The tutorial introduces the main principles behind GRID computing and the paradigm shift in approaching the use of services. The second parts introduces two particular notions on services; the focus of the discourse is then on a particular set of services governed by partiality and intermittence. A service is called partial when only some post-conditions are satisfied, while an intermittent or sporadic service guarantees all post-conditions for a limited time interval. This is mainly due to the volatile aspect of available resources and continuously changing users' profiles. General topics discussed upon are commonly known or dedicated mechanisms to implement these principles. While the tutorial is a mix between the state of the art and the challenging open issues, the details follow the topics pertaining to various facets of the SLA/QoS, as apparently these notions must be revisited. The third part emphasizes the potential change in defining and interpreting measurements on particularly related network components or/and network activities. The following facets of the presented service paradigms will be analyzed; for some items more challenging questions rather than existing solutions will be debated.

• Partially and intermittently available services,
• Partially and intermittently available resources
• Service Level agreement for partial and intermittent available behaviors
• Adaptive Time-out Discovery mechanisms for intermittent services
• Hybrid sequential and parallel processing in intermittently available services
• Sporadic data transfer services
• Semantic of high availability for partially available and intermittent services
• Adaptive applications to partially and intermittently available services
• Security for intermittently available and partial services
• User behavioral models for partial and intermittently offered services
• Implications on current communication protocols, security, and middleware
• Implications on service assurance, QoS, and traffic modeling

The tutorial will conclude on the ongoing work related to these type of services and the potential obstacles in reusing the already known similar SLA/QoS mechanisms use din traditional network services.

Bio: Petre is a Senior Technical Leader with Cisco Systems, Inc., being responsible for policy-based strategic architectures and protocols for network management, QoS, SLA, and Performance, Programmable Networks and Services, Provisioning under QoS constraints, and Consistent Service Manageability. He's industrial research interests include mobile systems, performance, scalability, and policy-related issues in GRID networks. He's >also working on particular issues in multimedia systems concerning traffic patterns and security. He worked on various industrial applications including CAD/CAM, nuclear plant monitoring, and real-time embedded software. In early 90's he worked on various Pan-Canadian projects related to object-oriented management applications for distributed systems, and to broadband services in multimedia applications. As a Researcher at the Computer Science Research Institute of Montreal he coordinated many projects on distributed software and management architectures. In this period he was an Adjunct Professor with McGill University, Montreal, Canada, and a Canadian representative in the European projects. Since 1998 until 2000 he was with AT&T Labs, as a senior technical manager, focusing on distributed QoS, SLA, and Performance in content delivery services.

He is actively involved in the innovative NGOSS industrial initiative in TeleManagement Forum. Petre is also a Rapporteur in Study Group 4 at ITU-T. He has been an invited speaker to many international conferences and presented tutorials; he chaired several international conferences and published many technical papers.

He is currently an Adjunct Professor at Concordia University, Montreal, Canada, a Senior IEEE member, and an ACM member.