To Schedule Tasks In Digital Organisms

With the development of Internet, the applications based on Internet becomes various and aboundant. The traditional Client/Server architecture model is not suitable for the development. The P2P technology, which can take good use of the resources distributed on a greate number of Internet personal computers, have greater potential on the extendability, utilizability and daptability.The traditional P2P system are mainly designed to solve the file's sharing and storage, cause that the most of researches focus on resource placement and orientation.Whereas the popularity of network streaming application (Such as VoD, IPTV and VoIP) brings out a new chanllege to P2P system. The streaming applications extend the resource sacle of traditonial P2P system, which regards the dynamic service as a kind of resource.There are great differences between file resource and service resource. The former is longeveous, dull in network-quality and steady. The latter, on the contrary, is shor-lived, sensitive and dynamic. So,The requirement of building up an efficient management system for the service resource becomes instant. A new P2SP system—DOS (Digital Organism System), which based on distributed file system and designed by 8010 research, offers an intergrated solution on distributed resource utilization and storage. Currently, DOS realizes and promots the most of functions of P2P system. Howerver,it has not realized the management for service resource. This document designes and realizes a subsystem of DOS, which aim at service resource management and scheduling.This subsystem build up a hiberarchy to organize the service providers into dynamic logic regions. All the service monitering work in one logic region is administrated by a supervisor node selected from that region. Each supervisor node sends analyzing reports to schedule management center for scheduleing decision and control. The subsystem realizes the efficiently management of service resource.Tests show that the subsystem's functions meet the design requirements. Furthermore, it can reduce the burden of server and network streaming loads.