| Description |
Scientific or industrial computing/data intensive applications for simulations and virtual prototyping may benefit of an on-demand basis from powerful computing or storage servers provided by external companies/service providers and remotely accessible via metropolitan or wide area networks. To enable a pervasive adoption of such an utility computing model, this should be supported by shared carrier networks by opposition to dedicated research networks. However, this raises two main issues, a business-related one and a technical-related one:
Are the concerned users ready to externalize their resource and to share their use on a converged network ? What would be the business relationship between application users, network operator and resource providers ?
What are the technical constraints on the networks resulting from this usage ? How do they vary from one application to the other ? Which kind of interactions between the applications and the network belonging to different administrative domains can exist to enforce the network service required by the distributed applications?
This workshop intends to address these issues, by discussing in particular the following points:
Interest and feasibility of the externalisation of computing and storage resources through a network, drivers and brakes for such an evolution.
Typology of distributed applications, factors of differentiation (number of sites and users, data volumes, sequence of application sessions, capacity to plan the phases, level of interactivity …)
Requirements on the quality of service at network level: constraints of bandwidth, latency, connectivity dynamics, security ….
The relevant contributions for this workshop are in particular among the following themes:
Description of practical experiences of distributed applications over long distance networks, performance assessment
Analysis of network requirements for specific applications in industrial or scientific domains, in areas such as aeronautics, automotive, energy, pharmaceutical, climatology, biology, physics, finance, etc.
Impact of network QoS (throughput, latency ….) over application quality of experiences: theoretical and/or experimental analysis
Scalability analysis of on-demand usage of distributed resources with respect to the network size, to the number of users, of applicative sessions, of resource servers, etc …
Application-network relationship: Overlay model versus integrated model
Co-allocation and orchestration of application and network resources, network (re)configuration driven by application needs
Security, trustability in utility computing
Market analysis and business models studies related to utility computing over converged networks
Associated service definition and service level agreements encompassing both network and computing/storage resources
|
|
