For safety-critical applications, accurately evaluating network dependability is crucial. This article, a special selection from the Symposium on Integrated Circuits and Systems design (SBCCI), describes a fault-injec...
详细信息
For safety-critical applications, accurately evaluating network dependability is crucial. This article, a special selection from the Symposium on Integrated Circuits and Systems design (SBCCI), describes a fault-injection environment for assessing CAN-based networks common in the automotive field. The approach focuses particularly on revealing how soft errors within the CAN protocol controllers affect system behavior.
With the increasing demand for vast storage repositories, network storage has become important for mass data storage and processing, telescopic addressing and availability, and the quality of service and security of d...
详细信息
With the increasing demand for vast storage repositories, network storage has become important for mass data storage and processing, telescopic addressing and availability, and the quality of service and security of data storage. This situation demands the emergence of new technology in the data storage field. In this paper, TH-MSNS, a SAN system, is introduced. This system was designed and implemented based on the Fiber Channel Protocol and its I/O route was tested. This paper introduces some of the key techniques in the network storage system, including an SCSI simulating target, intelligent and uniform storage management architecture, and the processing flow of the read/write commands. The software for the new storage area network system was implemented as a module in the kernel mode to improve its efficiency. The SCSI target adopts a layered design and standardized interface, which is compatible with various types of SCSI devices and can use different network protocols. The storage management software adopts distributed architecture, which enables higher interoperability and compatibility with various kinds of management protocols. TH-MSNS boasts characteristics such as high adaptability, high efficiency, high scalability, and high compatibility and is easy to maintain.
Many SoC applications require guaranteed levels of service and performance. Can networks on chips ( NoCs) enable such guarantees? Here, the authors demonstrate that the AEthereal network can. This particular NoC, deve...
详细信息
Many SoC applications require guaranteed levels of service and performance. Can networks on chips ( NoCs) enable such guarantees? Here, the authors demonstrate that the AEthereal network can. This particular NoC, developed at Philips Research Laboratories, encompasses hardware, a programming model, and a design flow. Read on to find out about the details.
In this paper, we propose a link layer design for mobile hotspots. We design a novel system architecture that enables highspeed Internet access in railway systems. The proposed design uses a number of repeaters placed...
详细信息
In this paper, we propose a link layer design for mobile hotspots. We design a novel system architecture that enables highspeed Internet access in railway systems. The proposed design uses a number of repeaters placed along the track and multiple antennas installed on the roof of a vehicle. Each packet is decomposed into smaller fragments and relayed to the vehicle via adjacent repeaters. We also use erasure coding to add parity fragments to original data. This approach is called information raining since fragments are rained upon the vehicle from adjacent repeaters. We investigate two instances of information raining. In blind information raining, all repeaters awaken when they sense the presence of the vehicle. The fragments are then blindly transmitted via awakened repeaters. A vehicle station installed inside the train is responsible for aggregating a large enough number of fragments. In the throughput-optimized information raining, the vehicle station selects a bipartite matching between repeaters and roof-top antennas and activates only a subset of the repeaters. It also dictates the amount of transmission power of each activated repeater. Both the bipartite matching and power allocations are individually shown to be NP-complete. Matching heuristics based on the Hungarian algorithm and Gale-Shapley algorithm are proposed. A simplex-type algorithm is proposed as the power allocation heuristics.
Background: Modern health care is provided with close cooperation among many different institutions and professionals, using their specialized expertise in a common effort to deliver best-quality and, at the same time...
详细信息
Background: Modern health care is provided with close cooperation among many different institutions and professionals, using their specialized expertise in a common effort to deliver best-quality and, at the same time, cost-effective services. Within this context of the growing need for information exchange, the demand for realization of data networks interconnecting various health care institutions at a regional level, as well as a national level, has become a practical necessity. Objectives: To present the technical solution that is under consideration for implementing and interconnecting regional health care data networks in the Hellenic National Health System. Methods: The most critical requirements for deploying such a regional health care data network were identified as: fast implementation, security, quality of service, availability, performance, and technical support. Results: The solution proposed is the use of proper virtual private network technologies for implementing functionally-interconnected regional health care data networks. Conclusions: The regional health care data network is considered to be a critical infrastructure for further development and penetration of information and communication technologies in the Hellenic National Health System. Therefore, a technical approach was planned, in order to have a fast cost-effective implementation, conforming to certain specifications.
The development of a Broadband Computer Communications network (Breitband-Wissenschaftsnetzes "B-WiN") became increasingly necessary due to the constantly growing need of facilities in research, science and ...
详细信息
Admission control and bandwidth allocation are important issues in telecommunications networks, especially when there are random fluctuating demands for service and variations in the service rates. In the emerging bro...
详细信息
ISBN:
(纸本)0444818340
Admission control and bandwidth allocation are important issues in telecommunications networks, especially when there are random fluctuating demands for service and variations in the service rates. In the emerging broadband communications environment these services are likely to be offered via an ATM network. In order to make ATM future safe, methods for controlling the network should not be based on the properties of present services. We propose one bandwidth allocation method which has this property. Our proposed approach is based on pricing bandwidth to reflect network utilization, with users competing for resources according to their individual bandwidth valuations. The prices may be components of an actual tariff or they may be used as control signals, as in a private network. Simulation results show the improvement possible with our scheme versus a leaky bucket method in terms of cell loss probability, and confirm that a small queue with pricing: can be efficient to multiplex heterogeneous sources.
This paper is a response to concerns about the performance of TCP/IP over high-speed backbone networks such as those based on ATM. We consider two versions of TCP: the widely used Tahoe version proposed by Van Jacobso...
详细信息
ISBN:
(纸本)044482023X
This paper is a response to concerns about the performance of TCP/IP over high-speed backbone networks such as those based on ATM. We consider two versions of TCP: the widely used Tahoe version proposed by Van Jacobson, and a more recent proposal (also by Van Jacobson) which uses fast retransmit to recover from losses. We attempt to develop a basic understanding, using analysis and simulation, of these protocols in the regime of high bandwidth-delay products and random loss by considering a single bottleneck link with one-way traffic. Our main results are as follows. For the Tahoe version, rapid growth in window size due to slow start leads to excessive buffer requirements when the bandwidth-delay product is large. This phenomenon does not occur (except as a transient) if fast retransmit is used. However, there are a number of shortcomings that are common to both versions of TCP considered here. Random packet loss leads to a significant deterioration in throughput when the product of the loss probability and the square of the bandwidth-delay product is larger than one. For multiple connections, we find that the throughput of a connection with propagation delay tau is roughly proportional to 1/tau(a), where less than or equal to a less than or equal to 2, sq that the how control mechanism discriminates significantly against connections with larger propagation delays. Our results show that modifications both in the end-to-end congestion control scheme and in the queueing disciplines in the switches may be required for good performance over high-speed backbones.
B-ISDN is envisioned to be the universal networking technology and ATM is the transfer mode of choice for B-ISDN. There has been a tremendous amount of work completed on ATM standards. However, enabling ATM networks r...
详细信息
ISBN:
(纸本)044482023X
B-ISDN is envisioned to be the universal networking technology and ATM is the transfer mode of choice for B-ISDN. There has been a tremendous amount of work completed on ATM standards. However, enabling ATM networks requires the design and development of several additional network services that are not within the scope of any standards activity. IBM's high speed networking architecture, Broadband network Services, complements emerging ATM standards and provides a complete architectural framework that enables high speed, multimedia networking while utilizing network resources efficiently. This paper presents a high level review of Broadband network Services and discusses how the challenges of high speed multimedia networks are addressed in this architecture.
The complexity of the path selection algorithms is a function of the network size, and can become significant for very large networks requiring a high call processing rate. A possible way to relax the computational re...
详细信息
ISBN:
(纸本)044482023X
The complexity of the path selection algorithms is a function of the network size, and can become significant for very large networks requiring a high call processing rate. A possible way to relax the computational requirement is to cluster the network into a. backbone and several subareas. This decomposition model is based on the typical highspeed network structure: the backbone is the transport network and the subareas include the access nodes. Path selection is run only on a subset of the network nodes, according to the localization of the source and the destination node. However, clustering a network is not an easy task since for large network it may lead to a decomposition that compromises the optimality of the path selection algorithm. In this paper we present a new automatic network clustering technique that leads in most cases to an optimal decomposition, that is the optimal path between two nodes can always be found. The use of this decomposition algorithm leads to a significant speed-up of the path selection algorithm.
暂无评论