Fast replica allocation method by parallel calculation on DAPDNA-2

Replica placement problem is to select a subset from a group of potential nodes to put replicas in Content Distribution Network (CDN). It is derived from the set cover problem which is known to be NP-hard. So it is difficult to calculate the large-scale replica placement problem on a program counter-based processor. Several greedy algorithms are proposed in order to decrease calculation time. However, it was proved mathematically that no greedy algorithm can obtain the optimal solution. This paper proposes a fast calculation method of the replica placement problem, which is implemented on reconfigurable processor DAPDNA-2 of IPFlex Inc. Our proposed method divides the combination optimally and performs pipeline operation. Beeler's algorithm can calculate all combinations in ascending order but it has data dependence. It's difficult to calculate any pattern because each data increases irregularly. In order to solve this problem, we propose the new algorithm that generates any order pattern. In addition, the optimal number of partitions depends on the number of combinations and calculation clocks of Beeler's algorithm. In order to solve this problem, we think about the optimal division number in theory. While the time complexity of conventional method is proportional to the number of combinations, that of proposed method is proportional to the square root of the number of combinations. Experimental results show that the proposed algorithm reduces the execution time by 40 times compared to Intel Pentium 4 (2.8GHz). copyright