Network as a resource: G-lambda project and its architecture

In computing systems, hardware components such as CPUs, memories and storages are "resources". Resources are managed by resource managers (c.f. a queuing system of a computing center), and a user can use them while they are allocated to the user. To provide a stable service using "resources" connected to a wide area network, network itself should be also considered as a "resource" and bandwidth should be allocated to users appropriately. Resource is provided to a user or upper layer middleware by a resource manager. Therefore, to realize "network as a resource", bandwidth of network should be managed by a network resource manager, and an interface to provide resource service to users or middleware is required. National Institute of Advanced Industrial Science and Technologies (AIST), KDDI R&D Laboratories, NTT and National Institute of Information and Communications Technology (NICT) are promoting collaborative project, called G-lambda, to establish such a standard interface, which is called GNS-WSI (Grid Network Service-Web Services Interface). In this paper, we will briefly introduce the architecture we assume in the G-lambda project to define GNS-WSI.