Execution coordination in mobile agent-based distributed job workflow execution

Mobile agent-based distributed job workflow execution requires the use of execution coordination techniques to ensure that an agent executing a subjob can locate its predecessors’ execution results. This paper describes the classification, implementation, and evaluation of execution coordination techniques in the mobile agent-based distributed job workflow execution system. First, a classification of the existing execution coordination techniques is developed for mobile agent systems. Second, to put the discussion into perspective, our framework for mobile agent-based distributed job workflow execution over the Grid (that is, MCCF: Mobile Code Collaboration Framework) is described. How the existing coordination techniques can be applied in the MCCF is also discussed. Finally, a performance study has been conducted to evaluate three coordination techniques using real and simulated job workflows. The results are presented and discussed in the paper.     

基于移动代理的工作流原型研究与设计

快速发展的移动代理和网格技术给工作流提供了新的集成各种分布式资源的计算环境。传统的工作流架构有很多弱点,比如在动态环境中缺少灵活性等。提出了一个集成Web Services,网格,工作流和移动代理技术的分布式工作流架构,可以通过移动代理间的通讯来达到工作流之间协同工作,而且这个架构增强了工作流的灵活性和可靠性,以这个新的模型为基础实现了一个具体的工作流工作站,且可实现系统的安全使用。     

Automatic distributed workflow generation with GridMD library

GridMD is a C++ class library intended for constructing simulation applications and running them in distributed environments. The library abstracts away from details of distributed environments, so that almost no knowledge of distributed computing is required from a physicist working with the library. She or he just uses GridMD function calls inside the application C++ code to perform parameter sweeps or other tasks that can be distributed at run-time. In this paper we briefly review the GridMD architecture. We also describe the job manager component which submits jobs to a remote system. The C++ source code of our PBS job manager may be used as a standalone tool and it is freely available as well as the full library source code. As illustrative examples we use simple expression evaluation codes and the real application of Coulomb cluster explosion simulation by Molecular Dynamics.     

Network-aware support for mobile distributed teams

An experiment evaluated network-aware support to increase understanding of the factors that are important for successful teamwork in mobile geographically dispersed teams of first responders. Participants performed a simulated search and rescue team task and were equipped with a digitized map and real-time situation updates on the location of other participants in a simulated disaster area. The connection to a server, however, was made deliberately error-prone, leading to occasional losses of network connections. Consequently, participants were not provided with real-time situation updates. To deal with this problem we equipped team members with a network-aware application that signaled network loss to them and adapted the graphical representation of the location of fellow team members accordingly to the quality of location information present. The experiment revealed that presenting complete and reliable geospatial information improves teamwork. Teams connected to a server over a fast and reliable link showed superior performance over teams with no network connection whatsoever to a server. The present study failed, however, to demonstrate the added value of network-aware support when teams had to collaborate in the presence of an unreliable communications infrastructure. Although participants indicated a slight preference for the network-aware application over a condition without support signaling network loss, no differences were observed in team process and outcome measures.     

A timed mobile agent planning approach for distributed information retrieval in dynamic network environments

The number of mobile agents and total execution time are two factors used to represent the system overhead that must be considered as part of mobile agent planning (MAP) for distributed information retrieval. In addition to these two factors, the time constraints at the nodes of an information repository must also be taken into account when attempting to improve the quality of information retrieval. In previous studies, MAP approaches could not consider dynamic network conditions, e.g., variable network bandwidth and disconnection, such as are found in peer-to-peer (P2P) computing. For better performance, mobile agents that are more sensitive to network conditions must be used. In this paper, we propose a new MAP approach that we have named Timed Mobile Agent Planning (Tmap). The proposed approach minimizes the number of mobile agents and total execution time while keeping the turnaround time to a minimum, even if some nodes have a time constraint. It also considers dynamic network conditions to reflect the dynamic network condition more accurately. Moreover, we incorporate a security and fault-tolerance mechanism into the planning approach to better adapt it to real network environments.     

Deploying personalized mobile services in an agent-based environment

The mobile Internet allows users to obtain digitized contents and services from wired and wireless networks virtually anywhere at any time via different handheld mobile devices. However, due to the distinct features of mobile users, mobile devices and wireless networks, deploying mobile services is not as straightforward as generally expected. To ensure the success of mobile services, this paper presents a multi-agent framework that considers different contexts to support personalized services on wireless networks. In the proposed approach, client users, content providers, and service providers are all considered as software agents. They interoperate on the same platform to request and deliver mobile services. The most important issues related to agent operations and context awareness in an agent world are also discussed and analyzed. To verify our framework, different application services are developed accordingly on a publicly available middleware platform. Experiments are conducted for both services to evaluate their corresponding performance. The preliminary results show that our multi-agent approach to personalization is promising and efficient in the deployment of mobile services.     

Algorithms for validating E-tickets in mobile computing environment

E-tickets issued online can be used by different vendors to provide services such as discount coupons for E-shopping, or an entrance for a football game. One major issue associated with E-tickets is the problem of validation, which becomes more severe in a mobile environment due to mobility of a mobile host (MH) and frequent failure/disconnection of validation servers known as mobile support stations (MSSs). Some additional problems include the identification of duplicate submissions of an E-ticket by a MH. Thus, this paper proposes two protocols using the Flat and Hierarchical schemes for E-ticket processing and validation in a mobile environment that guarantee at least and at most once property (each E-ticket is validated at least and at most once). The protocols have been validated and compared through complexity analysis and experiments.     

Multi-objective list scheduling of workflow applications in distributed computing infrastructures

Executing large-scale applications in distributed computing infrastructures (DCI), for example modern Cloud environments, involves optimization of several conflicting objectives such as makespan, reliability, energy, or economic cost. Despite this trend, scheduling in heterogeneous DCIs has been traditionally approached as a single or bi-criteria optimization problem. In this paper, we propose a generic multi-objective optimization framework supported by a list scheduling heuristic for scientific workflows in heterogeneous DCIs. The algorithm approximates the optimal solution by considering user-specified constraints on objectives in a dual strategy: maximizing the distance to the user’s constraints for dominant solutions and minimizing it otherwise. We instantiate the framework and algorithm for a four-objective case study comprising makespan, economic cost, energy consumption, and reliability as optimization goals. We implemented our method as part of the ASKALON environment (Fahringer et al., 2007) for Grid and Cloud computing and demonstrate through extensive real and synthetic simulation experiments that our algorithm outperforms related bi-criteria heuristics while meeting the user constraints most of the time.     

Improvement of group performance of job distributed mobile robots by an emotionally biased control system

This paper deals with the implementation of emotions in mobile robots performing a specified task in a group in order to develop intelligent behavior and easier forms of communication. The overall group performance depends on the individual performance, group communication, and the synchronization of cooperation. With their emotional capability, each robot can distinguish the changed environment, can understand a colleague robot’s state, and can adapt and react with a changed world. The adaptive behavior of a robot is derived from the dominating emotion in an intelligent manner. In our control architecture, emotion plays a role to select the control precedence among alternatives such as behavior modes, cooperation plans, and goals. Emotional interaction happens among the robots, and a robot is biased by the emotional state of a colleague robot in performing a task. Here, emotional control is used for a better understanding of the colleague’s internal state, for faster communication, and for better performance eliminating dead time. This work was presented in part at the 12th International Symposium on Artificial Life and Robotics, Oita, Japan, January 25–27, 2007     

SHadoop: Improving MapReduce performance by optimizing job execution mechanism in Hadoop clusters

As a widely-used parallel computing framework for big data processing today, the Hadoop MapReduce framework puts more emphasis on high-throughput of data than on low-latency of job execution. However, today more and more big data applications developed with MapReduce require quick response time. As a result, improving the performance of MapReduce jobs, especially for short jobs, is of great significance in practice and has attracted more and more attentions from both academia and industry. A lot of efforts have been made to improve the performance of Hadoop from job scheduling or job parameter optimization level. In this paper, we explore an approach to improve the performance of the Hadoop MapReduce framework by optimizing the job and task execution mechanism. First of all, by analyzing the job and task execution mechanism in MapReduce framework we reveal two critical limitations to job execution performance. Then we propose two major optimizations to the MapReduce job and task execution mechanisms: first, we optimize the setup and cleanup tasks of a MapReduce job to reduce the time cost during the initialization and termination stages of the job; second, instead of adopting the loose heartbeat-based communication mechanism to transmit all messages between the JobTracker and TaskTrackers, we introduce an instant messaging communication mechanism for accelerating performance-sensitive task scheduling and execution. Finally, we implement SHadoop, an optimized and fully compatible version of Hadoop that aims at shortening the execution time cost of MapReduce jobs, especially for short jobs. Experimental results show that compared to the standard Hadoop, SHadoop can achieve stable performance improvement by around 25% on average for comprehensive benchmarks without losing scalability and speedup. Our optimization work has passed a production-level test in Intel and has been integrated into the Intel Distributed Hadoop (IDH). To the best of our knowledge, this work is the first effort that explores on optimizing the execution mechanism inside map/reduce tasks of a job. The advantage is that it can complement job scheduling optimizations to further improve the job execution performance.     

Precise execution offloading for applications with dynamic behavior in mobile cloud computing

In order to accommodate the high demand for performance in smartphones, mobile cloud computing techniques, which aim to enhance a smartphone’s performance through utilizing powerful cloud servers, were suggested. Among such techniques, execution offloading, which migrates a thread between a mobile device and a server, is often employed. In such execution offloading techniques, it is typical to dynamically decide what code part is to be offloaded through decision making algorithms. In order to achieve optimal offloading performance, however, the gain and cost of offloading must be predicted accurately for such algorithms. Previous works did not try hard to do this because it is usually expensive to make an accurate prediction. Thus in this paper, we introduce novel techniques to automatically generate accurate and efficient method-wise performance predictors for mobile applications and empirically show they enhance the performance of offloading.     

Analysis of mobile agents in network fault management

Network domains have become more and more advanced in terms of their size, complexity and the level of heterogeneity. Comprehensive fault management is the most significant challenge in network management. Fault management can help increase the availability of the network by quickly identifying the faults and then, proactively, start the recovery process. Current centralized configured network management systems suffer from problems such as insufficient scalability, availability and flexibility as networks become more distributed. Mobile agents (MAs), with integral intelligence, can present a reasonable new technology that will help to achieve distributed management, several researchers have embraced these approaches. In this paper, we introduce a general analytical model for network management client/server (CS) and MA paradigms. We express how to build up an analytical framework, which can be used to quantitatively assess the performances of the MA and CS paradigms under different scenarios. We present some numerical and experimental results that demonstrate the applicability of our proposed framework, which will be based on a combination of MA and CS schemes called Adaptive Intelligent Mobile Agent.     

Appraising fairness in languages for distributed programming

The relations among various languages and models for distributed computation and various possible definitions of fairness are considered. Natural semantic criteria are presented which an acceptable notion of fairness should satisfy. These are then used to demonstrate differences among the basic models, the added power of the fairness notion, and the sensitivity of the fairness notion to irrelevant semantic interleavings of independent operations. These results are used to show that from the considerable variety of commonly used possibilities, only strong process fairness is appropriate forCSP if these criteria are adopted. We also show that under these criteria, none of the commonly used notions of fairness are fully aceptable for a model with an n-way synchronization mechanism. The notion of fairness most often mentioned for Ada is shown to be fully acceptable. For a model with nonblockingsend operations, some variants of common fairness definitions are appraised, and two are shown to satisfy the suggested criteria. Krzysztof R. Apt was born in 1949 in Poland. Received his Ph.D. in 1974 from Polish Academy of Sciences in Warsaw in mathematical logic. From 1974 until 1981 worked at various scientific institutions in the Netherlands and from 1981 until 1987 at C.N.R.S. in Paris, France. Spent 1985 as a visiting scientist at IBM Research Centre in Yorktown Heights, U.S.A. Currently holding an Endowed Professorship at the Department of Computer Sciences at the University of Texas at Austin; also a senior research scientist at the Centre for Mathematics and Computer Science in Amsterdam, the Netherlands. His research interests include program correctness and semantics, methodology of distributed computing, use of logic as a programming language and non-standard forms of reasoning. He has served on editorial boards of a number of journals and program committees of numerous conferences in computer science. Lectured in a dozen countries on four continents. Also, he has run two marathons and crossed Sumatra on a bicycle. Shmuel Katz received his B.A. in Mathematics and English Literature from U.C.L.A., and his M.Sc. and Ph.D. in Computer Science (1976) from the Weizmann Institute in Rehovot, Israel. From 1976 to 1981 he was a researcher at the IBM Israel Scientific Center. Presently, he is a Senior Lecturer in the Computer Science Department at the Technion in Haifa, Israel. In 1977–78, he visited for a year at the University of California, Berkeley, and in 1984–85 was at the University of Texas at Austin. He has also been a consultant for the MCC Software Technology Program. His research interests include the methodology of programming, specification methods, program verification and semantics, distributed programming, data structures, and programming languages. Nissim Francez received his B.A. in Mathematics and Philosophy from the Hebrew University in Jerusalem, and his M.Sc. and Ph.D. in computer science (1976) from the Weizmann Institute of Science, Rehovot, Israel. In 1976–77 he spent a postdoctoral year at Queen's university, Belfast, where he was introduced by C.A.R. Hoare to CSP. In 1977–78 he was an assistant professor at USC, Los Angeles. From 1978 he is with the Computer Science Department at the Technion. In 1982–83 he was on a sabbatical leave at IBM T.J. Watson Research Center. He has been a consultant for MCC's software technology program, working on multiparty activities in distributed systems. He had summer appointments in Harvard University, IBM T.J. Watson Research Center, Utrecht University, CWI (Amsterdam) and at MCC. He also served in several program committees. His research interests include program verification and the semantics of programming languages, mainly for concurrent and distributed programming. Is also interested in logic programming and recursive query evaluation and in compiler constration. He is the author of the first book onFairness. Unfortunately, he is incapable of Marathon running...     

Cost-effective approach inter-LMA domain management and distributed mobility control scheme in proxy mobile IPv6 networks

The Proxy Mobile IPv6 (PMIPv6) is designed to provide network-based mobility management support to a mobile node (MN) without any involvement of the MN in mobility related signaling; hence, the proxy mobility entity performs all related signaling on behalf of the MN. The new principal functional entities of PMIPv6 are the local mobility anchor (LMA) and the mobile access gateway (MAG). In PMIPv6, all the data traffic sent from the MN gets routed to the LMA through a tunnel between the LMA and the MAG, but it still has a single point of failure (SPOF) and a bottleneck state of traffic. To solve these problems, various approaches directed towards PMIPv6 performance improvements, such as route optimization, have been proposed. However, these approaches add additional signaling to support the MN׳s mobility, which incurs extra network overhead and still provides difficulty when applied to multiple-LMA networks. In this paper, we propose an improved route optimization in PMIPv6-based multiple-LMA networks. All LMAs were connected to the proxy internetworking gateway (PIG), which performs inter-domain distributed mobility control. Each MAG keeps the information of all LMAs in the PMIPv6 domain, so it is possible to perform fast route optimization. Therefore, this method supported route optimization without any additional signaling because the LMA receives the state information of route optimization from the PIG.     

Lower bounds for the broadcast problem in mobile radio networks

Summary.  In this paper, we prove a lower bound on the number of rounds required by a deterministic distributed protocol for broadcasting a message in radio networks whose processors do not know the identities of their neighbors. Such an assumption captures the main characteristic of mobile and wireless environments [3], i.e., the instability of the network topology. For any distributed broadcast protocol Π, for any n and for any D≦n/2, we exhibit a network G with n nodes and diameter D such that the number of rounds needed by Π for broadcasting a message in G is Ω(D log n). The result still holds even if the processors in the network use a different program and know n and D. We also consider the version of the broadcast problem in which an arbitrary number of processors issue at the same time an identical message that has to be delivered to the other processors. In such a case we prove that, even assuming that the processors know the network topology, Ω(n) rounds are required for solving the problem on a complete network (D=1) with n processors. Received: August 1994 / Accepted: August 1996     

Mutual tests using immunity-based diagnostic mobile agents in distributed intrusion detection systems

Distributed intrusion detection systems have several advantages over centralized systems, such as scalability, adaptability, and fault tolerance. A current research topic in distributed systems is self-monitoring to identify corrupted intrusion detection systems. One way of self-monitoring is for intrusion detection systems to check each other. As we describe herein, this can be done by mobile agents using an immunity-based diagnostic method modeled on idiotypic network theory. In simulations, the credibility of normal intrusion detection systems remained near 1, while it fell to about 0 for corrupted intrusion detection systems, thus enabling identification of the latter. We also confirmed what effects some parameters have on the diagnostic capability.This work was presented, in part, at the 8th International Symposium on Artificial Life and Robotics, Oita, Japan, January 24–26, 2003     

Distributed reformation of core-based group-shared multicast trees in mobile ad hoc networks

This paper proposes a method to reduce the cost of a core-based group-shared multicast tree, where the cost is evaluated by the total bandwidth consumption of multicasting packets among all group members. Due to the broadcast nature of radio transmissions, we find that the challenge of determining minimum cost multicast tree can be approximated by finding the multicast tree with a minimum number of non-leaves (the minimum non-leaf multicast tree problem). However, we also find that the minimum non-leaf multicast tree problem is NP-complete. Thus, a method is proposed to dynamically reduce the number of non-leaves in an existing multicast tree. Experimental results show that our method reduces the cost of the multicast tree in both geometrically and randomly distributed network models and the random waypoint mobility model.     

A lightweight and distributed middleware to provide presence awareness in mobile ubiquitous systems

Several researchers have identified the need to count on presence awareness in ubiquitous systems that support mobile activities, particularly when these systems are used to perform loosely-coupled mobile work. In such a work style, mobile users conduct face-to-face on-demand interactions, therefore counting on awareness information about the position and availability of potential collaborators becomes mandatory for these applications. Most proposed solutions that provide user presence awareness involve centralized components, have reusability limitations, or simply address a part of that service. This article presents a lightweight and fully distributed middleware named Moware, which allows developers to embed presence awareness services in mobile ubiquitous systems in a simple way. The article also describes the Moware architecture, its main components and strategies used to deal with several aspects of the presence awareness support. These design strategies can be reused by software designers to provide presence awareness capabilities into middleware and specific software applications. Moware services were embedded in a mobile ubiquitous system that supports inspectors during the construction inspection process. The preliminary results indicate that the middleware was easy to use for developers, and its services were useful for the end-users.     

基于移动代理的网格作业管理模型的研究

网格作业管理是网格的最主要的功能之一,主要涉及到网格计算环境下作业调度和分配、作业分解和迁移、负载平衡等问题.介绍了一种引用移动代理Agent技术的网格作业管理模型,并针对该系统依次详细地描述了其中的作业提交与发布、作业调度以及作业的分解和迁移,最后给出基于移动代理的网格作业管理模型的原型系统,并对该系统作单机与联机的性能进行分析和比较.