy rules
12. Optimistic data sharing control • Explicit sharing contract • Explicit Trust representation • Audit log against the contract • Update trust value based on audit Hien Thi Thu Truong, Claudia-Lavinia Ignat, Mohamed-Rafik Bouguelia and Pascal Molli. A Contract-extended Push-Pull-Clone Model. CollaborateCom 2011
13. Coordination in dynamic situation Application to crisis management • Issues – Changing goals – No central authority – Privacy/governance issues • Proposal – A Dynamic activity model – Design time = Execution time – Temporal dependencies – P2p Sharing between organisations
14. Service Composition • Development of applications at the scale of the web • How to develop services that are – Efficient, Safe, Secure, Flexible, Adaptative • How to ensure/enforce some properties on service composition – Transactional properties, Security properties, Privacy/Quality of service From programmableweb.com
15. Engineering Service Composition • Designing the composition (What, How) • Defining the contracts • Verifying the contracts • Deploying/executing • Monitoring/Auditing • Responding to contract violation IT Layer Deployment/Execution Business Layer Modeling/Specifying/Checking
16. Declarative Service Composition Composition Design Composition Monitoring Composition Implementatio n & Validation Composition Execution Ehtesham Zahoor, Olivier Perrin, Claude Godart: An Event-Based Reasoning Approach to Web Services Monitoring. ICWS 2011: 628-635 Closing the loop is a challenge
17. Our landscape evolves quickly • The web as a platform – To collaborate – To produce and share data – To produce and deliver services • Elastic Computing – P2P – Cloud Infrastructure
18. The scale is changing • Regarding data production – Wikipedia – Climate Change Report • Regarding coordination – Gestion de crise • For both – OSS
19. Self services are coming
20. Heterogeneity is back again
21. New ecosystems have appeared • Smartphone and tablet apps • They are based on services (REST API)
22. The Coast Project Collaborative data management Data centric service composition Trustworthy collaborative systems
23. Collaborative data management • Large scale data sharing • Collaborative editing • Real time p2p collaborative editing • Conflict management
24. Problems • How to ensure eventual consistency with optimistic replication ? • How to avoid conflicts when documents are edited concurrently • How to handle conflicts efficiently ? • Up to what scale ?
25. Our approach • Develop new optimistic replication algorithms – For new data types – For new scales • Study group awareness mechanisms – For communities – For heterogeneous settings
26. Problems • How to build applications through service assembly • How to represent and validate – Terms of Service (that should be enforced) – Service Level Agreement (the promise) • What kind of properties can we ensure on service composition • How to govern their evolution
27. Our approach • Define models to describe Service ToS and SLA that allow to create – Safe composition – Certified composition • Define and validate method for service based software engineering • Incorporate the economic dimension of service delivery (probably easier)
28. Trustworthy Collaborative Systems • Collaborative system : composition of collaborative systems • Define the confidence attributes – Trust and/or Control – Certification • Confidence is a relation – Service/Service – User/User – User/Service
29. An approach grounded on contract • Confidence contracts in collaborative environments – Between participants – Between services – Between users and the environment • Validation of contracts and of their composition • A posteriori contracts verification
30. Scientific challenge • Definition and validation of confidence attributes for – Data sharing – Collaboration – Service composition • Real time or deferred control Optimistic approach for safety and security
31. Methods • Theory – Algorithms and models • Experience – We need to invent protocols • That scale • That demonstrate usability • Design Science – We invent and validate artifacts based on research hypothesis
32. Applications • Software development environments – Cloud based forge as • A composition of services for collaboration • A support for trusted composition of trusted services • Large scale multi organisational cooperative environment Cloud9Ide
33. Valorisation • The Cloud is an opportunity – Iterative and incremental approach • Available platform to integrate and experiment our services – Collaborative editing. – Context and awareness management. – Coordination and data services. – Service contract management.
34. Questions ?
35. Trust (Ben Schneiderman) • Guideline 1.1. Disclose patterns of past performance. • Guideline 1.2. Provide references from past and current users. • Guideline 1.3. Get certifications from third parties. • Guideline 1.4. Make it easy to locate, read, and enforce policies involving privacy and security. • Guideline 2.1. Clarify each participant’s responsibilities. • Guideline 2.2. Provide clear guarantees with compensation. • Guideline 2.3. Support dispute resolution and mediation services.