Earticle

In this paper, we simulate and comparatively analyze the sensor network protocols with additional aggregate queue and transmit duplicate layers on MICA and MICAz platform with different radio models using PROWLER and RMASE. The simulation results show that the MICAz motes give low latency, high throughput, high energy consumption, low efficiency but better lifetime while the MICA motes give high success rate and low loss rate. It has been, thus, concluded that in case of all the radio models the MICA is preferably better than MICAz in applications where energy is a constraint otherwise MICAz is better as it gives greater lifetime.

In the current dominant IaaS framework, the centralized VM image management mechanism leads to massive consumption of shared bandwidth resources of an IaaS data center (IDC) in face of a large amount of concurrent deployment tasks and has a limited image management ability in large scale IDCs formed by multiple physic server clusters, causing many problems for IDCs, such as, the performance degrading of existing application systems, the long deployment time of new coming ones and the poor scalability. In view of the defects of it, this paper proposes a decentralized and historical VM image file-based image management mechanism. In this mechanism, the Native Image Repository is configured for every physic server cluster in an IDC and managed independently. Each physic server cluster uses VM images in the Native Image Repository to complete deployment tasks assigned to it without the need to transmit the images of customer application systems. Experimental results prove that the proposed mechanism is able to decrease the amount of data transmission involved in deployment tasks and save shared bandwidth resources in an IDC, shortening deployment time and not exerting adverse effects to the performance of the existing application systems.

The computational grid is a collection of resources from multiple administrative domains and selection of appropriate resources for a client application is a critical task. The process of resource selection should not be a simple matchmaking mechanism for a reliable execution environment. Therefore, appropriate resources are selected from the available grid resources that satisfy the application requirements of trust, computation capacity and network capacity and establish dynamic service level agreements to provide guaranteed services to the client. The proposed framework implements dynamic service level agreements by monitoring the job execution and extending the resource usage by predicting the future load on the resources. Simulation results show that there is a significant improvement in the success rate of the jobs submitted to the grid and considerable reduction in the SLA creation time.

The MapReduce framework has been widely used to process and analyze largescale datasets over large clusters. As an essential problem, join operation among large clusters attracts more and more attention in recent years due to the utilization of MapReduce. Many strategies have been proposed to improve the efficiency of distributed join, among which bloomfilter is a successful one. However, the bloomfilter’s potential has not yet been fully exploited, especially in the MapReduce environment. In this paper, three strategies are presented to build the bloomfilter for the large datasets using MapReduce. Based on these strategies, we design two algorithms for two-way join and one algorithm for multi-way join. The experimental results show that our algorithms can significantly improve the efficiency of current join algorithm. Moreover, cost models of these algorithms are characterized in order to find out the way of improving the performance of two-way and multi-way joins.

For internet servers, how to handle colossal data volumes and maintain data consistency is a major challenge. To ensure each database maintains the latest information and messages do not conflict with one another, we have a new type of database NoSQL (Not Only SQL). A number of atomic broadcast protocols, including libPaxos, Mencius and RingPaxos, have been established to help maintain transmission consistency for cloud computing services. This paper introduces a new atomic broadcast protocol to improve the performance of existing protocols. Our new protocol divides the network topology into various areas to process messages in a distributed way and uses a rotating mechanism to effectively balance the server loads. Experimental evaluation has been conducted to compare the performance of different protocols. The results show that our new protocol yields higher throughput and lower latency under different client loads in contrast to other protocols.

With the development of cloud computing, high energy consumption issue has attracted more and more attentions recently. To achieve the goal of energy conservation, accurately measuring the power consumption of distributed resource is of significant importance. Conventional power models can only provide fine-grained power measurement for physical devices instead of virtualized resources. In this paper, we proposed a novel power measuring technique, which uses performance monitoring counters (PMC) as basic indicator of virtual machine’s power consumption. Meanwhile, we also designed a scheduling algorithm to reduce the measuring errors caused by recursive power consumption. Theoretical analysis indicates that the proposed algorithm can provide bounded error when metering virtual machine’s power. Massive experiments are conducted by using various benchmarks on different platforms, and the results shown the proposed technique is effective and low-cost for measuring the power consumption of virtualized resources in large-scale cloud systems.