Jiong Yu

A New Self-Deployment Algorithm in Hybrid Sensor Network

Sensor network is usually composed of static nodes or mobile nodes. In order to achieve maximum performance, deployment of nodes should maximize coverage. The deployment of static sensor network is simple, but random location can not ensure the better connectivity. At the same time, the cost of mobile sensor network which consists of only mobile nodes is much high. In this paper, a virtual force based deployment-enhanced algorithm (VFDEA) was proposed to obtain better behavior in hybrid sensor network. Coverage holes and network partition were eliminated in the first step and the connectivity of all nodes is achieved. Overlapped coverage is removed in the second step and coverage percentage increases. The validity of the algorithm is verified by serials of simulation.

Energy-efficient algorithms for distributed storage system based on block storage structure reconfiguration

As the underlying core infrastructure for cloud computing, distributed storage systems like the Hadoop Distributed File System (HDFS) are the foundation of all kinds of cloud services. However, designers of the ever-expanding systems have ignored the problem of high energy consumption, causing serious environmental and economic problems. The data availability and performance Quality of Service (QoS) requirements make it hard to use existing energy-saving technologies to solve the problem. After researching the data block?s storage structure and mechanism, and the relationship between the server?s status and the data block?s availability, the method to solve the problem of ensuring data availability and performance QoS requirements is proposed. The energy-saving model for the distributed storage system is defined. The algorithm divides the RACK into two distinct storage areas, Active-Zone and Sleep-Zone, reconfiguring the data storage structure using the block storage structure reconfiguration algorithm. To save energy, we turn the servers in Sleep-Zone to sleep mode while the workload is low. Numerical analysis and experimental results demonstrate that the energy-efficient algorithms improved the energy efficiency for the distributed storage system.

A QoS-aware Dynamic Data Replica Deletion Strategy for Distributed Storage Systems under Cloud Computing Environments

As the underlying core infrastructure and important component of cloud computing, distributed storage system is the foundation of all kinds of cloud services. While its replica mechanism enhanced the reliability and performance of system, the extra block replicas consume more disk space and maintenance cost significantly increased too. In this paper, we propose a novel QoS-aware Dynamic Data Replicas Delete Strategy for disk space and maintenance cost saving purpose. Experimental results demonstrate that the DRDS algorithm can save disk space and maintenance costs for distributed storage system while the available and performance QoS requirements are ensured.

A green framework for DBMS based on energy-aware query optimization and energy-efficient query processing

Traditional database systems result in high energy consumption and low energy efficiency due to the lack of consideration of energy issues and environmental adaptation in the design process. In this study, we report our recent efforts on this issue, with a focus on energy-aware query optimization and energy-efficient query processing. Firstly, a method of modeling energy cost of query plans during query processing based on their resource consumption patterns is proposed, which helps predict energy cost of queries before execution. Secondly, as the traditional query optimizer focuses on solely optimizing for performance and ignores energy-efficient query plans, a query-plan evaluation model is proposed after a comprehensive study of plan evaluation principles. Using the cost model as a basis, the evaluation model can utilizes the trade-offs between power and performance of plans, and helps the query optimizer select plans that meet performance requirements but result in lower energy cost. Finally, a green database framework integrated with the two above models is proposed to enhance a commercial DBMS. Experimental results reveal that, with reliable and accurate statistical data, the proposed framework in this study can achieve significant energy savings and improve energy efficiency. Impact of cache structures on various costs of query processing should be studied.The proposed energy cost model can make an accurate prediction of energy cost.The query-plan evaluation model can help the optimizer select energy-efficient plans.

Multiple DAGs Scheduling Based on Lowest Transportation and Completion Time Algorithm on the Cloud

According to multiple DAG work Flow scheduling problem in heterogeneous distributed environments, in this paper, proposed a scheduling algorithm based on minimize the data transmission time and task completion time, which can deal with the problem that multiple DAGs workflow have the same priority, and gives the multi-priority multi-DAG mixed scheduling algorithm. Compared with E-Fairness algorithm, the experiments show that on the basis of fairness to ensure multiple DAGs scheduling, this algorithm can avoid additional data transfer overhead, shorten the entire workflow execution Make span, and improve resource utilization.

An Enhanced Self-deployment Algorithm in Mobile Sensor Network

In order to enforce the sensing performance of mobile sensor network (MSN), an enhanced self-deployment (ESD) algorithm is proposed in this paper. By designing the virtual repulsive force between nodes, the movement equation and the virtual attracting-field in the sensing area, the problems of the network partition and coverage holes which are aroused by conventional potential-field-based deployment approach is solved. Simulation results show that ESD algorithm exhibits excellent performance and expands the generalization of MSN greatly.

Study on the Non-Dedicated Grid Scheduling System

Effective task scheduling and deployment is hard to achieve in a grid environment, especially where computing resources are heterogamous and local tasks have a high priority in the non-dedicated grid. We introduce a scheduling system of the non-dedicated grid and propose a novel task scheduling algorithm. The simulations show that the algorithm overweighs traditional algorithms a lot in the makespan and the scalability of the scheduling algorithm.

The Priority Tasks Scheduling Algorithm Based on Grid Resource Prediction

According to the dependence and deadline of the grid workflow tasks, the effective degrees and MIPS of the grid resources, a new algorithm called the priority tasks scheduling algorithm based on the grid resource prediction is presented. The algorithm uses DAG to find the critical path, obtain the deadline of every task and compute their PRI (priority). The algorithm takes the below problems into consideration: the request of user, the type of resources and re-scheduling of failed tasks. The result shows that the algorithm is effective.

An Enhanced Precise Self-deployment Algorithm in Mobile Sensor Network

To reduce the cost of abundant mobile sensors deployed in unknown environment, a target-oriented self-deployment algorithm (TSA) based on attracting force line is proposed. By designing the attracting force line and constructing virtual attracting force and repulsive force, the paths between sink and targets are set up automatically. Simulation results show that the proposed algorithm compared to similar methods achieves shorter average moving distance with shorter maximum moving distance and shortens the deployment time.

Grid Workflow Scheduling Based on the Resource Combination Reliability

To the time-constrained workflow scheduling in grids, this paper proposes a new scheduling algorithm in terms of the finite-state continuous-time Markov process through selecting a resource combination scheme which has the lowest expenditure under the certain credit level of the resource reliability on the critical path in the DAG-based workflow. The simulation shows the validity of theory analysis.