Improving WSN Base Station Survivability with Rejuvenation Technology
He Xin,Gui Xiaolin,Wei Wei
DOI: https://doi.org/10.1080/15501320802561708
IF: 1.938
2009-01-01
International Journal of Distributed Sensor Networks
Abstract:The conventional security technology for WSN is mainly focused on improvement to the system's anti-invasion ability; however, seldom concerns how the system should go on providing external services after the invasion and failure, leading to the system's low survivability. In addition, base station is a keyport for WSN. Therefore, software rejuvenation technology is introduced into the construction of a survival model for WSN base station, in order to improve the overall WSN survival performance. The model includes the monitor, the analysis unit, the rejuvenation operation unit, the evaluation unit, strategy Database, etc. Its operation principles are as follows: first, the monitor can gets context information of the system recourses through agents (software sensors), and makes prompt examination and data collecting. Then, the analysis unit will make policy on the collected data, and judge the existence cause and harm degree of abnormal resource consumption by means of knowledge provided by strategies database, in order to perform the rejuvenation operation in time. When the rejuvenation occurs, it is necessary to clear the internal status (e.g., trash recycling, refreshing operation system tables, reinitializing internal data construction etc., adopting different operations in the presence of different strategies), to cut off exterior malicious links, to release system resources, and to restore system performance. Besides, when the rejuvenation operation is performed, the reconfiguration operation can also be adopted in order that patches are put on according to the system loophole, to avoid wrong reappearance caused by system interior drawbacks or the continuous occurrence of the same exterior attacks. Finally, according to the restored system performance, an evaluation of the original strategy which influenced the operation this time should be made and recorded in the strategies database by the evaluation unit, making the frame model adjust itself to a varied environment. Thus, the uninterrupted services can be provided by the system in face of attack or failure by the model, which monitors system performance to predict the rejuvenation interval and performs rejuvenation operation. Otherwise, the model is described by using the Semi-Markov Process and it is discovered that the detection probability of the system compromised state is the key parameter of improving the model's survivability by analyzing the model's stable health state probability. Too convienient for simulation, the base station survival model is described with Stochastic reward network and the experimental results gained by simulation tool SPNP show that the system detection probability is in inverse proportion of the failure probability. According to thought of rejuvenation technology, the failure probability can be reduced significantly by improving system detection probability so as to enhance the survivability of the system, and further improve the overall survival performance of WSN.