ForCES (Forwarding and Control Element Separation) architecture brought forward by IETF shows great advantages for network devices such as routers, wireless base stations and etc to be flexible, programmable, and cost...
详细信息
ISBN:
(纸本)9780769535012
ForCES (Forwarding and Control Element Separation) architecture brought forward by IETF shows great advantages for network devices such as routers, wireless base stations and etc to be flexible, programmable, and cost-effective. This paper researches on the transaction mechanism in ForCES router, based on the analysis of transaction requirement of ForCES, two phase commit algorithm is introduced and a state-machine based transaction implementation scheme is proposed. ForCES router test bed is built. The experiment result verifies the correctness and feasibility of the transaction mechanism proposed and implemented.
Peer-to-peer systems use redundant data replicas to maintain high availability and improve on user response times. The consistency and coherency of these data replicas need to be maintained over time in face of variou...
详细信息
ISBN:
(纸本)9783540922940
Peer-to-peer systems use redundant data replicas to maintain high availability and improve on user response times. The consistency and coherency of these data replicas need to be maintained over time in face of various system changes like node joins, failures and network outages. In this paper we present a robust approach to guarantee eventual coherency of replicas in a multi-location large scale peer-to-peer distributed file system. We use a combination of data pull and push mechanisms, and a last coherent time stamp on each replica. These mechanisms ensure that no user read operation ever retrieves data that is older than a configurable upper bound in time.
This paper addresses maximum a posteriori (MAP) estimation algorithm for multiple frequency offsets in distributed multiple inputs and multiple outputs (MIMO) system in the presence of prior information about frequenc...
详细信息
ISBN:
(纸本)9781424436927
This paper addresses maximum a posteriori (MAP) estimation algorithm for multiple frequency offsets in distributed multiple inputs and multiple outputs (MIMO) system in the presence of prior information about frequency offsets. In this paper, the prior information is considered and the Cramer-Rao Bound (CRB) for the problem at hand is derived. Additionally, taking advantage of prior information, the algorithm of maximum a posteriori is investigated. First, the variance of prior information can be estimated according to the range of frequency offsets. Then the MAP estimation will be used to estimate the frequency offsets. Compared with the maximum-likelihood estimation, simulation results illustrate the performance of MAP estimator achieve the CRB and the proposed algorithm can improve the efficiency of synchronization obviously, especially in low Signal to Noise Ratio (SNR) environment.
The problem of security in communication has been researched to Ad Hoc network in this paper, and it also has been proposed a distributed intrusion detection model which is based on the fuzzy judgment. The judging rul...
详细信息
ISBN:
(纸本)9781424436927
The problem of security in communication has been researched to Ad Hoc network in this paper, and it also has been proposed a distributed intrusion detection model which is based on the fuzzy judgment. The judging rules of the aggressive behavior are summarized in this model according to the different offensive features in routing layer. Moreover, the log is established according to IP packets in routing layer, and the fuzzy evaluation to the data is carried on. On the base of this, a way to enhance the correct rate of the detection by joint detection mechanism of multi-point cooperation has been introduced in this paper. To distribute the network resource dynamically and reasonably, all nodes are treated dynamically with safety classification. Taking AODV routing protocol as an example, the working mechanism of the model is introduced, and the software NS2 is used to put up network simulation platform. The experimental results show that the detection model can guarantee the network security better.
k-Anycast is a communication paradigm that is proposed to deliver a packet to any threshold k members of a set of communication group hosts. Most k-Anycast routing protocols are centralized and they are not suitable f...
详细信息
ISBN:
(纸本)9781424436927
k-Anycast is a communication paradigm that is proposed to deliver a packet to any threshold k members of a set of communication group hosts. Most k-Anycast routing protocols are centralized and they are not suitable for high dynamic networks (such as wireless sensor networks and wireless ad hoe networks) because they assume that each routing node has global network information. A distributed k-Anycast routing protocol based on mobile agents is proposed in this paper. The protocol forms multiple components and each component has at least k members. Each component can be treated as a virtual server, so k-Anycast service is distributed to each component. In this protocol, mobile agents are applied to initiate or manage components and routing table, so each routing node needn't global network information, only needs to exchange routing information with its neighbors, so the protocol saves much communication cost and adapts to high dynamic networks. Compared to flooding method and centralized k-Anycast routing protocols, performances of the protocol are improved in terms of cost, time delay and scalability.
For a complex distributed system to be dependable it must be continuously monitored, so that its failures and 'imperfections can be discovered and corrected in a timely manner. This work is concerned with the moni...
详细信息
ISBN:
(纸本)9789639799769
For a complex distributed system to be dependable it must be continuously monitored, so that its failures and 'imperfections can be discovered and corrected in a timely manner. This work is concerned with the monitoring of large, open and heterogeneous systems, at their application level. Our objective is a monitoring technique that satisfies the following properties: scalability with respect to the size of the system and with the complexity of the monitoring task;the ability to deal reliably with heterogeneous components;and the ease and flexibility of deployment. Our approach to monitoring is based on a middleware called Law-Governed Interaction (LGI), which is a decentralized coordination and control mechanism.
In wireless cellular communication, it is essential to find effective means for power control of signals received from randomly dispersed users. In this paper, a novel distributed power control game algorithm, focusin...
详细信息
This position paper outlines the need and rational for developing large-scale emulation facilities structured to allow the scientific method tenets to be met on a per experiment basis. The work specifically focuses on...
详细信息
ISBN:
(纸本)9781424439997
This position paper outlines the need and rational for developing large-scale emulation facilities structured to allow the scientific method tenets to be met on a per experiment basis. The work specifically focuses on the need to develop emulation-based test beds on the 1000+ machine scale, as expressed within the US. Defense Advanced Research Program Agency's (DARPA) BAA-08-43 Broad Agency Announcement of May 2008 for a National Cyber Range, the University of Victoria's Fall 2008 application to the Canadian Foundation for Innovation for a Canadian at-scale Emulation Laboratory (CASElab), and the recent HP-Intel-Yahoo global cloud computing test bed initiative. The work places these proposed large-scale facilities both within the general context of the standard research tools (i.e., analytical analysis, simulation studies, ad hoc testing, and smaller-scale emulation), as their placement against other available test beds, most notably Emulab, DETERlab, and PlanetLab.
Dynamic scheduling has been always a challenging problem for real-time distributed systems. EDF (Earliest Deadline First) algorithm has been proved to be optimal scheduling algorithm for single processor real-time sys...
详细信息
ISBN:
(纸本)9781615676606
Dynamic scheduling has been always a challenging problem for real-time distributed systems. EDF (Earliest Deadline First) algorithm has been proved to be optimal scheduling algorithm for single processor real-time systems and it performs well for distributed systems also. Limitation of EDF is that it cannot perform well when the system becomes overloaded. In this paper, we have proposed a scheduling algorithm for client/server distributed system with soft timing constraints. The algorithm is proposed with modification in conventional EDF algorithm. The proposed algorithm is simulated;results are obtained and measured in terms of SR (Success Ratio) & ECU (Effective CPU Utilization). Finally the results are compared with EDF algorithm in the same environment. It has been observed that the proposed algorithm is equally efficient during underloaded conditions and it performs much better during overloaded conditions.
This paper considers the implementation of parallel Grobner bases algorithms on distributed memory parallel computers like todays high performance computing clusters (HPC) in the Java computer algebra library (JAS). T...
详细信息
ISBN:
(纸本)9781424435692
This paper considers the implementation of parallel Grobner bases algorithms on distributed memory parallel computers like todays high performance computing clusters (HPC) in the Java computer algebra library (JAS). The implementation of Grobner bases algorithms is part of the essential building blocks for any computation in algebraic geometry. There are various implementations of these algorithms in procedural programming languages. Our aim is an implementation in a modern object oriented programming language with generic data types, as it is provided by Java programming language. The proposed middle-ware uses a distributed list for the communication of the reduction bases and a distributed thread pool for running the reduction engines in different computers. Java object serialization is used to encode polynomial objects for network transport via Java networking facilities.
暂无评论