Considering the low efficiency of massive data retrieving in traditional parallel processing, by taking advantage of the great availability of cloud computing paradigm, we propose a hybrid solution based on Map-Reduce model and distributed computing framework--Spark. Moreover, we design and implement this solution in our lab. The results show that the solution can effectively improve the performance of massive data retrieving.

Full Function Point measurement method (COSMIC-FFP) is too abstract and difficult to be applied to object-oriented software. By analyzing the notional relationship between main components of UML and main elements of COSMIC-FFP, we establish 22 measurement mapping rules, propose UML-Based COSMIC-FFP method and list the measurement process and measurement steps. An example is given and the results show that this method can measure the functional size of object-oriented software objectively and effectively.

This paper proposes a combined local best particle swarm optimization algorithm (CLBPSO) which combined with local optimum particle information. And it gives three ways of combination local information. Experimental results indicate that the CLBPSO algorithm improves the search performance on the benchmark functions significantly. On the basis of experimental results, we will also compare these three methods with each other to find the best one.

As the characteristics of power line channel varies over time, makes it be different from the tradition networks. In this paper, we study the performance of a form of slotted binary exponential random backoff algorithm using in HomePlug protocol. The performance of the backoff algorithm is evaluated by simulation. The results show that it makes the network performance become worse with the expansion of network.