Broadcast disks: data management for asymmetric communication environments
SIGMOD '95 Proceedings of the 1995 ACM SIGMOD international conference on Management of data
The art of computer programming, volume 3: (2nd ed.) sorting and searching
The art of computer programming, volume 3: (2nd ed.) sorting and searching
Proceedings of the ninth international conference on Information and knowledge management
A generalized air-cache design for efficiently broadcasting on multiple physical channels
Proceedings of the 2001 ACM symposium on Applied computing
A near optimal algorithm for generating broadcast programs on multiple channels
Proceedings of the tenth international conference on Information and knowledge management
Efficient Data Allocation over Multiple Channels at Broadcast Servers
IEEE Transactions on Computers
Index and Data Allocation on Multiple Broadcast Channels Considering Data Access Frequencies
MDM '02 Proceedings of the Third International Conference on Mobile Data Management
The integration of telecommunication and dissemination networks
International Journal of Communication Networks and Distributed Systems
Hi-index | 0.24 |
Data broadcast is known as a scalable way to transmit database items to large client populations through a wireless channel. However, with a large set of items, the expected delay of receiving desired data increases due to the sequential nature of the broadcast channel. One possible solution is to increase the number of available channels and allocate items evenly over these channels, then cyclically broadcast data in each channel. In view of data access skew (the access frequencies of data are usually different), some channels may be reserved exclusively for those few frequently requested items, while the infrequently accessed bulk of data are allocated on other channels. In this paper, an O(N log K) restricted dynamic programming (RDP) algorithm is proposed to partition N items over K channels assuming data access is skewed with the object of minimizing the average expected delay (aed) of clients. To speed up the DP process, for any partition, we predict a possible location which may be very close to the optimal cut by using a low bound as the actual aed for the remaining items. Thus, the number of comparisons in DP can be restricted to a small interval around the predicted cut point. To further reduce the costs in DP, the hierarchical property of optimal solution is adopted. Simulation results show that, the hit rate obtained by RDP is higher than 90% and it also outperforms the existing algorithm 200%. We extend the work of RDP and develop an O(N log N log K) PKR algorithm; simulation results show that the solution is optimal.