Search

We introduce a new parallel and distributed algorithm for the solution of the satisfiability problem. It is based on an algorithm portfolio and is intended to be used for servicing requests in a distributed cloud. The core of our contribution is the modeling of the optimal resource sharing schedule in parallel executions and the proposition of heuristics for its approximation. For this purpose, we reformulate a computational problem introduced in a prior work. The main assumption is that it is possible to learn optimal resource sharing from traces collected on past executions on a representative set of instances. We show that the learning can be formalized as a set coverage problem. Then we propose to solve it by approximation and dynamic programming algorithms based on classical greedy algorithms for the maximum coverage problem. Finally, we conduct an experimental evaluation for comparing the performance of the various algorithms proposed. The results show that some algorithms become more competitive if we intend to determine the trade-off between their quality and the runtime required for their computation.

The goal pursued in this study is to optimise the service provisioning and the energy consumed in a special existing cloud system named SlapOS. The SlapOS cloud innovates in considering that the data centre can be composed by dedicated and volunteer machines. We will use the term volunteer clouds for referring to such organisations. Volunteer clouds offer potential advantages for cloud elasticity, storage of big data and the minimisation of the energy consumption. But, we must manage the potential unavailability of volunteer machines. In this paper, we focus on the design of energy-efficient scheduling approaches in such clouds. Our first contribution is to formulate the scheduling challenges in two new computational problems. These problems are NP-hard. Then, we propose an integer linear programming (ILP) formulation and several greedy heuristics. Finally, we evaluate our approaches throughout various simulations of the SlapOS system in a realistic volunteer computing context.

This paper deals with the graph reduction approach for the QoS prediction of cloud-services’ compositions. Most works with graph reduction are based on a deterministic modelling of the QoS. Though interesting, these representations are not suitable for cloud environments because they neglect the QoS variations. On clouds, we propose to use a probabilistic modelling in which the QoS of an operation is drawn from a finite set of discrete values, according to a Probability Mass Function (PMF). This paper proposes to adapt an existing graph reduction algorithm for such representations. Our first contribution is to show that there is a potential combinatorial explosion when running the reduction algorithm with the probabilistic representation. Our second contribution is to propose two heuristics in which the combinatorial explosion is controlled. The heuristics are based on two classical ideas in arithmetic: truncation and rounding. Our last contribution is an experimental comparison of the heuristics.

For over a decade, the data requirements of e-Science applications increase drastically with the emergence of data-intensive applications. Several tools and frameworks have been developed to manage and handle the big amount of data for the grid platforms. However, the use of these tools by the basic scientist and the grid computing community is not well adopted because of the complexity of the installation and configuration processes. Recently, an open source distributed operating system for clouds emerged, namely SlapOS. The aim of SlapOS is to hide the complexity of IT infrastructures and software deployments from users. In this work, we propose a cloud-hosted data grid using the SlapOS cloud. Through a software as a service (SaaS) solution, users can request and install automatically any data movement and sharing tools like Stork and Bitdew without any intervention of a system administrator. The entire solution is now running in production into the SlapOS cloud at Paris 13 University. Intensive experiments have been conducted on the Grid ‘5000 testbed to validate our approach.

This paper shows how to parallelize a compute intensive application in mathematics (Group Theory) for an institutional Desktop Grid platform coordinated by a meta-grid middleware named BonjourGrid. The paper is twofold: it shows how to parallelize a sequential program for a multicore CPU which participates in the computation; and it demonstrates the effort for launching multiple instances of the solutions for the mathematical problem with the BonjourGrid middleware. BonjourGrid is a fully decentralized Desktop Grid middleware. The main results of the paper are: a) an efficient multi-threaded version of a sequential program to compute Littlewood-Richardson coefficients, namely the Multi-LR program and b) a proof of concept, centered around the user needs, for the BonjourGrid middleware dedicated to coordinate multiple instances of programsfor Desktop Grids and with the help of Multi-LR. In this paper, the scientific work consists in starting from a model for the solution of a compute intensive problem in mathematics, to incorporate the concrete model into a middleware and running it on commodity PCs platform managed by an innovative meta Desktop Grid middleware.