The parallel implementation of a 3D-BEM algorithm was investigated. This was done on multiprocessor workstations in order to demonstrate that parallelization is today possible and useful not only on supercomputers and...
详细信息
The parallel implementation of a 3D-BEM algorithm was investigated. This was done on multiprocessor workstations in order to demonstrate that parallelization is today possible and useful not only on supercomputers and large scale massively parallel machines. The serial implementation was investigated for parallel potential. It is shown that function level parallelism in the input phase does not further accelerate computation. Good results can be obtained by parallelization of the matrix filling due to a high degree of data independence in this subproblem.
The widespread proliferation of Internet connections has made current computer networks more vulnerable to intrusions than before. In network intrusions' there may be multiple computing nodes that are attacked by ...
详细信息
ISBN:
(纸本)0769522947
The widespread proliferation of Internet connections has made current computer networks more vulnerable to intrusions than before. In network intrusions' there may be multiple computing nodes that are attacked by intruders. The evidences of intrusions have to be gathered from all such attacked nodes. An intruder may move between multiple nodes in the network to conceal the origin of attack., or misuse some compromised hosts to launch the attack on other nodes. To detect such intrusion activities spread over the whole network, we present a new intrusion detection system (IDS) called distributed Intrusion Detection using Mobile Agents (DIDAM). DIDAM uses a set of software entities called mobile agents that can move from one node to another node within a network, and perform the task of aggregation and correlation of the intrusion related data that it receives from another set of software entities called the static agents. Mobile agents reduce network bandwidth usage by moving data analysis computation to the location of the intrusion data, support heterogeneous plat-forms, and offer a lot of flexibility in creating a distributed IDS. DIDAM utilizes the above-mentioned beneficial features offered by mobile agent technology and addresses some of the issues with centralized IDS models. The detailed architecture and implementation of a prototype of DIDMA are described It has been tested using some well-known attacks and performances have been com-pared with centralized IDS models.
With the rapid emergence of peer-to-peer applications, more and more applications will generate many-to-one rather than one-to-one traffic flows. While a number of previous works have reported the measured characteris...
详细信息
ISBN:
(纸本)0769526373
With the rapid emergence of peer-to-peer applications, more and more applications will generate many-to-one rather than one-to-one traffic flows. While a number of previous works have reported the measured characteristics o available If network bandwidth, these measurements are primarily conducted for one-to-one flows or for a physical network link. This paper addresses this void by reporting the characteristics of aggregate available bandwidth in many-to-one dataflows over TCP. Based on extensive measurements conducted in the PlanetLab and public FTP servers, we analyze the statistical properties of aggregate data flows originating from multiple senders destined to the same receiver. Our results show that many-to-one data flows exhibit substantially more consistent statistical properties over a long time scale, and thus could open a new way to achieve probabilistic performance guarantees in multimedia applications.
This paper presents a new rapid thread replacement mechanism which is important in multithread technology. Analysis to the memory system indicates that the memory utilization decreases with the increase of cache hit r...
详细信息
ISBN:
(纸本)0818678763
This paper presents a new rapid thread replacement mechanism which is important in multithread technology. Analysis to the memory system indicates that the memory utilization decreases with the increase of cache hit ratio. The parallelism between thread computation and thread replacement is found by analyzing their working processes. Based on these, we advance a rapid multithread replacement mechanism which overlaps the thread replacement with thread computation. More especially, with finite hardware contexts, this mechanism can play the same role of infinite contexts by tolerating the replacement overhead. By modifing the general thread switching model, we bulid the thread replacement model and evaluate this mechanism in theory and experiment methods. At last, we discuss the hardware implementation and put forward the problems to be resolved in the future.
Although some known parameter study tools [1] have successfully been used to solve standard parameter study problems, they are not yet sufficient to handle the current generation of complex parameter investigations. F...
详细信息
Although some known parameter study tools [1] have successfully been used to solve standard parameter study problems, they are not yet sufficient to handle the current generation of complex parameter investigations. Furthermore, time constraints typically require that multiparametric tasks are executed on a large number of parallel computational resources. Grid Computing potentially helps by enabling access to vast computational resources. But in practice the application of multiple Grid resources to complex tasks leads to considerable programming effort for the scientist. One way to reduce both the work and the cost associated with performing parametric investigations of complex scientific experiments could be to automate the creation of complex modules for the computation and the dynamic control of the study, for example through the use of the results from previous computational stages. In this paper we propose a concept for the design and implementation of an automated parametric modeling system for producing complex dynamically-controlled parameter studies without the use of any specific programming language in the process of parameterization, computation and the control of the experiment.
Most previous parallel logic programming systems have been built on top of classical operating systems. The advances in the area of parallel operating systems have made it possible to explore new execution models that...
详细信息
With the capability of employing virtually unlimited compute resources, the cloud evolved into an attractive execution environment for applications from the High Performance Computing (HPC) domain. By means of elastic...
详细信息
ISBN:
(纸本)9789897583650
With the capability of employing virtually unlimited compute resources, the cloud evolved into an attractive execution environment for applications from the High Performance Computing (HPC) domain. By means of elastic scaling, compute resources can be provisioned and decommissioned at runtime. This gives rise to a new concept in HPC: Elasticity of parallel computations. However, it is still an open research question to which extent HPC applications can benefit from elastic scaling and how to leverage elasticity of parallel computations. In this paper, we discuss how to address these challenges for HPC applications with dynamic task parallelism and present TASKWORK, a cloud-aware runtime system based on our findings. TASKWORK enables the implementation of elastic HPC applications by means of higher-level development frameworks and solves corresponding coordination problems based on Apache ZooKeeper. For evaluation purposes, we discuss a development framework for parallel branch-and-bound based on TASKWORK, show how to implement an elastic HPC application, and report on measurements with respect to parallel efficiency and elastic scaling.
With recent improvements in single CPU performance, several issues become more important in multiprocessor design. Two of these are interprocessor communication and locality. In parallelsystems with fast CPUs, locali...
详细信息
ISBN:
(纸本)0818626720
With recent improvements in single CPU performance, several issues become more important in multiprocessor design. Two of these are interprocessor communication and locality. In parallelsystems with fast CPUs, locality is vital to performance. However, traditional parallel programming models such as shared memory or message passing do not naturally lead to programs that exhibit locality. In this paper, the Seamless model for interprocessor communication is presented which is based on locality and that allows the programmer to explicitly manipulate a program's locality to optimize performance. Additionally, this model can support latency tolerance with proper hardware support. Extensions to the C programming language that support this model are also presented. Finally, a parallel program utilizing this model is provided to illustrate the paradigm.
distributed dataflow systems such as Apache Spark or Apache Flink enable parallel, in-memory data processing on large clusters of commodity hardware. Consequently, the appropriate amount of memory to allocate to the c...
详细信息
ISBN:
(纸本)9798400707469
distributed dataflow systems such as Apache Spark or Apache Flink enable parallel, in-memory data processing on large clusters of commodity hardware. Consequently, the appropriate amount of memory to allocate to the cluster is a crucial consideration. In this paper, we analyze the challenge of efficient resource allocation for distributed data processing, focusing on memory. We emphasize that in-memory processing with in-memory data processing frameworks can undermine resource efficiency. Based on the findings of our trace data analysis, we compile requirements towards an automated solution for efficient cluster resource allocation.
The vast majority of research projects involving Grid Computing have focused on the development and standardization of the middleware that allows Grids to function. Despite the maturity of the middleware as the Grid c...
详细信息
ISBN:
(纸本)0769526373
The vast majority of research projects involving Grid Computing have focused on the development and standardization of the middleware that allows Grids to function. Despite the maturity of the middleware as the Grid core, the actual application of Grids to solving concrete problems is far from being a common practice, much less a standard one. In this context, we present ResGrid, a Grid-aware toolkit for reservoir uncertainty analysis. With the help of ResGrid, a reservoir engineer can transparently take advantage of Grid resources and services for compute-intensive and data-intensive uncertainty analysis while achieving a better understanding of multiphase reservoir modeling. ResGrid improves the efficiency of reservoir analysis, and its flexible architecture allows its utilization by other application areas such as coastal modeling and climate simulations. ResGrid is our first step towards developing generic Grid-aware toolkits that could be applied to any science in a cost- and time-efficient manner.
暂无评论