MGSCP: A scalable service migration framework for spatial information grid

Service migration is important for the execution of distributed spatial tasks because the spatial data published in autonomous Grid node is often not permitted to be freely transferred in Spatial Information Grid. In this paper, we design and implement a scalable service migration framework. In our framework, all spatial services are composed of fine-grained meta-functional components and managed by GIS Function Center. A service can be constructed by functional-component-based workflow, executed by service workflow engine and its´ template can be persistent stored in Service Template Library. On these bases, we presented a practical service migration process composed of three sub-processes: the first to dynamically find the best source node, the second to transfer the related function packages and register function interfaces and the third to transfer the service template and deploy it without reloading service context. We also give a spatial analysis service migration example to demonstrate the process.