We have proposed and implemented a distributed asynchronous Web-based training (WBT) system. In order to improve the scalability and robustness of this system, all exercises and functions, such as scores user's an...
详细信息
ISBN:
(纸本)9780889867048
We have proposed and implemented a distributed asynchronous Web-based training (WBT) system. In order to improve the scalability and robustness of this system, all exercises and functions, such as scores user's answers are realized on mobile agents. these agents are distributed to computers, and they can be constructed with a P2P network that modified Content-Addressable Network (CAN). In this paper, we present the exercise management scheme for the proposed WBT system. In a WBT system based on client/server model, management of exercises is simply achieved by manipulating data, since all contents are concentrated in one server computer. In the proposed system, however, we need to pay attention to distributed agents for management of exercises. In order to achieve operation of exercise management, i.e., adding, deleting and updating exercises on the distributed WBT system, we use steps with considering multi-agent based distributed environment: specifying an agent which provides an exercise and searching its node as pre-operation, and sending an agent to the expected node and notifying other cooperating agents as post-operation.
In this paper, we study a parallel job scheduling model which takes into account both computation time and the overhead from communication between processors. Assuming that a job Jj has a processing requirement pj and...
详细信息
ISBN:
(纸本)9780889867048
In this paper, we study a parallel job scheduling model which takes into account both computation time and the overhead from communication between processors. Assuming that a job Jj has a processing requirement pj and is assigned to kj processors for parallel execution, then the execution time will be modeled by tj = p j / kj+ (kj - 1) c, where c is the constant overhead cost associated with each processor other than the master processor. In this model, (kj - 1)c represents the cost for communication and coordination among the processors. this model attempts to accurately portray the actual execution time for jobs running in parallel on multiple processors. Using this model, we will study the online algorithm Earliest Completion Time (ECT) and show a lower bound for the competitive ratio of ECT for m ≥ 2 processors. For m ≤ 4, we show the matching upper bound to complete the competitive analysis for m = 2,3,4. For large m, we conjecture that the ratio approaches 30/13 ≈ 2.30769.
Both web services framework and peer-to-peer networks provide a discovery process, but as current P2P systems focus more on the discovery of content in the form of common files (possibly associated with metadata), a c...
详细信息
ISBN:
(纸本)9780889867048
Both web services framework and peer-to-peer networks provide a discovery process, but as current P2P systems focus more on the discovery of content in the form of common files (possibly associated with metadata), a centralized UDDI registry serves content in the form of metadata describing web services. thus, the intersection between P2P and Web services is clear, so that it recently gained lots of interest: there is the need of finding a way to exploit the high potential of web services without paying in terms of scalability or single-point-of-failures. Our aim in this paper is two-fold: on the one hand, we want to present a high level protocol for the addressing and the discovery of web services that is totally decentralized and based on a structured P2P topology. On the other hand, we introduce a way to solve the exact match key-based routing problem, that afflicts structured P2P systems. this model guarantees that all addressed web services will be discovered in a logarithmic number of hops, as it is based on a distributed Hash Table. One of the key innovation is the possibility of customizing the categorization of web services using tags, which are becoming very common in the so called Web 2.0.
Passive testing is a technique suitable for continuous, non-intrusive, autonomous testing of qualitative behavioural properties (correctness) of a deployed distributed system. A real passive tester has to face observa...
详细信息
ISBN:
(纸本)9780889866379
Passive testing is a technique suitable for continuous, non-intrusive, autonomous testing of qualitative behavioural properties (correctness) of a deployed distributed system. A real passive tester has to face observational uncertainty, which may lead to false verdicts. We submit the novel idea of a self-tuned passive tester, which is able to adapt to a priori unknown, and possibly changing delays in communication channels. We propose the structure and algorithms of a passive tester that "tunes itself" basing solely on its own, locally issued verdicts. this seemingly counter-intuitive principle is shown by simulation to be viable and effective.
Multiple-disk architectures are an attractive approach to meet high performance I/O demands in I/O intensive applications such as search engines, web servers and information retrieval systems. this requires that the i...
详细信息
ISBN:
(纸本)9780889867048
Multiple-disk architectures are an attractive approach to meet high performance I/O demands in I/O intensive applications such as search engines, web servers and information retrieval systems. this requires that the issues of dynamic load balancing and access parallelism be addressed, which is the goal of this paper. We address the problem of document declustering in a keyword-based information retrieval system for parallel architectures consisting of a single processor and multiple disks. We propose and evaluate experimentally four similarity-based methods, viz., set, multiset, vector, and euclidean, for declustering documents. Interestingly, our results show that for single keyword queries as well as boolean and queries the set and multiset methods generally outperform the vector and euclidean methods with set being the best for the so-called simple plan. We also introduce a highest-frequency first retrieval scenario and compare the methods under this scenario, and find that set and multiset methods are still generally superior to the other methods withthe multiset outperforming the set method. We compare these methods withthe (theoretically) optimal values, which are practically impossible to achieve. Finally, we approximated the multiset method using the harmonic mean and found that the results were slightly inferior than multiset method, but still better than the vector and euclidean methods.
Two of the main characteristics of computation grids are their heterogeneity and the sharing of resources between different users. this is the cost of the tremendous computing power offered by such platforms. Scheduli...
详细信息
ISBN:
(纸本)9780889867048
Two of the main characteristics of computation grids are their heterogeneity and the sharing of resources between different users. this is the cost of the tremendous computing power offered by such platforms. Scheduling several applications concurrently in such an environment is thus challenging. In this paper we propose a first step towards the scheduling of multiple parallel task graphs (PTG), a class of applications that can benefit of large and powerful platforms, by focusing on the allocation process. We consider the application of a resource constraint on the schedule and determine the number of processors allocated to the different tasks of a PTG while respecting that constraint. We present two different allocation procedures and validate them in simulation over a wide range of scenarios with regard to their respect of the resource constraint and their impact on the completion time of the scheduled applications. We find that our procedures provide a guarantee on the resource usage for a low cost in terms of execution time.
As data collection technologies advance, almost every field of scientific research sees an increase in the amount and frequency of data that can be collected. In many cases, the first best method for preprocessing two...
详细信息
ISBN:
(纸本)9780889867048
As data collection technologies advance, almost every field of scientific research sees an increase in the amount and frequency of data that can be collected. In many cases, the first best method for preprocessing two and three dimensional data sets is to examine them visually, using some sort of plotting software. Although display technology is advancing and monitors are becoming available with increasing resolutions, this increase in available resolution does not necessarily match the rapid increase in data collection. this mismatch leaves researchers with only three choices for visual processing: to view the data in many small chunks, to view only a representative subset of the data, or to create a custom display for viewing much larger data sets at higher resolutions. this paper discusses one method of fullfilling the third option. the process used here is to create a customized display using nine consumer level monitors mounted in a three by three configuration resulting in a display with nine times the resolution of a single monitor. A custom viewing environment is also developed within this project that is designed to lend itself better to the tiled display than would an existing three dimensional data visualization application. the data used for this project consists of a large set of Oceanographic sensor readings taken from several deployments throughout the Gulf of Maine.
Structured peer-to-peer overlay networks provide a means for interconnecting a large set of computing nodes called peers in a distributed, scalable and self-organizing manner. Typically in such systems, each peer main...
详细信息
ISBN:
(纸本)9780889867048
Structured peer-to-peer overlay networks provide a means for interconnecting a large set of computing nodes called peers in a distributed, scalable and self-organizing manner. Typically in such systems, each peer maintains contact information about a small subset of other peers called neighbors. these neighbors are often chosen to ensure that each lookup operation takes a logarithmic number of overlay hops under normal conditions. Relying solely on the number of overlay hops as the metric for the quality of the lookup operation is not sufficient both for the applications using the overlay network and, for the effective use of the underlying communication network which is typically the Internet. DKS [1] is a P2P middleware built upon the principle we called distributed K-ary search, which ensures that any lookup takes O(log k(N)) overlay hops, where N is the number of nodes in the system and K is the search arity. In a DKS overlay network, proximity is not considered. As a result, end-to-end latencies could be significantly large. In this paper we present EPDKS, a novel structured peer-to-peer overlay network that is based on the distributed K-ary search principle. EPDKS takes proximity into account by combining round-trip times and the number of IP hops for interconnecting peers, thereby reducing end-to-end latencies and making effective use of the underlying network. Furthermore, we propose improvements for some of the algorithms currently implemented in the DKS. Simulation results confirm our expectations that EPDKS ensures lower end-to-end latencies and smaller number of IP hops when compared to DKS.
In this paper detailed analysis of discrete state observers from the point of view of potential time expenses required for the observer computation is presented. the aim of this article is to present, what is the infl...
详细信息
ISBN:
(纸本)9780889866331
In this paper detailed analysis of discrete state observers from the point of view of potential time expenses required for the observer computation is presented. the aim of this article is to present, what is the influence of choice of observer parameters on computation time consumption. the attention is paid most of all on the analysis of the possibility of application of the control systems in a distributed version in the context of discrete state reconstruction. All of numerical tests are conducted in the LAM/MPI environment.
this paper presents an analysis of the parallel binary merge sort algorithm and a comparison of its behavior in a generalized multiprocessor framework and a simplified Grid Environment. Based on the well known method ...
详细信息
暂无评论