Performance criteria for graph clustering and Markov cluster experiments
Performance criteria for graph clustering and Markov cluster experiments
Graph Clustering Via a Discrete Uncoupling Process
SIAM Journal on Matrix Analysis and Applications
| Hi-index | 0.00 |
Typically, the information security system is a sort of load balancing one, so server must handle all the communication traffics in intranet system. The server and system easily can be unstable because of increased volume overloaded by an internal connection. The delay and even the network failure of system can occurs server down. To solve these problems, the clustering resources must be distributed effectively to the internal infrastructure system. Nowadays high availability clustering and load balacing (HALB) is necessary for availability of clustering mechanism of information sysetm. HALB is a system structure and functional hierarchy for interactive load euquality and load balancing system of each network area. If one server of some network domain fails, the other server of another domain migrates as a backup server under the multiplexing clustering mechanism through the load-balancing and high availability function. To achieve the best distributed computing, this mechanism is based on the value of critical path through dynamic load euquality policy and efficiency management of server resources. This paper provides a experimental test and recomands HALB as a test result, because the test shows that the HALB method is more effective in the load balancing between servers and packet filtering schedulings than that of traditional.