Research On Service Discovery And Load Balancing In The Gird Environment

Military Grid develops from the basis of civilian grid, it is a new technology in theMilitary Information Systems and its core is a software system which is huge,collaborative and distributed in each node in the Military grid. It has the ability ofintegrating all the resources of hardware and software based on the existing facilitieswhich is used to transmit and process information to let the resources can be widelyused.There are all kinds of resources which has the features of distributive in the areaand heterogeneous in the operating platform in the grid, the main studying issue is howto organize and find the service in the grid more effectively; the advantages anddisadvantages of services forms of organization and find service quickly determineswhether the grid is efficient and friendly. Based on the research of Military Grid, wedesigned a hierarchical model of grid service discovery and made a deep study in themechanism of service discovery and the strategy of the load balancing with the currentgrid model and the discovery algorithm.The main work of this paper are as follows: we analyze detailedly the keytechnology of the discovery method in the grid such as the description of gridservices,the forms of organization of the grid service, the service discovery algorithmand the load balancing strategy which are the basis of the designing of service discoverymodel in this paper; By comparing the advantages and disadvantages of the distributedmodel and the hierarchical model, we build a new hierarchical model of grid servicediscovery which is consisted of three layers: the virtual organization layer, ontologylayer and management layer, and we also design a service discovery algorithm and aload balancing strategy based on the model. Finally, we make some experiments toshow that the algorithm in this paper can improve the situation of the traditional servicediscovery algorithm which is lack of semantic and the search result is not as well as the user expected, make the computer could deal the service more automatic and intelligent;The service requests are limited in some ontology domain, thereby it can not onlyreduce the search space and improve the efficiency of the service discovery, but also canreduce the cost of network; it can reduced the possibility of some parts of nodes in thegrid because of overload and led to a decline in performance or even collapse, while theother parts of the nodes are always in the idle state, so that it make every node in thegrid to achieve a relative sense of equilibrium by adopting the load balancing strategy.