Computer Networks (1976)最新文献

A process offering a service in a distributed system receives request messages from client processes. For proper system behavior, it is often necessary for requests to be serviced in the same order that they are sent. Ensuring this order requires an elaborate scheme, unless certain restrictions are imposed on the distributed system.

We present some restrictions that ensure that requests are received at a server in the same order that they are sent. With these restrictions, each server can safely service requests in the order of arrival. The restrictions either apply to the message routing, or to the client/server communication. The restrictions on message routing are often satisfied by systems based on local networks, and the restrictions on client/server communication are satisfied by some existing message passing schemes.

This paper describes the “TPA/70-X.25 gateway-network” of the International Institute for Applied Systems Analysis (IIASA) in Austria, and of the Institute for Computerization and Automation of the Hungarian Academy of Sciences (SZTAKI) in Hungary, and its promoting role in the on-line exchange of scientific information among national and international institutes and organizations. It presents a short overview of the major categories of transborder data flows relevant to IIASA's work, and how the gateway-network handles them. Finally, some operational and technical aspects of this East-West network of gateways are discussed.

The CSIRO Division of Computing Research has developed an Australia-wide packet switching network, CSIRONET. An interface between CSIRONET and Digital Equipment Corporation's ANF-10 network products has been developed. This interface has been used to connect CSIRONET to the ANF-10 based networks of two Australian Universities. The connections support internetwork terminal and file traffic.

We describe and analyze an efficient collision-free channel-access protocol for cable or radio networks with an arbitrary spatial (one-, two- or three-dimensional) fixed node configuration.

The protocol is robust and provides distributed access-control under a myriad of possible priority disciplines, including Fixed, Fair Round-Robin and Prioritized Round-Robin disciplines. The protocol optimally employs available information on network topology, to provide performance characteristics (throughput and delay) that are at least as good as and in most cases much better than those of other published protocols that employ less information. The performance advantage is especially significant for networks with a large number of nodes.

We present a contention transmission algorithm for local networks which is related to the p-persistent family of algorithms previously analyzed by Kleinrock and Tobagi. Our algorithm incorporates an automatic acknowledgement signal which is sent by the receiving station immediately after each packet is correctly received. The algorithm is suitable for radio environments in which transmitting stations can “capture” nearby receivers. Assuming that stations are not placed too close each other, we prove that an acknowledgement signal is received by a transmitting station only if the packet has been received correctly by the intended receiver. In a cable network where transmitters do not capture receivers, the acknowledgement signal is guaranteed (in the same sense) regardless of the distances between stations. This result depends on very weak assumptions about the type of data encoding used on the channel. We discuss the more interesting aspects of implementing this scheme, and we outline how higher-level protocols can make use of the acknowledgement signal and its properties.