Exploiting composition of mobile devices for maximizing user QoS under energy constraints in mobile grid
Introduction
In the grid, the computing resources are autonomously managed at different locations in a distributed manner. They are aggregated through a global Internet-wide network. Mobile devices have evolved from being merely data-access devices to being capable of processing and storing significant amounts of data. Additionally, each new generation of mobile devices has better connectivity technologies, more storage memory and faster processors than previous generations. The Mobile Grid architecture embraces mobile resources to Grid, leverages the existing global Grid infrastructure capabilities by integrating mobile devices in a transparent way. Mobile Grid enables both the mobility of the users requesting access to a fixed Grid and the resources that are themselves part of the Grid [1]. For the formation of the mobile grid, there are two possible roles of mobile devices in grid. First, mobile devices can be used as interfaces to the grid. Thus, a mobile device can initiate the use of grid resources, monitor the jobs being executed remotely, and take any results from the grid in an ‘anytime, anywhere’ mode. Secondly and more interestingly, mobile devices can be assumed to participate in grid as computing resource providers, not just service recipients. Laptops and PDAs can provide aggregated computational capability when gathered in hotspots, forming a Grid on site. Recent advancement of technologies on mobile devices and wireless communications makes this scenario more feasible.
Currently, little research focuses on mobile devices as resource or service providers in mobile grid environment. In this paper, we consider mobile devices acting as service providers for grid users. The mobile grid user submits jobs to mobile grid sites, and one mobile grid site performing the job is comprised of mobile devices. The group of mobile devices in the mobile grid system is composed as a single service provider. The mobile device composing is achieved via the cooperation among a selected set of devices. The mobile grid user request can be associated with some end-to-end QoS requirements (like bandwidth, latency and price). The mobile device service composition has to ensure that the QoS values of the selected services match the user requirements.
Considering the difficulties of service composition of the mobile devices in mobile grid, this paper presents exploiting composition of mobile devices for maximizing user QoS under energy constraints in mobile grid. Mobile device service composition process includes two parts: mobile device service provisioning through device service market and mobile device resource allocation through device resource market. Interactions among mobile grid user agent, mobile device service agent and mobile device resource agent are mediated by means of market mechanisms. Mobile device service composition optimization maximizes the interests of mobile grid user agent, mobile device service agent and mobile device resource agent. Utility function is used to specify QoS requirement of mobile grid users and benefit of mobile grid resource providers. The problem of services composition of mobile devices is formulated by utility optimization. The paper also presents a mobile grid services composition algorithm to maximize user QoS under energy constraint. In the simulation, the performance evaluation of proposed iterative algorithm for services composition of mobile devices is conducted and compared with other related works.
The rest of the paper is structured as followings. Section 2 discusses the related works. Section 3 presents exploiting composition of mobile devices for maximizing user QoS under energy constraints in mobile grid. Section 4 describes mobile device service composition algorithms. In Section 5 the experiments are conducted and discussed. Section 6 gives the conclusions to the paper.
Section snippets
Related works
Mobile Grid enables both the mobility of the users requesting access to a fixed Grid and the resources that are themselves part of the Grid. Currently, some research focuses on resource scheduling and service provisioning in mobile grid environment. Lee et al. [2] present a group-based fault tolerance scheduling algorithm. The algorithm classifies mobile devices into several groups considering characteristic parameters of mobile Grid. Then, it uses an adaptive replication algorithm for enduring
Model description
The modeling of the service composition of mobile device in mobile grid is based on analyzing the related work such as [2], [3], [5], [9], which consider resource allocation and service provisioning in mobile grid environment. The mobile grid environment is shown Fig. 1. The mobile grid system consists of mobile devices and ordinary grid nodes. Wireless access points provide the access for mobile device, while ordinary grid nodes provide computing resources (e.g., CPU, memory, bandwidth and
Mobile device services composition algorithms in mobile grid
In mobile device service composition algorithm, different agents are used namely mobile device service agents, mobile device resource agents and mobile grid user agent. Mobile device resource market scheduler and mobile device service market scheduler are used for mobile device service provisioning and mobile device resource allocation. The mobile device service market scheduler starts a listening thread that listens to the requests from mobile grid users. It receives the QoS requirements of
Environment setup
In this section, we evaluate the performance of proposed mobile device services composition algorithm (MDSCA) using the JAVASIM [18] simulator. The experiments are conducted to compare our mobile device services composition algorithms (MDSCA) with other related algorithms for mobile grid environment. We focus on the two most important aspects of node mobility, i.e., the mobility model and the node speed. The wireless bandwidth is set to 11 Mbps. It is assumed that 60 mobile devices, each of
Conclusions
This paper presents exploiting composition of mobile devices for maximizing user QoS under energy constraints in mobile grid. Mobile device service composition process includes two parts: mobile device service provisioning through device service market and mobile device resource allocation through device resource market. Interactions among mobile grid user agent, mobile device service agent and mobile device resource agent are mediated by means of market mechanisms. Mobile device service
Acknowledgements
The authors thank the editor in chief and the anonymous reviewers for their helpful comments and suggestions. The work was supported by the National Natural Science Foundation (NSF) under grants (No. 61171075), Special Fund for Fast Sharing of Science Paper in Net Era by CSTD (FSSP) No. 2013014311021, Program for the High-end Talents of Hubei Province, Specialized Research Fund for the Doctoral Program of Higher Education under Grant No. 20120143110014 and the Open Fund of the State Key
References (24)
- et al.
Utility based QoS optimisation strategy for multi-criteria scheduling on the grid
J. Parallel Distrib. Comput.
(2007) - Hassan Jameel, Umar Kalim, Ali Sajjad, Sungyoung Lee, Taewoong Jeon, Mobile-to-Grid Middleware: Bridging the Gap...
- et al.
Group-based scheduling algorithm for fault tolerance in mobile grid
- et al.
Scheduling tasks in mobile grid environment using mobility based resource prediction
- et al.
The proxy-based mobile grid
- et al.
Mobility-aware efficient job scheduling in mobile grids
- T. Alwada’n, H. Janicke, O. Aldabbas, et al., New framework for policy support for mobile grid services, 2011 6th...
- J.M. Rodriguez, A. Zunino, M. Campo, Mobile grid seas: simple energy-aware scheduler, in Proc. 3rd High-Performance...
- et al.
Balanced scheduling algorithm considering availability in mobile grid
- et al.
Mobility and battery power prediction based job scheduling in mobile grid environment
Job management in mobile grid computing
Cited by (26)
Augmenting computing capabilities at the edge by jointly exploiting mobile devices: A survey
2018, Future Generation Computer SystemsCitation Excerpt :A set of experiments were designed to show the impact in energy consumption of SMDs with selfish and altruist owners profiles. In [59,60] the price of resources is defined by means of a non-cooperative bargaining game played by resource providers and resource consumers, and intermediated by resource brokers. In all cases, the price/reward for the usage of resources is obtained by assuming that tasks resource usage, i.e., execution time, energy spent, is known before the actual usage occurs which turn the approaches hard to be practiced.
A computational experiment-based evaluation method for context-aware services in complicated environment
2016, Information SciencesCitation Excerpt :To improve service quality consistently, service performance should be evaluated quantitatively. Considerable work has been conducted on the evaluation of service quality, focusing on finding the most appropriate QoS (Quality of Service) model [4,6,13,18]. It is the standard way to evaluate how each service solution can satisfy specific customer requirements.
Battery-aware centralized schedulers for CPU-bound jobs in mobile Grids
2016, Pervasive and Mobile ComputingCitation Excerpt :For these reasons, the scheduling logic is considered centralized. This is the case of the works by Li and Li [11,35,36], where a market based model for resource allocation takes place at the mediator component, called mobile Grid proxy. Gosh and Das’ [34,9] proposals are also centralized because of the existence of a central component called Job Allocator (JA).
Designing service-based applications in the presence of non-functional properties: A mapping study
2016, Information and Software TechnologySolving Task Scheduling Problems in Dew Computing via Deep Reinforcement Learning
2022, Applied Sciences (Switzerland)