Large archival storage systems experience long periods of idleness broken up by rare data accesses. These systems can loose data for a variety of reasons. Such failures can occur at the device level or at the block level. To deal with these failures, we need to detect them early enough to be able to use the redundancy built into the storage system. We propose a system where drives are periodically accessed to detect drive failure, and where we read all the archived data on all the disks to detect block failures and compensate for them. We research the impact of "disk scrubbing" and its scheduling.

No abstract

Wireless Sensor Network (WSN) is an essential technology for the Internet-of-Things (IoT) and intelligence-based applications. In the case of Intelligent Transportation Systems (ITS), the WSNs play an important role in safety and efficient traffic management. Therefore, there is enormous demand for energy efficient WSNs for dynamic resource allocation in vehicles and infrastructures. This work presents a Multi-Input Multi-Output (MIMO) technique model in WSNs, which addresses the Cluster Head (CH) recognition issue for MIMO sensor networks by using Back Propagation Neural Network (BPNN). The conventional CH identification suffers from a lack of location identification due to the dynamic and real-time environment. Thus, to obtain more precise positioning accuracy, the proposed work uses BPNN combined with a distributed gradient drop technique to calculate the position of the unknown CH. This reduces the distance estimation error, and the particle swarm optimization technique is further used to obtain the optimal weight and threshold of the network. The work is validated by using mathematical analysis, simulations, and comparison with existing techniques. The proposed model shows a better performance in terms of energy consumption, error rate, and computation time.