International Journal of High Performance Computing and Networking http://www.inderscience.com/browse/index.php?journalID=61&year=2011&vol=7&issue=1 Inderscience Publishers Ltd en-uk International Journal of High Performance Computing and Networking 1740-0562 1740-0570 © 2011 Inderscience Publishers Ltd editor@inderscience.com https://www.inderscience.com/images/files/coverImgs/ijhpcn_scoverijhpcn.jpg http://www.inderscience.com/browse/index.php?journalID=61&year=2011&vol=7&issue=1 http://www.inderscience.com/link.php?id=38704 In this study, we model pricing of grid/distributed computing resources as a problem of real option pricing. Grid resources are non-storable compute commodities (e.g., CPU cycles, memory, etc.). The non-storable characteristic feature of the grid resources hinders it from fitting into a risk-adjusted spot price model for pricing financial options. Grid resources users pay upfront to acquire and use grid compute cycles in the future, for example, six months. The user expects a high and acceptable degree of satisfaction expressed as the quality of service (QoS) assurance. This requirement further imposes service constraints on the grid because it must provide a user-acceptable QoS guarantee to compensate for the upfront value. This study integrates three threads of our research; pricing the grid compute cycles as a problem of real option pricing, modelling grid resources spot price using a discrete time approach, and addressing uncertainty constraints in the provision of QoS using fuzzy logic. We have proved the feasibility of this model through experiments and we have presented some of our pricing results and discussed them. Grid resources valuation with fuzzy real option
David Allenotor, Ruppa K. Thulasiram
International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 1 - 7
In this study, we model pricing of grid/distributed computing resources as a problem of real option pricing. Grid resources are non-storable compute commodities (e.g., CPU cycles, memory, etc.). The non-storable characteristic feature of the grid resources hinders it from fitting into a risk-adjusted spot price model for pricing financial options. Grid resources users pay upfront to acquire and use grid compute cycles in the future, for example, six months. The user expects a high and acceptable degree of satisfaction expressed as the quality of service (QoS) assurance. This requirement further imposes service constraints on the grid because it must provide a user-acceptable QoS guarantee to compensate for the upfront value. This study integrates three threads of our research; pricing the grid compute cycles as a problem of real option pricing, modelling grid resources spot price using a discrete time approach, and addressing uncertainty constraints in the provision of QoS using fuzzy logic. We have proved the feasibility of this model through experiments and we have presented some of our pricing results and discussed them.

]]> 10.1504/IJHPCN.2011.038704 International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 1 - 7 David Allenotor Ruppa K. Thulasiram Department of Computer Science, University of Manitoba, R3T 2N2, Canada. ' Department of Computer Science, University of Manitoba, R3T 2N2, Canada grid computing grid resource pricing financial options modelling fuzzy real options QoS quality of service distributed computing spot price models uncertainty constraints fuzzy logic. 2011-02-23T23:20:50-05:00 7 1 1 7 2011-02-23T23:20:50-05:00 http://www.inderscience.com/link.php?id=38706 Excessive run time variability of parallel application codes on commodity clusters is a significant challenge. To gain insight into this problem, our earlier work developed tools to emulate parallel applications (PACE) by simulating computation and using the cluster's interconnection network for communication, and further study parallel application run time sensitivity effects to controlled network performance degradation (PARSE). This work expands our previous efforts by presenting a metric derived from PARSE test results conducted on several widely used parallel benchmarks and application code fragments. The metric suggests that a parallel application's sensitivity to network performance variation can be quantified relative to its behaviour in optimal network performance conditions. Ideas on how this metric can be useful to parallel application development, cluster system performance management and system administration are also presented. A network performance sensitivity metric for parallel applications
Jeffrey J. Evans, Cynthia S. Hood
International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 8 - 18
Excessive run time variability of parallel application codes on commodity clusters is a significant challenge. To gain insight into this problem, our earlier work developed tools to emulate parallel applications (PACE) by simulating computation and using the cluster's interconnection network for communication, and further study parallel application run time sensitivity effects to controlled network performance degradation (PARSE). This work expands our previous efforts by presenting a metric derived from PARSE test results conducted on several widely used parallel benchmarks and application code fragments. The metric suggests that a parallel application's sensitivity to network performance variation can be quantified relative to its behaviour in optimal network performance conditions. Ideas on how this metric can be useful to parallel application development, cluster system performance management and system administration are also presented.

]]> 10.1504/IJHPCN.2011.038706 International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 8 - 18 Jeffrey J. Evans Cynthia S. Hood Electrical and Computer Engineering Technology, Purdue University, West Lafayette, IN 47907, USA. ' Department of Computer Science, Illinois Institute of Technology, Chicago, IL 60616, USA parallel applications run time sensitivity high performance networks network evaluation performance management run time variability commodity clusters simulation performance degradation performance variation. 2011-02-23T23:20:50-05:00 7 1 8 18 2011-02-23T23:20:50-05:00 http://www.inderscience.com/link.php?id=38707 We propose two deterministic key predistribution schemes in a wireless sensor network (WSN), in which sensor nodes are deployed randomly. Both the schemes are based on combinatorial designs, called partially balanced incomplete block designs (PBIBD). An important feature of our scheme is that every pair of nodes within communication range can communicate directly, making communication faster and efficient. The number of keys per node is of the order of $\sqrt N$, where $N$ is the number of nodes in the network. Our second design has the added advantage that we can introduce new nodes in the network keeping the key pool fixed. This makes the network scalable. We study the resiliency of the network under node compromise and show that our designs give better results than the existing ones. Key pre-distribution using partially balanced designs in wireless sensor networks
Sushmita Ruj, Bimal Roy
International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 19 - 28
We propose two deterministic key predistribution schemes in a wireless sensor network (WSN), in which sensor nodes are deployed randomly. Both the schemes are based on combinatorial designs, called partially balanced incomplete block designs (PBIBD). An important feature of our scheme is that every pair of nodes within communication range can communicate directly, making communication faster and efficient. The number of keys per node is of the order of $\sqrt N$, where $N$ is the number of nodes in the network. Our second design has the added advantage that we can introduce new nodes in the network keeping the key pool fixed. This makes the network scalable. We study the resiliency of the network under node compromise and show that our designs give better results than the existing ones.

]]> 10.1504/IJHPCN.2011.038707 International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 19 - 28 Sushmita Ruj Bimal Roy School of Information Technology and Engineering, University of Ottawa, 800 King Edward Ottawa, Ontario K1N6N5, Canada. ' Applied Statistics Unit, Indian Statistical Institute, 203 B T Road, Kolkata 700 108, India combinatorial design PBIBD resiliency wireless sensor networks WSNs wireless networks key predistribution sensor node deployment partially balanced designs. 2011-02-23T23:20:50-05:00 7 1 19 28 2011-02-23T23:20:50-05:00 http://www.inderscience.com/link.php?id=38708 Ensuring that a software system meets its objectives requires continuous monitoring. In practice, monitoring is either insufficient to effectively detect and diagnose failures, or is too costly to use in production. An alternative is adaptive monitoring, where the system is monitored at a minimal level to determine system health, and if a problem is suspected, the monitoring level is automatically increased to determine faults. To model the system at different monitoring levels, we employ statistical techniques to identify stable relationships in the monitored data. These relationships characterise normal operation and can help detect anomalies. We describe our approach in the context of a J2EE-based system. We show that adaptive monitoring is a cost-effective alternative to continuous detailed monitoring. We inject 29 different faults, and show that we detect the faults in 80% of cases and shortlist the faulty component in 65% of the detected cases. Leveraging many simple statistical models to adaptively monitor software systems
Mohammad A. Munawar, Paul A.S. Ward
International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 29 - 39
Ensuring that a software system meets its objectives requires continuous monitoring. In practice, monitoring is either insufficient to effectively detect and diagnose failures, or is too costly to use in production. An alternative is adaptive monitoring, where the system is monitored at a minimal level to determine system health, and if a problem is suspected, the monitoring level is automatically increased to determine faults. To model the system at different monitoring levels, we employ statistical techniques to identify stable relationships in the monitored data. These relationships characterise normal operation and can help detect anomalies. We describe our approach in the context of a J2EE-based system. We show that adaptive monitoring is a cost-effective alternative to continuous detailed monitoring. We inject 29 different faults, and show that we detect the faults in 80% of cases and shortlist the faulty component in 65% of the detected cases.

]]> 10.1504/IJHPCN.2011.038708 International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 29 - 39 Mohammad A. Munawar Paul A.S. Ward Department of Electrical and Computer Engineering, University of Waterloo, 200 University Avenue, Waterloo, Ontario, N2L 3G1, Canada. ' Department of Electrical and Computer Engineering, University of Waterloo, 200 University Avenue, Waterloo, Ontario, N2L 3G1, Canada self-managing systems adaptive monitoring statistical models fault diagnosis software monitoring modelling. 2011-02-23T23:20:50-05:00 7 1 29 39 2011-02-23T23:20:50-05:00 http://www.inderscience.com/link.php?id=38709 Self-stabilisation is a theoretical framework for fault-tolerance without external assistance. Adoption of self-stabilisation in distributed systems has received considerable research interest over the last decade. In this paper, we propose a self-stabilising algorithm for 3-edge-connectivity of an asynchronous distributed model of computation. A self-stabilising depth-first search algorithm is run concurrently to build a depth-first search spanning tree of the system. Once such a tree is constructed, all the 3-edge-connected components of the system can be detected in O(h) rounds, where h is the height of the depth-first search tree. The result of computation is kept in a distributed fashion in the sense that, upon stabilisation of the algorithm, each processor knows all other processors that are 3-edge-connected to it. The space complexity of our algorithm is O(n² log Δ) bits per processor, where Δ is an upper bound on the degree of a processor. A self-stabilising algorithm for 3-edge-connectivity
Abusayeed M. Saifullah, Yung H. Tsin
International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 40 - 52
Self-stabilisation is a theoretical framework for fault-tolerance without external assistance. Adoption of self-stabilisation in distributed systems has received considerable research interest over the last decade. In this paper, we propose a self-stabilising algorithm for 3-edge-connectivity of an asynchronous distributed model of computation. A self-stabilising depth-first search algorithm is run concurrently to build a depth-first search spanning tree of the system. Once such a tree is constructed, all the 3-edge-connected components of the system can be detected in O(h) rounds, where h is the height of the depth-first search tree. The result of computation is kept in a distributed fashion in the sense that, upon stabilisation of the algorithm, each processor knows all other processors that are 3-edge-connected to it. The space complexity of our algorithm is O(n² log Δ) bits per processor, where Δ is an upper bound on the degree of a processor.

]]> 10.1504/IJHPCN.2011.038709 International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 40 - 52 Abusayeed M. Saifullah Yung H. Tsin Computer and Inf. Science and Engineering, University of Florida, E301 CSE Building, P.O. Box 116120, Gainesville, FL 32611, USA; Computer Science and Engineering Department, Washington University, 1 Brookings Drive, Campus Box 1045, St. Louis St Louis, MO 63130, USA. ' School of Computer Science, University of Windsor, 401 Sunset Avenue, Windsor, Ontario N9C 3K7, Canada distributed systems transient faults fault tolerance self-stabilisation legitimate states illegitimate states depth-first search tree cut-pairs 3-edge-connected components time complexity 3-edge connectivity. 2011-02-23T23:20:50-05:00 7 1 40 52 2011-02-23T23:20:50-05:00 http://www.inderscience.com/link.php?id=38710 In this paper, we first apply the hierarchical concept to the care-of prefix (CoP) scheme as HCoP and enhance HCoP with a novel binding update tree (BUT) structure as HCoP-B for network mobility (NEMO) management of the nested mobile network. Second, we further extend HCoP-B to support the seamless handoff of the nested NEMO. As compared to schemes such as reverse routing header (RRH), route optimisation using tree information option (ROTIO) and HCoP with numerical performance evaluations, HCoP-B achieves the shortest handoff latency and significantly reduces the consumed network bandwidth of global binding update messages for route optimisations (RO) of all correspondent nodes (CN) after the nested mobile network hands over to a new AR. Besides, HCoP-B also achieves shorter playback disruption time and buffering time than ROTIO does, which is the only one scheme mentioned how to achieve seamless handoff for the NEMO in the literature, for ongoing real-time multimedia applications whenever the mobile subnet in the old nested mobile network hands over to a new one. On design of a route-optimised and seamless HCoP-B scheme for nested mobile networks
Ing-Chau Chang, Chia-Hao Chou, Lin-Huang Chang
International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 53 - 62
In this paper, we first apply the hierarchical concept to the care-of prefix (CoP) scheme as HCoP and enhance HCoP with a novel binding update tree (BUT) structure as HCoP-B for network mobility (NEMO) management of the nested mobile network. Second, we further extend HCoP-B to support the seamless handoff of the nested NEMO. As compared to schemes such as reverse routing header (RRH), route optimisation using tree information option (ROTIO) and HCoP with numerical performance evaluations, HCoP-B achieves the shortest handoff latency and significantly reduces the consumed network bandwidth of global binding update messages for route optimisations (RO) of all correspondent nodes (CN) after the nested mobile network hands over to a new AR. Besides, HCoP-B also achieves shorter playback disruption time and buffering time than ROTIO does, which is the only one scheme mentioned how to achieve seamless handoff for the NEMO in the literature, for ongoing real-time multimedia applications whenever the mobile subnet in the old nested mobile network hands over to a new one.

]]> 10.1504/IJHPCN.2011.038710 International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 53 - 62 Ing-Chau Chang Chia-Hao Chou Lin-Huang Chang Department of Computer Science and Information Engineering, National Changhua University of Education, No. 1, Jin De Road, Paisha Village, Changhua 500, Taiwan. ' Department of Computer Science and Information Engineering, National Changhua University of Education, No. 1, Jin De Road, Paisha Village, Changhua 500, Taiwan. ' Department of Computer and Information Science, National Taichung University, 140 Min-Shen Road, Taichung 40306, Taiwan care-of prefix CoP hierarchical CoP HCoP-B NEMO binding update tree BUT seamless handoff high performance networking routing optimisation nested mobile networks network mobility mobility management playback disruption time buffering time multimedia applications. 2011-02-23T23:20:50-05:00 7 1 53 62 2011-02-23T23:20:50-05:00 http://www.inderscience.com/link.php?id=38711 Middlewares are software infrastructures used to cluster heterogeneous and geographically distributed computational resources so as to exploit them as computing systems able to run large-scale applications. Although their main aim is to transform the internet into a sort of computational grid to which everyone can connect in order to execute distributed applications, a number of problems have still to be solved in order to make middlewares actually effective in such a context. For instance, exploiting computational resources available within 'departmental' organisations can be still considered a difficult task, since such resources are usually represented by computing nodes which belong to non-routable, private networks and are connected to the internet through publicly addressable IP front-end nodes. This paper presents a Java middleware that can support the execution of large-scale applications over heterogeneous multidomain, non-routable networks. In fact, the middleware can be also exploited to relieve programmers of the classic burden tied to the deployment of PVM run-time libraries and program executables among computational resources belonging to distinct network administrative domains. Parallel and distributed computing on multidomain non-routable networks
Franco Frattolillo, Federica Landolfi
International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 63 - 73
Middlewares are software infrastructures used to cluster heterogeneous and geographically distributed computational resources so as to exploit them as computing systems able to run large-scale applications. Although their main aim is to transform the internet into a sort of computational grid to which everyone can connect in order to execute distributed applications, a number of problems have still to be solved in order to make middlewares actually effective in such a context. For instance, exploiting computational resources available within 'departmental' organisations can be still considered a difficult task, since such resources are usually represented by computing nodes which belong to non-routable, private networks and are connected to the internet through publicly addressable IP front-end nodes. This paper presents a Java middleware that can support the execution of large-scale applications over heterogeneous multidomain, non-routable networks. In fact, the middleware can be also exploited to relieve programmers of the classic burden tied to the deployment of PVM run-time libraries and program executables among computational resources belonging to distinct network administrative domains.

]]> 10.1504/IJHPCN.2011.038711 International Journal of High Performance Computing and Networking, Vol. 7, No. 1 (2011) pp. 63 - 73 Franco Frattolillo Federica Landolfi Department of Engineering, University of Sannio, Corso Garibaldi 107, 82100 Benevento, Italy. ' Department of Engineering, University of Sannio, Corso Garibaldi 107, 82100 Benevento, Italy Java middleware cluster grids cluster computing parallel computing distributed computing multidomain networks non-routable networks. 2011-02-23T23:20:50-05:00 7 1 63 73 2011-02-23T23:20:50-05:00