Scheduling in Sensor Grid Middleware for Telemedicine Using ABC Algorithm

Hindawi Publishing Corporation International Journal of Telemedicine and Applications Volume 2014, Article ID 592342, 7 pages http://dx.doi.org/10.1155/2014/592342 Research Article Scheduling in Sensor Grid Middleware for Telemedicine Using ABC Algorithm T. Vigneswari1,2 and M. A. Maluk Mohamed2 1 Kings College of Engineering, Punalkulam, Tamilnadu 613303, India System Software Group, M.A.M College of Engineering, Tiruchirappalli, Tamilnadu 621105, India 2 Correspondence should be addressed to T. Vigneswari; Received 28 June 2014; Revised 2 November 2014; Accepted 11 November 2014; Published 3 December 2014 Academic Editor: Fei Hu Copyright © 2014 T. Vigneswari and M. A. M. Mohamed. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Advances in microelectromechanical systems (MEMS) and nanotechnology have enabled design of low power wireless sensor nodes capable of sensing different vital signs in our body. These nodes can communicate with each other to aggregate data and transmit vital parameters to a base station (BS). The data collected in the base station can be used to monitor health in real time. The patient wearing sensors may be mobile leading to aggregation of data from different BS for processing. Processing real time data is compute-intensive and telemedicine facilities may not have appropriate hardware to process the real time data effectively. To overcome this, sensor grid has been proposed in literature wherein sensor data is integrated to the grid for processing. This work proposes a scheduling algorithm to efficiently process telemedicine data in the grid. The proposed algorithm uses the popular swarm intelligence algorithm for scheduling to overcome the NP complete problem of grid scheduling. Results compared with other heuristic scheduling algorithms show the effectiveness of the proposed algorithm. 1. Introduction Telemedicine plays a very important role in patient management and has been effectively used for intrahospital transport of patients. Live monitoring of patients across hospitals creates new challenges. Similar issues arise as where to process the data captured in real time. Sensor grid can overcome some of the challenges faced in telemedicine. Human life can be made more comprehensive by equipping it with modern diagnosis which utilizes the advancement in the field of information and communication and this arena is termed as telemedicine. Sensor grid is one among the computing methods which can be used efficiently to provide such a service. Grid computing [1, 2] is a growing distributed computing paradigm which enables coordinated sharing of heterogeneous resources across the globe. Sensor networks have [3] group of sensor nodes that are deployed to receive real time values about the parameters which they sense. Sensor grid [4] is the integration of sensor network and grid computing by which both of the paradigms can complement each other with their own strengths. The advantages of merging sensor network and grid computing to form sensor grids [5, 6] include the following. (i) Large amount of real time data generated by sensors can be processed and stored in the grid. (ii) Set of sensors can be shared by different user based on the application they are using. (iii) Pervasive seamless access to sensor data is made possible. A sensor grid architecture to monitor patients who have undergone transplantation surgery is proposed in our previous work which includes the architecture for the sensor middleware [7, 8]. Scheduling component is a significant constituent in the proposed middleware. In general, scheduling discovers resources and allocates suitable tasks on appropriate resource to meet the requirement of the job handled. The major criteria for a good scheduling algorithm are response time, optimized resource utilization, load balancing, and meeting QoS constraints. 2 The scheduler in grid assigns jobs to the resources in an optimum way. Jobs arrive at the grid environment specifying the requirement about the resources. The assignment of jobs to the resources should be optimal to minimize the makespan, minimize the cost of allocated resources, and maximize the throughput [9]. Grid monitoring [10] collects the status and performance details of a large-scale distribution system. The parameters such as load on the system, number of jobs in the running state, and the performance of each job in running state are gathered to notify the behavior of the grid environment to the consumers. Various scheduling algorithms have been proposed in literature for the grid environment. All algorithms fall into the taxonomy proposed in [11]. The first level in taxonomy is local versus global scheduling algorithms. In local scheduling, scheduler schedules the processes available on a single CPU. Global scheduler allocates processes to multiple CPUs. Grid scheduling uses global scheduling. Global scheduling uses static and dynamic scheduling. In static scheduling, the information about the jobs and resources are available at the time of scheduling. In dynamic scheduling, it is impossible to predict the arrival time and resource request of jobs, earlier to the time of scheduling. First in first out (FIFO), balance constrained techniques, and cost constrained techniques are some of the techniques which have been used in grids [12]. Due to the NP completeness of grid scheduling, metaheuristics techniques like particle swarm optimization (PSO), ant colony optimization (ACO), genetic algorithm (GA), and artificial bee colony (ABC) have been used effectively for grid scheduling [13–16]. This paper proposes ABC algorithm that can be used in scheduling of resource for the middleware discussed in [8] to provide a scheduling solution which optimizes the makespan of the submitted task. 2. Related Works Telemedicine architecture using a sensor with P2P overlay has been proposed and a middleware has been developed for the abovementioned architecture. Among the middleware components, scheduling plays an inevitable role such that a proper scheduling may save a life. The data sets from the sensor should be scheduled to a computational resource for execution in the SaaS. The SaaS may send alert if the vital signs are abnormal and doctor reviews the patient immediately. An analysis of execution times of scheduling algorithms in [17] shows that taboo search algorithm is a suitable choice for applications including telemedicine involving static scheduling. The creation time is relatively small and simultaneously average execution time of the schedule is minimal. The total processor cycle consumption model [18] has shown to be useful for independent and coarse-grain task scheduling, that is, scheduling in which the computation time in grid nodes is superior to data transmission time. A resource-performance-flu (...truncated)