Yongcai Tao

Scalable DHT- and ontology-based information service for large-scale grids

Current grid information services are centralized or hierarchical and prove inefficient as the scale of the grid rapidly increases. The introduction of the P2P DHT technique into grids brings an encouraging path. However, current applications of the P2P DHT technique to grids do not consider the Virtual Organization (VO) management mode of grid resources. Frequent joining and leaving of resources requires strong self-organization capacity of the system to maintain the rigid structure. Moreover, arranging a moderate identifier space for a DHT ring is knotty. A large identifier space will make some nodes overloaded, while a small identifier space will bring forth the same problem as the millennium bug. In addition, current grid services are described in XML-based standards, which only support syntactic keyword and taxonomy-based queries without reasoning ability at the semantic level, leading to poor precision and integrality. To address these issues, the paper proposes a scalable DHT- and ontology-based Information Service (DIS) for a grid system, which organizes resources into a DHT ring based on the VO mode. To save the identifier space, only stable VOs can join DIS via a new DHT node, whereas volatile VOs join DIS via being the sub-domain of other VOs. Furthermore, ontology-based information integration is adopted in DIS and a novel ontology for grid resources is designed, which supports semantic-based information queries. By integrating DHT- and semantic-based query techniques, DIS speeds up the information query and improves the query precision and integrality. DIS is evaluated over the ChinaGrid environment and experimental results show that DIS provides rapid resource query, high throughput and strong scalability.

Dependable Grid Workflow Scheduling Based on Resource Availability

Due to the highly dynamic feature, dependable workflow scheduling is critical in the Grid environment. Various scheduling algorithms have been proposed, but seldom consider the resource reliability. Current Grid systems mainly exploit fault tolerance mechanism to guarantee the dependable workflow execution, which, however, wastes system resources. The paper proposes a dependable Grid workflow scheduling system (called DGWS). It introduces a Markov Chain-based resource availability prediction model. Based on the model, a reliability cost driven workflow scheduling algorithm is presented. The performance evaluation results, including the simulation on both parametric randomly generated DAGs and two real scientific workflow applications, demonstrate that compared to present workflow scheduling algorithms, DGWS improves the success ratio of tasks and diminishes the makespan of workflow, so improves the dependability of workflow execution in the dynamic Grid environments.

GNSD: A Novel Service Discovery Mechanism for Grid Environment

Due to the highly distributed and dynamic features, service discovery is a key requirement in grid computing. The paper first introduces the existing service discovery mechanisms in distributed computing, and then points out the problems in MDS which is widely used in grid systems. In order to address these issues, GNSD, a novel service discovery mechanism is proposed, which is based on the OSPF (Open Shortest Path First) flooding mechanism and combines the advantages of tree architecture and flat architecture. The time and space complexity of both GNSD and MDS are also analyzed. Experimental results show that GNSD not only accelerates the service discovery but also relieves the burden of servers at the cost of little extra storage and communication traffic. Furthermore, it is more robust. As a conclusion, GNSD is an efficient service discovery mechanism

PGWFT: a petri net based grid workflow verification and optimization toolkit

Graphical workflow modeling tools, such as UML and DAG, can facilitate users to express workflow process logic, but lack of abilities to carry out simulation and correctness checking. In this paper, we propose a service composition oriented grid workflow model and its related six elementary workflow patterns: sequence, condition, iteration, concurrency, synchronization, and triggering. Based on this, we present a Petri net based grid workflow verification and optimization toolkit, called PGWFT, to help common users describe workflow instances and perform verification and optimization analysis. The experimental results show that our workflow verification and optimization mechanisms are feasible and efficient.

Scalable dht-based information service for large-scale grids

Current grid information service is centralized or hierarchical and proves inefficient as grid scale rapidly increases. The introduction of P2P techniques into grids breaks an encouraging path. However, frequent join and departure of resource nodes require strong self-organization capacity of system to maintain their rigid structure. Moreover, arranging identifier space for P2P nodes is knotty and has great impact on system performance. If the identifier space is too large, some nodes will be overloaded. On the contrary, small identifier space will bring the same problem as millennium bug. To address the issues, this paper proposes a scalable DHT-based (Distributed Hash Table) Information Service (DIS) for grid system, which organizes grid resources into a DHT ring based on VO (Virtual Organization). To save the identifier space while retaining the scalability and system performance, only stable VOs can join DIS via a new DHT node, whereas volatile VOs join DIS through being the sub-domain of other VO. Experimental results show that DIS provides rapid resource query, strong scalability and high throughput, meanwhile avoiding the key node failure as well as the bottleneck problem.

Performance Modeling of Resource Failures in Grid Environments

Grid is susceptible to a number of software and hardware failures, so a deep understanding of and modeling the grid resource failures are a challenge and have significant influence on grid researching. However, due to various reasons such as commercial secret and security, it is difficult to obtain real historical logs of grids. Therefore, an accurate model of resource failures is critically useful. In the paper, through analyzing the grid log data, we detail the suitability of three potential statistical distributions for each data set: Weibull, Zipf’s law and Pareto. Then, this paper develops a grid resource failure simulator. Finally, with the different failure patterns generated by the failure simulator, the paper evaluates several common scheduling algorithms used in grid systems.

An optimistic checkpoint mechanism based on job characteristics and resource availability for dynamic grids

In the paper, based on the job characteristics and resources availability, an optimistic checkpoint mechanism for dynamic grids(OCM4G) is proposed. It can determine whether to checkpoint a given job running on a given resource node and establish optimal aperiodic checkpoint intervals by applying the knowledge of job characteristics and resource availability. We evaluate OCM4G over a real grid environment (ChinaGrid) and the results show that OCM4G achieves better performance than the periodic checkpoint and the analytical method of calculating aperiodic checkpoint intervals.

DGSS: A Dependability Guided Job Scheduling System for Grid Environment

Due to the diverse failures and error conditions in grid environments, node unavailability is increasingly becoming severe and poses great challenges to reliable job scheduling in grid environment. Current job management systems mainly exploit fault recovery mechanism to guarantee the completion of jobs, but sacrificing system efficiency. To address the challenges, in this paper, a node TTF (Time To Failure) prediction model and job completion prediction model are designed. Based on these models, the paper proposes a dependability guided job scheduling system, called DGSS, which provides failure avoidance job scheduling. The experimental results validate the improvement in the dependability of job execution and system resources utilization.