Cloud computing allows for high volume storage and flexible access over wireless networks. However, current algorithms require all data to be fully downloaded in one access event. This is limiting because wireless bandwidth is sometimes insufficient and full-downloading is not power efficient. This technology is a scheme that permits partial downloading and full reconstruction of data in distributed networks. This technology optimizes power and bandwidth management to increase the efficiency of cloud computing networks with many users and highly limited bandwidth.
Existing schemes require all symbols from each data node of a distributed network to be downloaded in one access event. This technology utilizes custom-designed regeneration codes and cross-layer wireless resource allocation to reliably and efficiently download a subset of symbols from many storage nodes and reconstruct the full data set. Special algorithms determine an optimal order to access data from the various nodes. This reduces the complexity of resource allocation on wireless networks increasing power and bandwidth efficiency while improving download speed and reliability. This technology may improve access to wireless distributed networks by mobile devices, sensor networks, and other low-power equipment.
The improved efficiency of the technology has been demonstrated using numerical simulations.
Tech Ventures Reference: IR CU12344