Title :
Optimal reconstruction bandwidth scheme for Zigzag-decodable code with combination property in cloud storage system
Author :
Mingjun Dai;Bailu Mao;Yongjun Fan;Xiaohui Lin;Hui Wang;Bin Chen
Author_Institution :
Shenzhen Key Lab of Advanced Communication and Information Processing, Shenzhen Key Lab of Media Security, College of Information Engineering, Shenzhen University, Shenzhen, China
Abstract :
A data reconstruction scheme for the storage code that possesses the combination property (CP) and zigzag decod-able (ZD) is proposed for distributed storage system. (n, k) CP is defined as follows: k message with each containing L bits are encoded to n ≥ k packets and then stored into n storage nodes, respectively. Any k from these n storage nodes can recover the original message. The code operates within binary field. Straightforward utilization of binary field to design storage code cannot support (n, k) CP for n greater than 3. Therefore, bit shift technique is needed, which requires some storage overhead to tackle the n constraint problem. In the data reconstruction process, however, it is shown that no overhead is required. In other words, our proposed scheme contacts k storage nodes and only fetches L bits from each storage node, namely the amount of bits fetched is equal to the amount of bits to recover which indicates optimal bandwidth consumption. Though the optimal bandwidth reconstruction scheme is designed dedicated for (8, 4) CP at current stage, it sheds light on further generalizing for arbitrary (n, k).
Keywords :
"Decoding","Systematics","Bandwidth","Encoding","Connected vehicles","Cloud computing","Complexity theory"
Conference_Titel :
Connected Vehicles and Expo (ICCVE), 2015 International Conference on
Electronic_ISBN :
2378-1297
DOI :
10.1109/ICCVE.2015.87