With the increase in use of distributed real-time systems (DRTS), the ability of networks to handle time-constrained traffic is becoming more and more important. This paper examines one of the most promising networks ...
详细信息
With the increase in use of distributed real-time systems (DRTS), the ability of networks to handle time-constrained traffic is becoming more and more important. This paper examines one of the most promising networks for future DRTS, namely ATM. On one hand, there exist applications with various types of time constraints associated with messages generated by control activities, database access activities, video-monitoring activities, etc., and on the other hand, two service categories, CBR and rt-VBR, have been defined by the ATM Forum to support real-time communications. We analyze some issues to be considered to provide ATM users with a formal approach to map timing constraints of messages onto ATM services. We also discuss the problem of globally allocating a set of ATM connections required by a distributed real-time application.
We present a real-time communication service suitable for distributedsystems built over ATM networks. This service offers end-to-end quality of service guarantees, and supports end-to-end priority inheritance. Priori...
详细信息
We present a real-time communication service suitable for distributedsystems built over ATM networks. This service offers end-to-end quality of service guarantees, and supports end-to-end priority inheritance. Prioritized handling of user messages is performed in each communication layer crossed, including the ATM level. Our service relies on a specific real-time ATM adaptation layer (AAL), that grants to preempt the transmission of a low priority message by a higher priority one. A transport service, in charge of admission control, resource reservation and connection setup is also presented. The real-time transport (RT-transport) service is implemented in the Chorus distributed operating system, as servers running in the microkernel address space. A real-rime interprocess communication (IPC) service runs over the RT-transport service. It offers location transparency to distributed communicating processes, and is designed to deal with deadline scheduling requirements. Our communication architecture allows communication user rates up to 130 Mbps over 155 Mbps ATM links. The end-to-end transfer delay is less than 400 /spl mu/s.
Multimedia communications have evolved rapidly during the last few years and they now require communication services with guaranteed quality of service (QoS). For the type of these new distributed multimedia applicati...
详细信息
Multimedia communications have evolved rapidly during the last few years and they now require communication services with guaranteed quality of service (QoS). For the type of these new distributed multimedia applications, it is crucial to monitor continuously the current QoS parameters of a connection to ensure that they comply with the pre-arranged negotiated values. The method proposed in this paper suggests the use of QoS monitoring agents in the ATM switches as well as in the end systems, in a way that allows automatic recovery, if possible, from all the QoS violations. The recovery will be achieved by redistributing the levels of QoS that should be supported immediately by a component to meet the end-to-end requirements. An initial architecture is proposed as well as the communication protocol between the agents and future directions are given.
The ATM Forum's PNNI (private network node interface) standard specifies a flooding mechanism to distribute topological state information amongst nodes participating in a PNNI network. While the flooding mechanism...
详细信息
The ATM Forum's PNNI (private network node interface) standard specifies a flooding mechanism to distribute topological state information amongst nodes participating in a PNNI network. While the flooding mechanism provides robust topology distribution, we show by simulation, that it can disproportionately overload lower-bandwidth links and, due to its inherent redundancy, can generate considerable computational overheads due to the processing of redundant topology updates. To address these issues, we introduce the hybrid spanning tree algorithm, a spanning tree-based topology distribution mechanism that has low computational maintenance and can support policy that restricts topology distribution control traffic from being carried over lower-bandwidth links. Unlike other spanning tree proposals, the hybrid spanning tree algorithm provides a simple and practical migration path to smoothly transition PNNI nodes executing the flooding algorithm to that of the hybrid spanning tree algorithm.
This paper introduces a new concept and model for building complex ATE systems. The underlying principles were discovered over several years but were driven home by the difficulty found in modernizing a legacy test sy...
详细信息
This paper introduces a new concept and model for building complex ATE systems. The underlying principles were discovered over several years but were driven home by the difficulty found in modernizing a legacy test system. The result of this effort produced a solution that provided answers to the following three questions: (1) How to achieve asset interchangeability in complex test and measurement systems; (2) How to place test system software in a modular component form that can be reused in different ATE or desk top environments, and (3) How to apply new software technologies in distributedcomputing to ATE systems.
Mobility will be one key issue for future broadband access. Consequently, many research projects are currently focusing on wireless ATM networks. Within the European ACTS research program the "Magic WAND" pr...
详细信息
Mobility will be one key issue for future broadband access. Consequently, many research projects are currently focusing on wireless ATM networks. Within the European ACTS research program the "Magic WAND" project is developing mobile terminals and access points based on wireless ATM technology. The terminals will communicate at a speed up to 20 Mbit/s in the 5 GHz band within multiple picocells of approximately 50 meters in diameter. This technology supports applications and services demanding both high bandwidth and mobile network access, two contradicting requirements in the past. We take a look at typical multimedia applications and evaluate the key components of the WAND system necessary to support their quality of service (QoS) requirements. One major challenge is to guarantee QoS transparently over the wireless link. We present our solutions for a QoS-supported mobile audio/videoconferencing subsystem and analyse how video compression algorithms behave under the conditions of a wireless medium.
Controlling torsional vibrations of nonlinear driven distributed parameter systems requires a good understanding of the dynamics as well as low dimensional models. Singular value analysis and proper orthogonal decompo...
详细信息
ISBN:
(纸本)078034247X
Controlling torsional vibrations of nonlinear driven distributed parameter systems requires a good understanding of the dynamics as well as low dimensional models. Singular value analysis and proper orthogonal decomposition are employed for dimension analysis and dimension reduction of a one-dimensional continuous system. The structure of an active damping system based on the reduced model is presented and the theoretical findings are verified with an experimental set-up of an extremely slender torsional string. Considerable improvements of the system dynamics are observed.
The reliability of a distributed database systems is the probability that a program which runs on multiple processing elements and needs to communicate with other processing elements for remote database will be execut...
详细信息
The reliability of a distributed database systems is the probability that a program which runs on multiple processing elements and needs to communicate with other processing elements for remote database will be executed successfully. This reliability varies according to 1) the topology of the distributed database system, 2) the reliability of the communication links, 3) the databases and program distribution among processing elements, and 4) the databases required to execute a program. This paper shows that solving this reliability problem is NP-hard even when the distributed database system is restricted to a series-parallel, a 2-tree, a tree, or a star structure. Two polynomial-time algorithms are proposed for computing the reliability of a distributed program which runs on a linear and a ring distributed database system, respectively.
In this paper, a simulation system with distributed intelligence realised by means of chaotic speed independent integrators is considered, A fixation method of transient processes in combinational logical circuits is ...
详细信息
In this paper, a simulation system with distributed intelligence realised by means of chaotic speed independent integrators is considered, A fixation method of transient processes in combinational logical circuits is discussed, The procedure of chaotic integration is proposed and a variant of circuit design of a chaotic integrator is presented, The connection between chaotic and fuzzy integration is investigated, The way of determination of the time moment of the end of transient processes in chaotic integrators is discussed and results of experimental investigations are presented, The problems connected with hardware realisation of functional integration elements of the simulation system for solving parabolic partial differential equations are considered, The results of computer simulation of the system with chaotic integrators for solving this problem are presented.
RPC (Remote Procedure Call) protocol is one of popular communication mechanisms. It is very simple and transparent to write distributed programs. RPC protocol user need not to have information on distributed environme...
详细信息
RPC (Remote Procedure Call) protocol is one of popular communication mechanisms. It is very simple and transparent to write distributed programs. RPC protocol user need not to have information on distributed environments and can easily construct distributed application systems. RPC protocol reduces overload of communications since it uses messages to communicate. This paper presents a group communications to cooperate with RPC. After creating a specific group by user request, they can communicate each other. A group RPC protocol can improve reliability, transparency and facility of classic RPC protocol. The group RPC proposed in this paper can be used in various applications such as video conference, replicated distributed database and management of distributed network.
暂无评论