Dynamic Provisioning for Community Services

verfasst von: Prof. Li Qi, Prof. Hai Jin

Verlag: Springer Berlin Heidelberg

Buchreihe : Advanced Topics in Science and Technology in China

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

Dynamic Provisioning for Community Services outlines a dynamic provisioning and maintenance mechanism in a running distributed system, e.g. the grid, which can be used to maximize the utilization of computing resources and user demands. The book includes a complete and reliable maintenance system solution for the large-scale distributed system and an interoperation mechanism for the grid middleware deployed in the United States, Europe, and China. The experiments and evaluations have all been practically implemented for ChinaGrid, and the best practices established can help readers to construct reliable distributed systems.

This book is intended for researchers, developers, and graduate students in the fields of grid computing, service-oriented architecture and dynamic maintenance for large distributed systems.

Li Qi is an Associate Professor and the Deputy Director of the R&D Center for the Internet of Things at the Third Research Institute of Ministry of Public Security (TRIMPS), China. Hai Jin is a Professor and the Director of Department of Computer Science, Huazhong University of Science and Technology, China.

Inhaltsverzeichnis

Frontmatter

1. Provisioning and Maintenance

Abstract

With the rapid development of computer network technology, more and more businesses need to provide a high-quality and uninterrupted service. However, during the running of the system, the re-configuration of resources, maintenance and update will be necessary. For the traditional services, the system maintenance process might inevitably lead to a decreased Quality of Service (QoS) or even service interruptions. When this problem is injected into a distributed system environment, the heterogeneity of the system and network latency and other issues will make this more difficult to maintain and provision to achieve efficient coordination. On the other hand, even the correct maintenance of the system may lead to non-availability. Therefore, in evaluating how to achieve normal running in a distributed system as well as in a grid management system for dynamic maintenance, in order to satisfy user demand for services, distributed computing has become an important part of research in the field. To well understand this problem, we will introduce the history, background, characteristics and challenges of maintenance and provisioning technology in this chapter.

Li Qi, Hai Jin

2. Grid and Distributed Systems

Abstract

Grid computing is becoming more and more attractive for coordinating large-scale heterogeneous resource sharing and problem solving. Of particular interest for effective grid, computing is a software provisioning mechanism. To efficiently maintain and provision software upon a grid infrastructure, the middleware employed to manage the system is greatly important. Thus, understanding the principles and logic of grid middleware is the first step in designing the provisioning and maintenance system.

Li Qi, Hai Jin

3. Highly Available Dynamic Deployment Infrastructure

Abstract

To effectively support dynamic provisioning, an adaptive infrastructure which enables host-swapping and fault tolerance is necessary. Addressing this challenge, we propose a Highly Available Dynamic Deployment Infrastructure (HAND), based on the Java Web Services Core of Globus Toolkit 4. HAND provides capability, availability and extensibility for dynamic deployment of Java Web Services in dynamic grid environments. We identify the factors that can impact dynamic deployment in static and dynamic environments. We also present the design, analysis, implementation and evaluation of two different approaches to dynamic deployment (service level and container-level). We also examine the performance of an alternative data transfer protocol for service implementations. The results demonstrate that HAND can deliver significantly improved availability and performance relative to other approaches.

Li Qi, Hai Jin

4. Service-Oriented Dependency-Aware Maintenance

Abstract

When the scale of a computational system grows from a single machine to a grid with thousands of nodes, the interdependencies among these resources and software components also become complicated. The maintenance of the increasingly distributed systems will face many challenges. One of the most important challenges is how to balance the efficiency of maintenance and the availability of a global system without shutting down the whole system. In this chapter, a novel mechanism, called Cobweb Guardian, is proposed to provide several solutions to avoid or reduce the effects of different dependencies, i.e., deployment dependency, invocation dependency and environment dependency. By using the Cobweb Guardian, administrators of a grid can execute the maintaining task safely in the runtime with high availability. The evaluation results show that our proposed dependency-aware maintenance can bring higher throughput and availability for a grid during maintenance in the runtime.

Li Qi, Hai Jin

5. Asynchronous Provisioning Strategy for Heterogeneous Grids

Abstract

Following upon the infrastructure and architecture discussed in the former chapters, this chapter proposes a distributed and asynchronous model in our Guardian maintenance system for dynamic provisioning that is well-suited to such an intricate setting. The solution introduces multiple maintaining layers so that all the maintenance instructions can be done in an asynchronous way. Experimenting with the new model, in which maintenances are executed in an interlaced time sequence, the downtime of the global system during service maintenance can be effectively reduced by preventing a Weakest Link Effect usually arising from individual heterogeneity or emergent faults. Thus, the consistency of alien nodes and the execution of new requests (after maintenance) can be guaranteed simultaneously. In particular, the implications of maintenance due to heterogeneous dependencies can be detected and resolved automatically in this framework. The practical evaluations prove that the Guardian system equipped with this model can improve the availability and reduce downtime effectively.

Li Qi, Hai Jin

6. Orchestrating Provisioning among Heterogeneous Middleware

Abstract

After several years of development, the computational infrastructure has been widely developed. In particular, when the grid technology grows to production level, users have several options to handle resourceless problems when submitting large-scale jobs to the infrastructure. The capability of dynamic provisioning, fast deployment of virtual machines and interoperation can help users complete their jobs correctly even if the local resources are not adequate. To orchestrate these approaches well and find the optimal solution for end users, this chapter proposes a novel orchestrating model to optimize these approaches efficiently. An image processing application running on three heterogeneous grid platforms is adopted to demonstrate the efficiency and capability of the proposed model. The result proved that the optimal solution can efficiently orchestrate the jobs among the grids.

Li Qi, Hai Jin

7. Implementation of Dynamic Provisioning with Community Services

Abstract

In the earlier chapters we explored the provisioning technology in terms of infrastructure, architecture, model and strategies. Combined with these features, we implemented the Guardian system to help the administrators, developers and end users to provision applications on ChinaGrid. After exploring the implementation details and actual workflow, readers can understand more about the service-oriented provisioning system. Furthermore, we will introduce practical applications that apply provisioning technologies to improve the efficiency and availability. These applications have been deployed on ChinaGrid.

Li Qi, Hai Jin

8. Conclusions and Future Challenges

Abstract

With the rapid development of a distributed computing system, an efficient and effective provisioning system is necessary to help different users easily access unlimited computing or storage power. This chapter reviews the whole book and concludes the contribution of our studies.

Li Qi, Hai Jin

Backmatter

Titel: Dynamic Provisioning for Community Services
verfasst von: Prof. Li Qi
Prof. Hai Jin
Verlag: Springer Berlin Heidelberg
Electronic ISBN: 978-3-642-34513-5
Print ISBN: 978-3-642-34512-8
DOI: https://doi.org/10.1007/978-3-642-34513-5