Execution Nodes Research Articles

Service-Centric Networking for Distributed Heterogeneous Clouds

Optimal placement and selection of service instances in a distributed heterogeneous cloud is a complex trade-off between application requirements and resource capabilities that requires detailed information on the service, infrastructure constraints, and the underlying IP network. In this article we first posit that from an analysis of a snapshot of today's centralized and regional data center infrastructure, there is a sufficient number of candidate sites for deploying many services while meeting latency and bandwidth constraints. We then provide quantitative arguments why both network and hardware performance needs to be taken into account when selecting candidate sites to deploy a given service. Finally, we propose a novel architectural solution for service-centric networking. The resulting system exploits the availability of fine-grained execution nodes across the Internet and uses knowledge of available computational and network resources for deploying, replicating and selecting instances to optimize quality of experience for a wide range of services.

An Enhanced Model for Stochastic Coordination

Applications developed over the cloud coordinate several, often anonymous, computational resources, distributed over different execution nodes, within flexible architectures. Coordination models able to represent quantitative data provide a powerful basis for their analysis and validation. This paper extends IMCreo, a semantic model for Stochastic reo based on interactive Markov chains, to enhance its scalability, by regarding each channel and node, as well as interface components, as independent stochastic processes that may (or may not) synchronise with the rest of the coordination circuit.

Hybrid Topology Control Based on Clock Synchronization in Wireless Sensor Network

Objectives: The main objective of this work is to lessen energy utilization of Wireless Sensor Network. The Radio Frequency Identification is the protocol of Wireless Sensor Network which provides channel access using TDMA protocol. Due to weak clock synchronization of sensor nodes TDMA protocols do not work efficiently. The improvement will be proposed in Radio Frequency Identification protocol to synchronize clocks of sensor nodes for efficient working of TDMA protocol. Methods: To reduce energy consumption of sensor networks various operation modes are used and these modes are sleep, active and ready mode. The TDMA protocol is used to assign time to sensor nodes for efficient execution of these modes. In the sensor network clocks of the sensor nodes are weakly synchronized due to which packet loss may occur due to which network lifetime get reduced. The proposed improvement and existing Radio Frequency Identification protocol is simulated in NS2 with finite number of sensor nodes on area of 800*800 meter. The LEACH protocol is used for clustering and for Cluster Head selection. Findings: In this paper, improvement is being proposed in Radio Frequency Identification protocol for clock synchronization and proposed improvement in based on time lay technique. In time lay technique base station takes initiative to synchronize clocks of the sensor nodes. It is been analyzed that proposed algorithm performs well in terms of energy consumption, throughput and delay.

Distributed Late-binding Scheduling and Cooperative Data Caching

Pull-based overlays are used in some of today’s largest computational grids. Job agents are submitted to resources with the duty of retrieving real workload from a central queue at runtime and executing it. This model helps overcome the problems of direct job submission in the highly complex grid environments, namely, heterogeneity, imprecise status information, relatively high failure rates and slow adaptation to changes of grid conditions or user priorities. This article presents a distributed scheduling architecture for such late-binding overlays. In this architecture, execution nodes share a distributed hash table and cooperatively perform job assignment. As our experiments prove, scalability problems of centralized matching are avoided, achieving low and predictable scheduling overheads even for execution of large workflows, and total turnaround times are improved. This is in line with the predictions of a theoretical model of grid workflow execution that the article also discusses. Scalability makes fine-grained scheduling possible and enables new functionalities, like a distributed data cache shared by the execution nodes, which helps alleviate the commonly congested storage services. In addition, we show that our system is more resilient to problems like communication breakdowns between computation centres. Moreover, the new architecture is better prepared to deal with demanding scenarios like intense demand of popular data files or remote data processing.

Characterization of grid computing resources using measurement-based evaluation

An important factor that needs to be considered by every Grid application end-user and systems (such as schedulers or mediators), during Grid resource selection and mapping to applications, is the performance capacity of hardware resources attached to the Grid, and made available through its Virtual Organizations. In this paper, we represent the performance of a computational Grid as a regression model that can be used to fine-tune the selection of suitable Grid resources. A study on the performance of distributed systems with respect to particular variations in parameters is presented. Our objective is to use a measurement-based evaluation technique to characterize the specific performance contribution of the individual Grid resource configurations. In the process, we identify the key primary parameters (or factors) that should be considered when selecting and allocating a computational node for user application execution.

Safe and efficient runtime testing framework applied in dynamic and distributed systems

This paper provides a standard-based and resource aware Runtime Testing Framework For Adaptable and Distributed Systems (RTF4ADS). Based on the runtime testing approach, RTF4ADS performs safely and efficiently tests on the final operational environment of such dynamic systems. Indeed, our proposal (1) looks for a minimal set of tests to re-execute written in a standardized notation, (2) assigns the involved test components in execution nodes while respecting resources and connectivity constraints and finally (3) performs distributed testing at runtime while it prevents test processes from interfering with business processes. Implementation details of the proposed research prototype are presented. To demonstrate the validity and the usefulness of RTF4ADS, a case study in the healthcare domain implemented using the Open Services Gateway Initiative (OSGi) platform is illustrated. The experiments that we conducted show a reasonable overhead introduced by RTF4ADS. They also demonstrate the efficiency of this framework in testing dynamic and distributed systems while reducing side effects of this validation technique.

A Constraint-based Decentralized Task Routing Approach to Large-Scale Scheduling in Cloud Environment

This article mainly focused on cloud scheduling with constraint based decentralized task routing. The job of scheduling tasks across various nodes in a hierarchical network scenario is an exigent problem. The concept of decentralized distribution scheme proposed in base paper is time-consuming since it has to compute the availability function for each and every node. In this paper we proposed a CBDA (Constraint Based Decentralized Algorithm) which offers the expediency of being quick and “Make-span minimization policy” is implemented to reduce the completion time of the currently executing nodes. In our presumption, the submission nodes are semi centralized and it can store the availability information of the nodes or routers within its area. This paper considers the allotment of the tasks to the execution nodes which are unoccupied by other tasks. The dynamic allotment of the tasks to the nodes in the tree based approach is the major criteria for selecting the desired node. This paper proposes a trade-off between fully centralized model and the decentralized model by implementing a new constraint based decentralization scheme which saves time consumption and enhances efficiency of task scheduling.

Parallel Fuzzy Inference System (PFIS) based Medical Image Denoising

Objective: To remove the noises occurred in the bio medical pictures with less computation time. Methods: An innovative technique called Parallel Fuzzy Inference System (PFIS) is introduced for image denoising in the medical images. This method takes input training images, process a dissimilar neighbourhood association between the center pixel and generates the fuzzy inference rules. These rules are distributed to the nodes for simultaneous execution. Then, the type-1 interval fuzzy set is submitted to the resultant defuzzifier module which will decode it into scalar value. If it is impulse pixel, the noise filter is used to reduce the noise. Results: The improvement of the PFIS technique is evaluated by using the medical images. PFIS method shows high efficiency when matched with the existing Type-2 Fuzzy Logic (TFL) based impulse detector for impulse noise removal. In the PFIS method, the fuzzy rules are generated analyzing the medical images and these rules are processed simultaneously. If the noise density is 75%, the Mean Squared Error (MSE) in PFIS is 59.48, the False Classification Ratio (FCR) is 3.25 and the computation time is 740 ms. According to the comparison and the results from the experiment shows that the proposed method has high efficiency. Conclusion: The findings demonstrate that the PFIS is presented and this method has high efficiency in terms of MSE, FCR and computation time.

WSN-based Control System of Co2Concentration in Greenhouse

Supplying proper CO2 in greenhouses is an important solution to improve yields and quality of crops. WSN has been widely used for production management in greenhouse. In this paper, a WSN system was developed for CO2 concentration control in greenhouses. The whole system consisted of three parts, including sensor and execution nodes, gateway node and remote management platform. The sensor and execution node could collect environmental information, and control the switch of CO2 air source by connecting to a solenoid valve. The monitored environmental parameters included air temperature and humidity, light intensity and CO2 concentration. The gateway node, as a bridge between sensor nodes and remote management platform, could organize a ZigBee network with sensor and execution nodes, and also keep communication with internet via general packet radio service (GPRS) module. The remote management platform was used to predict the photosynthetic rate and calculate the optimal CO2 concentration based on real time...

Runtime testing approach of structural adaptations for dynamic and distributed systems

Adaptable and distributed component-based systems need to be checked and validated in order to ensure their correctness and trustworthiness when structural dynamic changes occur. Runtime testing, seen as an online validation technique, is used in this work for this aim. Indeed, we propose a generic and resource aware test platform that: 1 looks for a minimal set of tests to re-execute written in the TTCN-3 standard notation; 2 assigns the involved test components in execution nodes while respecting resources and connectivity constraints and finally; 3 performs runtime tests with reducing interference risks by supplying different test isolation strategies depending on testability degrees of each component under test. To show the relevance of our solution, a prototype is proposed and illustrated via a case study in the healthcare domain implemented using the OSGi platform.

Optimal service distribution in WSN service system subject to data security constraints.

Services composition technology provides a flexible approach to building Wireless Sensor Network (WSN) Service Applications (WSA) in a service oriented tasking system for WSN. Maintaining the data security of WSA is one of the most important goals in sensor network research. In this paper, we consider a WSN service oriented tasking system in which the WSN Services Broker (WSB), as the resource management center, can map the service request from user into a set of atom-services (AS) and send them to some independent sensor nodes (SN) for parallel execution. The distribution of ASs among these SNs affects the data security as well as the reliability and performance of WSA because these SNs can be of different and independent specifications. By the optimal service partition into the ASs and their distribution among SNs, the WSB can provide the maximum possible service reliability and/or expected performance subject to data security constraints. This paper proposes an algorithm of optimal service partition and distribution based on the universal generating function (UGF) and the genetic algorithm (GA) approach. The experimental analysis is presented to demonstrate the feasibility of the suggested algorithm.

An Executive Approach to Achieve Mutual Exclusion in Distributed Data using Topology and Association Rule.

In this paper, a new approach is being proposed to achieve mutual exclusion in distributed system using computer network and topology of nth nodes. In this executive approach nodes communicate among themselves using message passing technique. In this executive approach, distributed system with n nodes is logically partitioned into number of sub distributed system having only m½ nodes, where m is obtained by adding a minimum number in n to make it next perfect square number only if n is not a perfect square. Proposed algorithm is a Token based approach and achieves token optimally in 2 messages only for the best case and in worst case a node achieves token in n messages only.

CMS users data management service integration and first experiences with its NoSQL data storage

The distributed data analysis workflow in CMS assumes that jobs run in a different location to where their results are finally stored. Typically the user outputs must be transferred from one site to another by a dedicated CMS service, AsyncStageOut. This new service is originally developed to address the inefficiency in using the CMS computing resources when transferring the analysis job outputs, synchronously, once they are produced in the job execution node to the remote site.The AsyncStageOut is designed as a thin application relying only on the NoSQL database (CouchDB) as input and data storage. It has progressed from a limited prototype to a highly adaptable service which manages and monitors the whole user files steps, namely file transfer and publication.The AsyncStageOut is integrated with the Common CMS/Atlas Analysis Framework. It foresees the management of nearly nearly 200k users' files per day of close to 1000 individual users per month with minimal delays, and providing a real time monitoring and reports to users and service operators, while being highly available. The associated data volume represents a new set of challenges in the areas of database scalability and service performance and efficiency. In this paper, we present an overview of the AsyncStageOut model and the integration strategy with the Common Analysis Framework. The motivations for using the NoSQL technology are also presented, as well as data design and the techniques used for efficient indexing and monitoring of the data. We describe deployment model for the high availability and scalability of the service. We also discuss the hardware requirements and the results achieved as they were determined by testing with actual data and realistic loads during the commissioning and the initial production phase with the Common Analysis Framework.

TAN: A Distributed Algorithm for Dynamic Task Assignment in WSNs

We consider the scenario of wireless sensor networks where a given application has to be deployed and each application task has to be assigned to each node in the best possible way. Approaches where decisions on task execution are taken by a single central node can avoid the exchange of data packets between task execution nodes but cannot adapt to dynamic network conditions, and suffer from computational complexity. To address this issue, in this paper, we propose an adaptive and decentralized task allocation negotiation algorithm (TAN) for cluster network topologies. It is based on noncooperative game theory, where neighboring nodes engage in negotiations to maximize their own utility functions to agree on which of them should execute single application tasks. Performance is evaluated in a city scenario, where the urban streets are equipped with different sensors and the application target is the detection of the fastest way to reach a destination, and in random WSN scenarios. Comparisons are made with three other algorithms: 1) baseline setting with no task assignment to multiple nodes; 2) centralized task assignment lifetime optimization; and 3) a dynamic distributed algorithm, DLMA. The result is that TAN outperforms these algorithms in terms of application completion time and average energy consumption.

Maximizing availability for task scheduling in computational grid using genetic algorithm

SUMMARYComputational grid provides a wide distributed platform for high‐end compute intensive applications. Grid scheduling is often carried out to schedule the submitted jobs on the nodes of the grid so that some characteristic parameter is optimized. Availability of the computational nodes is one of the important characteristic parameters and measures the probability of the node availability for job execution. This paper addresses the availability of the grid computational nodes for the job execution and proposes a model to maximize it. As such, the task scheduling problem in grid is nondeterministic polynomial‐time hard, and often, metaheuristics techniques are applied to solve it. Genetic algorithm, a metaheuristic technique based on evolutionary computation, has been used to solve such complex optimization problem. This work proposes a technique for the grid scheduling problem using genetic algorithm with the objective to maximize availability. Simulation experiment, to evaluate the performance of the proposed algorithm, is conducted, and results reveal the effectiveness of the model. A comparative study has also been performed. Copyright © 2014 John Wiley & Sons, Ltd.

ACM HotMobile 2013 demo

A key concern in utilizing external resources for mobile application execution is a potential loss of data privacy. User data needs to be shipped to possibly untrusted remote nodes for execution. We introduce the concept of data partitioning, where in the user can conveniently identify the sensitive parts of the data, which is then prevented from being shipped to untrusted remote servers.

Implementation of seismic data inversion as grid-backed web service

The essence of probability approach in solving inverse problems is use of a posteriory probability density function (APDF). APDF bears a priory system signature and the way the corresponding model parameter set describes the observational data. APDF importance sampling is carried out independently in different points thus the problem is a good target for distributed computing. The software implementation make use of virtualization to lower the hardware cost: all Globus Toolkit services, except for GridFTP, run as virtual guests on execution nodes. Due to very insignificant resources utilization the guests make no footprint on node’s computation power. The execution nodes local resource manager is TORQUE. The main thought for web interface is to hide complex middleware interaction from user. Such interaction usually takes place during start-up and result collection and analysis procedures. The implemented minimalistic web service allows defining of models, estimation of calculation time for user-specified parameter set, task start-up, management, results acquisition and visualization. The specific pilot problem we implemented is determination of seismic anisotropy of mantle and upper crust by teleseismic data.

A New Approach for Dynamic Job Scheduling in a Volunteer Desktop Grid

Nodes in a volunteer desktop grid are based on the volunteer participation of desktop nodes owned by the individual users. The volunteer nodes contribute their idle resources for public execution during their free time and withdraw during their busy time due to high priority node owner’s private execution. This property makes desktop grid dynamic in nature. Job scheduling is mainly influenced by two factors: node dynamism and heterogeneity. The job generation rate at each node is different from other nodes in the desktop Grid and hence the load at each node changes with time. This situation leads to increased computational demands at some nodes than from others and makes the grid often to get into unbalanced environment. Job migration to remote nodes involves job transmission latency. Since grid is a dynamic environment, when the job reaches to the remote node for execution the node might become busy and the selected target node may not complete the execution of the job at the expected speed. Therefore, the selection of a target node for job migration plays an important role in improving the overall performance of the desktop Grid. In this paper we present a new approach for dynamic job scheduling that considers node dynamism and job transmission latency into account for making scheduling decisions. The algorithm is compared against the Resource Exclusion and non migration algorithms and the simulation results shows that the proposed algorithm has got considerable improvement over the other two.

A task routing approach to large-scale scheduling

Scheduling many tasks in environments of millions of unreliable nodes is a challenging problem. To our knowledge, no work in the literature has proposed a solution that also supports many policies with very different objectives. In this paper, we present a decentralized scheduling model that overcomes these problems. A hierarchical network overlay supports a scalable resource discovery and allocation scheme. It uses aggregated information to route tasks to the most suitable execution nodes, and is easily extensible to provide very different scheduling policies. For this paper, we implemented a policy that just allocates tasks to idle nodes, a policy that minimizes the global makespan and a policy that fulfills deadline requirements. With thorough simulation tests, we conclude that our model allocates any number of tasks to several million nodes in just a few seconds, with very low overhead and high resilience. Meanwhile, policies with different objectives implemented on our model perform almost as well as their centralized counterpart.

Selecting the most fitting resource for task execution

In the computing grid, task scheduling, in order to discover resources for user’s requirements, is important. In general, the task scheduler assigns tasks to a proper resource node for execution, and the resource nodes with better performance would be assigned first. When task loading is heavy and all resource nodes with better performance are assigned, other tasks have to be assigned to the resource nodes with inferior performance. Therefore, if a task is assigned to a resource node without considering the performance factor, the overall execution time will increase obviously. To solve this problem, a task scheduling algorithm that searches for the proper resource for task execution is proposed in this paper. Simulation results show that the proposed algorithm works well in dynamic environments.