CoPart: a context-based partitioning technique for big data

The MapReduce programming paradigm is frequently used in order to process and analyse a huge amount of data. This paradigm relies on the ability to apply the same operation in parallel on independent chunks of data. The consequence is that the overall performances greatly depend on the way data are partitioned among the various computation nodes. The default partitioning technique, provided by systems like Hadoop or Spark, basically performs a random subdivision of the input records, without considering the nature and correlation between them. Even if such approach can be appropriate in the simplest case where all the input records have to be always analyzed, it becomes a limit for sophisticated analyses, in which correlations between records can be exploited to preliminarily prune unnecessary computations. In this paper we design a context-based multi-dimensional partitioning technique, called CoPart, which takes care of data correlation in order to determine how records are subdivided between splits (i.e., units of work assigned to a computation node). More specifically, it considers not only the correlation of data w.r.t. contextual attributes, but also the distribution of each contextual dimension in the dataset. We experimentally compare our approach with existing ones, considering both quality criteria and the query execution times.

Data mining and visualization from planetary missions: the VESPA-Europlanet2020 activity

This paper presents the VESPA (Virtual European Solar and Planetary Access) activity, developed in the context of the Europlanet 2020 Horizon project, aimed at providing tools for analysis and visualization of planetary data provided by space missions. In particular, the activity is focused on minor bodies of the Solar System.The structure of the computation node, the algorithms developed for analysis of planetary surfaces and cometary comae and the tools for data visualization are presented.

A Novel Model for Multi-Parametric Load Balance in Grid Computing

Grid computing is a computing model in which computing resources are geographically dispersed. A computing node can share these resources as well as can transfer applications at other nodes to execute it. Due to the absence of centralized authority in grid, some resources may be overloaded and others may be under loaded. To obtain high performance, load balance strategy is necessarily needed. Load balance strategy can be affected by different parameters like network parameters, application characteristics, computing node capacity etc. In this paper, we consider using three parameters namely network parameters, computation node capacity and application characteristic to obtain effective load balance. All these parameters will pass to AHP (Arithmetic hierarchy process) for automatic decision making to select better resources for high performance service.

On Construction of a Diskless Cluster Computing Environment in a Computer Classroom

This paper’s objective is to implement and evaluate a high-performance computing environment by clustering idle PCs (personal computers) with diskless slave nodes on campuses to obtain the effectiveness of the largest computer potency. Two sets of Cluster platforms, BCCD and DRBL, are used to compare computing performance. It’s to prove that DRBL has better performance than BCCD in this experiment. Originally, DRBL was created to facilitate instructions for a Free Software Teaching platform. In order to achieve the purpose, DRBL is applied to the computer classroom with 32 PCs so to enable PCs to be switched manually or automatically among different OS (operating systems). The bioinformatics program, mpiBLAST, is executed smoothly in the Cluster architecture as well. From management’s view, the state of each Computation Node in Clusters is monitored by “Ganglia”, an existing Open Source. The authors gather the relevant information of CPU, Memory, and Network Load for each Computation Node in every network section. Through comparing aspects of performance, including performance of Swap and different network environment, they attempted to find out the best Cluster environment in a computer classroom at the school. Finally, HPL of HPCC is used to demonstrate cluster performance.