- Email: [email protected]
Towards a Middleware for Resource Sharing in Collaboration of Pervasive Computing
Available online at www.sciencedirect.com
ScienceDirect Procedia Computer Science 50 (2015) 87 – 92
2nd International Symposium on Big Data and Cloud Computing (ISBCC’15)
Towards a Middleware for Resource Sharing in Collaboration of Pervasive Computing A. Samydurai* a, C. Vijayakumaran,a G. Kumaresan,a K. Revathib
b
a Dept of CSE, Valliammai Engineering college, SRM Nagar, Kattankuulathur, Chennai-603203, India. Dept of CSE, Dhanalakshmi college of Engineering, Dr VPR Nagar, Manimangalam, Chennai- 601301, India
Abstract The way in which the computing devices and applications are developed and used does not satisfy the needs of the users calling for and shots of the potential for pervasive computing. It is crucial that the provides have full visibility of their context i.e., the set of available and relevant resources depending on access control rules, client locations user preferences, privacy requirements, terminal characteristics and current state of hosting environments. The applications developed for pervasive computing are very hard, but they are necessary to build ubiquitous computing environment. Few researchers have indeed focused on some of these requirements but there is a practical difference between theory and its achievement. A new model is proposed for pervasive computing. A solution to the existing problems is proposed through this paper. Some of the ways to improve the performance environments through the use of middleware is also proposed in this research work. © 2015 2015 The © The Authors. Authors. Published Publishedby byElsevier ElsevierB.V. B.V.This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of scientific committee of 2nd International Symposium on Big Data and Cloud Computing Peer-review under responsibility of scientific committee of 2nd International Symposium on Big Data and Cloud Computing (ISBCC’15). (ISBCC’15) Keywords: PRIM; Pervasive Computing; Middleware; Cluster
1. Introduction
Rapid growth of technology and devices makes to realize the pervasive computing in reality. Irrespective of the limitations of devices like low battery, minimum memory and expensive connections, the requirement of portable access to information becomes mandatory and unavoidable for the growing business cause. The motivation of pervasive computing is to force the computer and computing facilities to live with people in an interconnecting
* Corresponding author. Tel.: +91-944-357-7501; fax: + 044 - 27451504. E-mail address:[email protected]
1877-0509 © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of scientific committee of 2nd International Symposium on Big Data and Cloud Computing (ISBCC’15) doi:10.1016/j.procs.2015.04.065
88
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
manner. However, the applications running in pervasive computing faces major challenges on three characteristics, namely, sensitivity of context, connectivity, and ability to analyze. To make the pervasive computing more successful, it is necessary that the proposed system architecture should consider these three characteristics as a primary parameter to build the system for optimum utilization. Various peer to peer computation and communication devices are creating this environment to facilitate user’s everyday tasks and to increase productivity as well. To meet the existing demands in portable connecting it is necessary to provide a trustworthy connections and sensible analyzing system to prove that the middleware is worth while to adhere too. The system proposed in this paper contributes towards the solution to provide better connectivity and sensitivity to the context for enabling the proper service provision with limited resource utilization. The development of middleware to support pervasive computing requires more emphasis on context service, self organizing, load balancer, clustering and communication. To deploy the middleware based application in pervasive computing environment effectively it is proposed a context aware middleware model. This paper proposes a PRIM (Pervasive computing Realtime In collaboration of device through Middleware) model to directly addresses the concerns of heterogeneous devices, limited connecting and autonomous administration to make it feasible to develop and deploy applications on a global scale. Pervasive computing4,5 on its general terms can be viewed as accessing data and the mechanisms needed to support a community of nomadic users. On device-centric view, it could be viewed as focusing on how best to deploy new function on a device exploiting its interface modalities for a specific task. Hardware and network infrastructure to realize this vision are increasingly becoming a reality. In addition to this, the network facilities like blue tooth and World Wide Web facilities along with cell-phones and palm-sized computing devices makes pervasive computing from research area to a commercial reality. Pervasive computing mainly focuses1 on three key areas. First, it defines the way by which the computing devices are viewed and use them within their environments to perform takes. Secondly, it defines the way that the applications are created and deployed. Thirdly, it concerns about the environment. The main contribution of this paper is to collaborate the mobile devices and PC’s in real-time and to make them adaptable and providing a general computing environment. The rest of the paper is structured in such a way to deal with the proposed system model for better middleware logic and performance characteristics associated with the system. The system design is illustrated in the following section followed by performance issues in the next section. 2. System
design
The proposed architecture for effective communication and efficient clustering logic is shown in Fig.1. The architecture PRIM major focuses on services and core components. The core components include the connection management, transaction management, and security management. It will also ensure the clustering and communication. The service oriented sector of this architecture will emphasis on contest sensitivity self organizing and load balancing . Application Object
Context service
Self organizing
Load Balancer
Core Component Connection Management
Transaction Management
Clustering
Security Management
Communication Agent Sensors
Fig. 1. PRIM Architecture
Operating System
Service
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
2.1. Content Service The main functionality of the context services is to identify the remote resource which is potential enough to support the requesting services. The context event service provider is a mechanism for asynchronous communication. It also maintains a trigger repository for notifying consumers about event that have occurred. 2.2. Service Self organizing The key services in the proposed architecture are the self organizing service. The system has a self organizer (a managerial unit) which will ensure to look after any system failure basically a fault tolerant system. To ensure supreme quality of services for the requesting entity, the unit will periodically monitors as well as organizes various events to supply the demand of request. It maintains various status calls like ‘Success’ or ‘Failure’ based on the service status. Any interruption to the service half way through being traced based on the status and the required healing action will be taken by managerial unit to recover and handover resources issues for ensuring the service is uninterrupted. 2.3. Load balancer The Load balancer unit of the system needs to look after the process of balancing the load request based on the network traffic and resource utilization. Either on the situation of failure or for effective usage of resources, the load balancer unit will redirect request to the allotted resources keeping all the informative parameters in to consideration. The load balancing unit will also ensure that the response for the request doesn’t extends from the response period time offered in general by the system. The load balancing units are tested by varying system parameters like the CPU utilization, number of threads; memory factory, throughput and response time, then parameter are calculated. 2.4 Core components The Core components of the middleware proposed in this system as basically three management units. There are connection management, transaction management, and security management. The connection management meant for the purpose of the establishing connection between the request entry and service providers. The transaction management will ensure the transfer of information and recovery of loss of data in case of failure. The security management system is to provide security to the data transferred and being transformed in the specified format for clear presentation. 2.5 Clustering and communication The Clustering and communication component of middleware makes the proposed system more viable. Any failure in the proposes of request may not end up with toss, by having clustering unit based on the level of completion on alternate performing unit is identified and the transfer of work will take place. The system will ensure that the interrupted process continue from the point where it has failed instead of starting it from scratch. The clustering and communication unit will take the necessary action to transfer the required data to alternate executing resources to complete the process and reproduce the expected results. During this execution, the process of completion should be taken the time factor into consideration. 3.
Experimental approach
As and when a client request for a service, the middleware will acknowledge with the client and identifies the task coordinator unit from the pool of service providers. The task coordinator will generate subtask of operation and recollect the subtask as specified in integrated resource algorithm.
89
90
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
Clients Middleware
Server Providers Fig. 2. Coordinator Service Provide
To facilitate the optimum service provision, the coordinator unit of service provider will apply the integration of resource action event to obtain the faster results as in Fig. 2. To integrate the T1, T2, ..., Tn to get a particular task T, will deploy the following algorithm which will act as transaction function. Procedure Integrate Resource () SPi = The service provider of task Ti SPq = The queuing devices of resources Action [i] = Action required to initiate task i. i=1 While i <= n SPi complete the task Ti If (i = n) Exit else action [ i+1 ] and Ti to SPq i = i+1 endif endwhile 4.
Related Work
A model is proposed1 which is characterized by a devices independent application development process which includes abstract specification of the application front end and application’s resources and service requirements. Another model is proposed2 to simplify the task of building reliable application, all operations on tasks to group several operations into a single atomic unit. The paper outlined3 the paradigm of potential in pervasive computing especially for work context characteristics by mobility and adhoc uses of mobile devices. The MARKS architecture4 which addresses the knowledge usability, resource discovery and self healing properties of pervasive computing and also proposed5 the discovery mechanism named SAFE–RD which is an integral part of the MARKS architecture. A system is designed6 which highlighted the key characteristics of pervasive middleware to support context awareness and service discover and adaptation. A context aware middleware7 for the just in time deployment of the component based applications supporting the context aware adaptation. The implementation of software infrastructure 8, 9 for building augmented reality applications for ubiquitous computing environments. A security middleware architecture10, 11 for heterogeneous pervasive devices is provided to collaborate with high security confidence. Two dynamic load balancing unit schemes12 for multiuser (muli-class) j o b s in heterogeneous distributed system. One tries to minimize the expected response time of entire system while the other tries to minimize the expected response time of the individual users. T echnologies, sensor networks, and their related applications in pervasive computing is presented15 to discuss about the four areas for new research interaction, software agents expert system and artificial intelligence. Computing
91
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
system16, physical integration and spontaneous interoperation. Also additionally Unicom software must operate in conditions of radical change. The authors have been proposed17, 18 a distributed environment with number of middleware, servers and clients. It utilizes QoS parameters such as object’s memory, load, CPU usage and object function execution to improve the QoS by providing fault tolerance. The middleware receives the inputs from the client and forward them to servers that utilize the five parameters of QoS in improving the reliability through high fault tolerance in distributed system. 5.
Performance Measurements of Load Balancing Unit
Dynamic load balancing the following five parameters. They are number of request, free memory, CPU utilization number of threads, parameters like the request vs. free memory, request vs.CPU request vs. threads and request vs. response time. We have to implement using software’s such as Linux platform and JBOSS. The JMX is used for management of load balancing and clustering unit. Servlet filter used to implement the load balancing unit. Accessing the middleware with any one of the free servers or devices. The input request for the system is given in the following Table1: performance analysis of load balancing unit. Table 1 Performances analysis of load balancing unit Number of request 1 45 69 125 165 460
CPU utilization 11 11 10 11 11 13
Free memory 87030112 84783320 85738712 91261256 89386528 79534384
Number of thread 135 138 145 145 145 147
Response time 1 2 2 2 2 2
The fig. 3 shows the available of free request increase the available memory is not reduced. It can be observed that at low and medium system loads the performance of static and dynamic is similar. Some times the high request expected free memory increased or decreased. Increase in the availability will not reduce the free memory. The fig. 4 shows the CPU utilization on dependence of number of request. The CPU utilization will be completed with minimum time even the number of request is more. The fig. 5 shows the created number of threads in corresponding request. Once the number of thread created for a single request, then there is number maximum deviation in the thread with the increase in the number of request. The Fig. 6 shows the response time for each request. If the request is one then the response time is one, if more than one request is given the response time will be low, further the number of request is increased then the response time will be reduced. Therefore the load balancing unit found the system which reduces response time hence increasing durability. 500
1500
400
1000
300 200
500
100
0
0
1
2
3
No. of R eques t
4
5
6
F ree Memory
Fig.3. Performance of Number of request vs. free memory
1
2
3
No. of R eques t
4
5
6
C PU Utiliz ation
Fig. 4. Performance of Number of request vs. CPU utilization
92
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
500
500
400
400
300
300
200
200 100
100
0
0 1
2
3
No. of R eques t
4
5
6
Number of Thread
Fig. 5. Performance of Number of request vs. Number of Threads
1
2
3
No. of R eques t
4
5
6
R es pons e Time
Fig. 6. Performance of Number of Request vs. Response Time
6. Conclusion
In this paper, the basic idea of pervasive computing through collaboration of devices was outlined. The paper mainly focuses on collaboration of devices and a proposal to support this collaboration. For pervasive computing to meet the exception of the mobile users fundamentally changes the need to occur in the way people perceive the roll of devices. The computing environment must be recognized as the extension of users surrounding not a virtual space for hosting and running program. Also the network environments must be cryptographically strong. This view is treated as a future work for the researches to develop in the area of pervasive computing. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Guruduth B, James B, Eugene G, Jonathan M, Jeremy S, and Deborra Z, Challenges: An Application Model for Pervasive Computing Proc.international conference on Mobile computing and networking, 2000, pp. 266-274. Robert G, Tom A, Brian B, and David Wl , A System Architecture for Pervasive Computing, Proc. on ACM SIGOPS European workshop: beyond the PC: new challenges for the operating system , 2000 , pp.177 - 182 . Jakob E. Bardram and Henrik Bærbak Christensen, Real-Time Collaboration in Activity Based Architectures, CfPC Technical Report, 2008, pp.1-10. Moushumi S, Shameem A, and Sheilkh I.A , MARKS (Middleware Adaptability for Resource Discovery, Knowledge Usability and Self-healing) for Mobile Devices of Pervasive Computing Environment, Proceedings of the Third international Conference on information TechnologyNewGenerations GCC’06, IEEE, 2006, pp. 306 – 313. Moushumi Sharmin, Shameem Ahmed, and Sheilkh I. Ahamed, SAFE – RD( Secure, Adaptive Faullt Tolerant,and Eficient Resource Discovery), international Conference on information Technology Coding and Computing (ITCC’05), IEEE, 2005, pp. 271- 276. Amitava Mukherjee, Debashis saha, Chinmohan Biswas, Present Scenarios future Challenges in pervasive Middleware, IEEE, 2006, pp.1-5. Di Zheng, Jun Wang, Weihong Han, Yan Jia Peng Zou, Towards A Context-aware middleware for deploying Component Based in Application in pervasive computing, Proc.of the 5thinternational conference on Grid and Corporative computing, 2006, pp.454-457. Richardo Couto Antunes De Rocha, Markus Endler, Middleware: Context Management in Heterogeneous, Evolving Ubiquitous Environment, IEEE Distributed System Online, April , 2006 ,Vol.7 No.4., pp-1. Eiji T., Andrej VDZ, Makoto K, Masahiro N, Tatsuo N, Object-Oriented Middleware Infrastructure for Distributed Augmented Reality, Proceedings of the Sixth IEEE international Symposium on Object-Oriented Real-Time Distributed Computing, 2003, pp.156-163. Nick Cook , Santosh Shrivastava and Stuart Wheater, Distributed Object Middlware to support Dependable Information sharing between organizations, Proceedings of the international Conference on Dependable system and Networks, IEEE, 2002, pp.1-10. Zhaoyu Liu and Dichao Peng, A Security- Supportive Middleware Architecture for pervasive Computing, Proceedings of the 2nd IEEE international Symposium on Dependable, Autonomic and secure Computing (DASC’06), IEEE, 2006, pp.137-144. Satish P. and Anthony T.C., Dynamic Multi-User Load Balancing in Distributed Systems, Proceeding of the International Symposium on Parallel and Distributed Processing IEEE Proceeding, 2007, pp.1-10. Karen Henricksen Jadwiga Indulska and Andry Rakotonirainy, Infrastructure for Pervasive Computing :Challenges, In Informatics Workshop on Pervasive Computing and Information Logistics, 2001, pp. 214-222. Albrecht S., Sarah S., Anatole G., and Florian M., Real-World Challenges of Pervasive Computing, Published by the IEEE CS, 2006, pp.91-94. M. Satyanarayanan, Pervasive Computing: Vision and Challenges, IEEE Personal Communication 2001, pp.1-10. Tim, K., Armando, F, System Software for Ubiquitous Computing, IEEE Pervasive Computing, 2002, Vol.1, No.1, pp. 70-81. A. Samydurai and A. Shanmugam, 2014. Fault Tolerant Middleware Framework to Improve QoS in Distributed Systems. Research Journal of Applied Sciences, vol. 9, pp.154-159. A. Samydurai and A. Shanmugam, Fault Tolerant Middleware and Server Side Data Replication Framework to Improve QoS in Distributed Systems, International Journal Wulfenia, 2013, no.4,vol.20, pp.121-135.
ScienceDirect Procedia Computer Science 50 (2015) 87 – 92
2nd International Symposium on Big Data and Cloud Computing (ISBCC’15)
Towards a Middleware for Resource Sharing in Collaboration of Pervasive Computing A. Samydurai* a, C. Vijayakumaran,a G. Kumaresan,a K. Revathib
b
a Dept of CSE, Valliammai Engineering college, SRM Nagar, Kattankuulathur, Chennai-603203, India. Dept of CSE, Dhanalakshmi college of Engineering, Dr VPR Nagar, Manimangalam, Chennai- 601301, India
Abstract The way in which the computing devices and applications are developed and used does not satisfy the needs of the users calling for and shots of the potential for pervasive computing. It is crucial that the provides have full visibility of their context i.e., the set of available and relevant resources depending on access control rules, client locations user preferences, privacy requirements, terminal characteristics and current state of hosting environments. The applications developed for pervasive computing are very hard, but they are necessary to build ubiquitous computing environment. Few researchers have indeed focused on some of these requirements but there is a practical difference between theory and its achievement. A new model is proposed for pervasive computing. A solution to the existing problems is proposed through this paper. Some of the ways to improve the performance environments through the use of middleware is also proposed in this research work. © 2015 2015 The © The Authors. Authors. Published Publishedby byElsevier ElsevierB.V. B.V.This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of scientific committee of 2nd International Symposium on Big Data and Cloud Computing Peer-review under responsibility of scientific committee of 2nd International Symposium on Big Data and Cloud Computing (ISBCC’15). (ISBCC’15) Keywords: PRIM; Pervasive Computing; Middleware; Cluster
1. Introduction
Rapid growth of technology and devices makes to realize the pervasive computing in reality. Irrespective of the limitations of devices like low battery, minimum memory and expensive connections, the requirement of portable access to information becomes mandatory and unavoidable for the growing business cause. The motivation of pervasive computing is to force the computer and computing facilities to live with people in an interconnecting
* Corresponding author. Tel.: +91-944-357-7501; fax: + 044 - 27451504. E-mail address:[email protected]
1877-0509 © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of scientific committee of 2nd International Symposium on Big Data and Cloud Computing (ISBCC’15) doi:10.1016/j.procs.2015.04.065
88
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
manner. However, the applications running in pervasive computing faces major challenges on three characteristics, namely, sensitivity of context, connectivity, and ability to analyze. To make the pervasive computing more successful, it is necessary that the proposed system architecture should consider these three characteristics as a primary parameter to build the system for optimum utilization. Various peer to peer computation and communication devices are creating this environment to facilitate user’s everyday tasks and to increase productivity as well. To meet the existing demands in portable connecting it is necessary to provide a trustworthy connections and sensible analyzing system to prove that the middleware is worth while to adhere too. The system proposed in this paper contributes towards the solution to provide better connectivity and sensitivity to the context for enabling the proper service provision with limited resource utilization. The development of middleware to support pervasive computing requires more emphasis on context service, self organizing, load balancer, clustering and communication. To deploy the middleware based application in pervasive computing environment effectively it is proposed a context aware middleware model. This paper proposes a PRIM (Pervasive computing Realtime In collaboration of device through Middleware) model to directly addresses the concerns of heterogeneous devices, limited connecting and autonomous administration to make it feasible to develop and deploy applications on a global scale. Pervasive computing4,5 on its general terms can be viewed as accessing data and the mechanisms needed to support a community of nomadic users. On device-centric view, it could be viewed as focusing on how best to deploy new function on a device exploiting its interface modalities for a specific task. Hardware and network infrastructure to realize this vision are increasingly becoming a reality. In addition to this, the network facilities like blue tooth and World Wide Web facilities along with cell-phones and palm-sized computing devices makes pervasive computing from research area to a commercial reality. Pervasive computing mainly focuses1 on three key areas. First, it defines the way by which the computing devices are viewed and use them within their environments to perform takes. Secondly, it defines the way that the applications are created and deployed. Thirdly, it concerns about the environment. The main contribution of this paper is to collaborate the mobile devices and PC’s in real-time and to make them adaptable and providing a general computing environment. The rest of the paper is structured in such a way to deal with the proposed system model for better middleware logic and performance characteristics associated with the system. The system design is illustrated in the following section followed by performance issues in the next section. 2. System
design
The proposed architecture for effective communication and efficient clustering logic is shown in Fig.1. The architecture PRIM major focuses on services and core components. The core components include the connection management, transaction management, and security management. It will also ensure the clustering and communication. The service oriented sector of this architecture will emphasis on contest sensitivity self organizing and load balancing . Application Object
Context service
Self organizing
Load Balancer
Core Component Connection Management
Transaction Management
Clustering
Security Management
Communication Agent Sensors
Fig. 1. PRIM Architecture
Operating System
Service
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
2.1. Content Service The main functionality of the context services is to identify the remote resource which is potential enough to support the requesting services. The context event service provider is a mechanism for asynchronous communication. It also maintains a trigger repository for notifying consumers about event that have occurred. 2.2. Service Self organizing The key services in the proposed architecture are the self organizing service. The system has a self organizer (a managerial unit) which will ensure to look after any system failure basically a fault tolerant system. To ensure supreme quality of services for the requesting entity, the unit will periodically monitors as well as organizes various events to supply the demand of request. It maintains various status calls like ‘Success’ or ‘Failure’ based on the service status. Any interruption to the service half way through being traced based on the status and the required healing action will be taken by managerial unit to recover and handover resources issues for ensuring the service is uninterrupted. 2.3. Load balancer The Load balancer unit of the system needs to look after the process of balancing the load request based on the network traffic and resource utilization. Either on the situation of failure or for effective usage of resources, the load balancer unit will redirect request to the allotted resources keeping all the informative parameters in to consideration. The load balancing unit will also ensure that the response for the request doesn’t extends from the response period time offered in general by the system. The load balancing units are tested by varying system parameters like the CPU utilization, number of threads; memory factory, throughput and response time, then parameter are calculated. 2.4 Core components The Core components of the middleware proposed in this system as basically three management units. There are connection management, transaction management, and security management. The connection management meant for the purpose of the establishing connection between the request entry and service providers. The transaction management will ensure the transfer of information and recovery of loss of data in case of failure. The security management system is to provide security to the data transferred and being transformed in the specified format for clear presentation. 2.5 Clustering and communication The Clustering and communication component of middleware makes the proposed system more viable. Any failure in the proposes of request may not end up with toss, by having clustering unit based on the level of completion on alternate performing unit is identified and the transfer of work will take place. The system will ensure that the interrupted process continue from the point where it has failed instead of starting it from scratch. The clustering and communication unit will take the necessary action to transfer the required data to alternate executing resources to complete the process and reproduce the expected results. During this execution, the process of completion should be taken the time factor into consideration. 3.
Experimental approach
As and when a client request for a service, the middleware will acknowledge with the client and identifies the task coordinator unit from the pool of service providers. The task coordinator will generate subtask of operation and recollect the subtask as specified in integrated resource algorithm.
89
90
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
Clients Middleware
Server Providers Fig. 2. Coordinator Service Provide
To facilitate the optimum service provision, the coordinator unit of service provider will apply the integration of resource action event to obtain the faster results as in Fig. 2. To integrate the T1, T2, ..., Tn to get a particular task T, will deploy the following algorithm which will act as transaction function. Procedure Integrate Resource () SPi = The service provider of task Ti SPq = The queuing devices of resources Action [i] = Action required to initiate task i. i=1 While i <= n SPi complete the task Ti If (i = n) Exit else action [ i+1 ] and Ti to SPq i = i+1 endif endwhile 4.
Related Work
A model is proposed1 which is characterized by a devices independent application development process which includes abstract specification of the application front end and application’s resources and service requirements. Another model is proposed2 to simplify the task of building reliable application, all operations on tasks to group several operations into a single atomic unit. The paper outlined3 the paradigm of potential in pervasive computing especially for work context characteristics by mobility and adhoc uses of mobile devices. The MARKS architecture4 which addresses the knowledge usability, resource discovery and self healing properties of pervasive computing and also proposed5 the discovery mechanism named SAFE–RD which is an integral part of the MARKS architecture. A system is designed6 which highlighted the key characteristics of pervasive middleware to support context awareness and service discover and adaptation. A context aware middleware7 for the just in time deployment of the component based applications supporting the context aware adaptation. The implementation of software infrastructure 8, 9 for building augmented reality applications for ubiquitous computing environments. A security middleware architecture10, 11 for heterogeneous pervasive devices is provided to collaborate with high security confidence. Two dynamic load balancing unit schemes12 for multiuser (muli-class) j o b s in heterogeneous distributed system. One tries to minimize the expected response time of entire system while the other tries to minimize the expected response time of the individual users. T echnologies, sensor networks, and their related applications in pervasive computing is presented15 to discuss about the four areas for new research interaction, software agents expert system and artificial intelligence. Computing
91
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
system16, physical integration and spontaneous interoperation. Also additionally Unicom software must operate in conditions of radical change. The authors have been proposed17, 18 a distributed environment with number of middleware, servers and clients. It utilizes QoS parameters such as object’s memory, load, CPU usage and object function execution to improve the QoS by providing fault tolerance. The middleware receives the inputs from the client and forward them to servers that utilize the five parameters of QoS in improving the reliability through high fault tolerance in distributed system. 5.
Performance Measurements of Load Balancing Unit
Dynamic load balancing the following five parameters. They are number of request, free memory, CPU utilization number of threads, parameters like the request vs. free memory, request vs.CPU request vs. threads and request vs. response time. We have to implement using software’s such as Linux platform and JBOSS. The JMX is used for management of load balancing and clustering unit. Servlet filter used to implement the load balancing unit. Accessing the middleware with any one of the free servers or devices. The input request for the system is given in the following Table1: performance analysis of load balancing unit. Table 1 Performances analysis of load balancing unit Number of request 1 45 69 125 165 460
CPU utilization 11 11 10 11 11 13
Free memory 87030112 84783320 85738712 91261256 89386528 79534384
Number of thread 135 138 145 145 145 147
Response time 1 2 2 2 2 2
The fig. 3 shows the available of free request increase the available memory is not reduced. It can be observed that at low and medium system loads the performance of static and dynamic is similar. Some times the high request expected free memory increased or decreased. Increase in the availability will not reduce the free memory. The fig. 4 shows the CPU utilization on dependence of number of request. The CPU utilization will be completed with minimum time even the number of request is more. The fig. 5 shows the created number of threads in corresponding request. Once the number of thread created for a single request, then there is number maximum deviation in the thread with the increase in the number of request. The Fig. 6 shows the response time for each request. If the request is one then the response time is one, if more than one request is given the response time will be low, further the number of request is increased then the response time will be reduced. Therefore the load balancing unit found the system which reduces response time hence increasing durability. 500
1500
400
1000
300 200
500
100
0
0
1
2
3
No. of R eques t
4
5
6
F ree Memory
Fig.3. Performance of Number of request vs. free memory
1
2
3
No. of R eques t
4
5
6
C PU Utiliz ation
Fig. 4. Performance of Number of request vs. CPU utilization
92
A. Samydurai et al. / Procedia Computer Science 50 (2015) 87 – 92
500
500
400
400
300
300
200
200 100
100
0
0 1
2
3
No. of R eques t
4
5
6
Number of Thread
Fig. 5. Performance of Number of request vs. Number of Threads
1
2
3
No. of R eques t
4
5
6
R es pons e Time
Fig. 6. Performance of Number of Request vs. Response Time
6. Conclusion
In this paper, the basic idea of pervasive computing through collaboration of devices was outlined. The paper mainly focuses on collaboration of devices and a proposal to support this collaboration. For pervasive computing to meet the exception of the mobile users fundamentally changes the need to occur in the way people perceive the roll of devices. The computing environment must be recognized as the extension of users surrounding not a virtual space for hosting and running program. Also the network environments must be cryptographically strong. This view is treated as a future work for the researches to develop in the area of pervasive computing. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Guruduth B, James B, Eugene G, Jonathan M, Jeremy S, and Deborra Z, Challenges: An Application Model for Pervasive Computing Proc.international conference on Mobile computing and networking, 2000, pp. 266-274. Robert G, Tom A, Brian B, and David Wl , A System Architecture for Pervasive Computing, Proc. on ACM SIGOPS European workshop: beyond the PC: new challenges for the operating system , 2000 , pp.177 - 182 . Jakob E. Bardram and Henrik Bærbak Christensen, Real-Time Collaboration in Activity Based Architectures, CfPC Technical Report, 2008, pp.1-10. Moushumi S, Shameem A, and Sheilkh I.A , MARKS (Middleware Adaptability for Resource Discovery, Knowledge Usability and Self-healing) for Mobile Devices of Pervasive Computing Environment, Proceedings of the Third international Conference on information TechnologyNewGenerations GCC’06, IEEE, 2006, pp. 306 – 313. Moushumi Sharmin, Shameem Ahmed, and Sheilkh I. Ahamed, SAFE – RD( Secure, Adaptive Faullt Tolerant,and Eficient Resource Discovery), international Conference on information Technology Coding and Computing (ITCC’05), IEEE, 2005, pp. 271- 276. Amitava Mukherjee, Debashis saha, Chinmohan Biswas, Present Scenarios future Challenges in pervasive Middleware, IEEE, 2006, pp.1-5. Di Zheng, Jun Wang, Weihong Han, Yan Jia Peng Zou, Towards A Context-aware middleware for deploying Component Based in Application in pervasive computing, Proc.of the 5thinternational conference on Grid and Corporative computing, 2006, pp.454-457. Richardo Couto Antunes De Rocha, Markus Endler, Middleware: Context Management in Heterogeneous, Evolving Ubiquitous Environment, IEEE Distributed System Online, April , 2006 ,Vol.7 No.4., pp-1. Eiji T., Andrej VDZ, Makoto K, Masahiro N, Tatsuo N, Object-Oriented Middleware Infrastructure for Distributed Augmented Reality, Proceedings of the Sixth IEEE international Symposium on Object-Oriented Real-Time Distributed Computing, 2003, pp.156-163. Nick Cook , Santosh Shrivastava and Stuart Wheater, Distributed Object Middlware to support Dependable Information sharing between organizations, Proceedings of the international Conference on Dependable system and Networks, IEEE, 2002, pp.1-10. Zhaoyu Liu and Dichao Peng, A Security- Supportive Middleware Architecture for pervasive Computing, Proceedings of the 2nd IEEE international Symposium on Dependable, Autonomic and secure Computing (DASC’06), IEEE, 2006, pp.137-144. Satish P. and Anthony T.C., Dynamic Multi-User Load Balancing in Distributed Systems, Proceeding of the International Symposium on Parallel and Distributed Processing IEEE Proceeding, 2007, pp.1-10. Karen Henricksen Jadwiga Indulska and Andry Rakotonirainy, Infrastructure for Pervasive Computing :Challenges, In Informatics Workshop on Pervasive Computing and Information Logistics, 2001, pp. 214-222. Albrecht S., Sarah S., Anatole G., and Florian M., Real-World Challenges of Pervasive Computing, Published by the IEEE CS, 2006, pp.91-94. M. Satyanarayanan, Pervasive Computing: Vision and Challenges, IEEE Personal Communication 2001, pp.1-10. Tim, K., Armando, F, System Software for Ubiquitous Computing, IEEE Pervasive Computing, 2002, Vol.1, No.1, pp. 70-81. A. Samydurai and A. Shanmugam, 2014. Fault Tolerant Middleware Framework to Improve QoS in Distributed Systems. Research Journal of Applied Sciences, vol. 9, pp.154-159. A. Samydurai and A. Shanmugam, Fault Tolerant Middleware and Server Side Data Replication Framework to Improve QoS in Distributed Systems, International Journal Wulfenia, 2013, no.4,vol.20, pp.121-135.