The authors propose that an environment for defining services must be domain-specific (as opposed to a general-purpose programming environment). The cornerstone of such an environment is an explicit model of the netwo...
详细信息
The authors propose that an environment for defining services must be domain-specific (as opposed to a general-purpose programming environment). The cornerstone of such an environment is an explicit model of the network-services domain. This domain model captures the object universe of both the network capabilities and the user environment (e.g., line, call, customer), within which services are defined. A framework for object-oriented conceptual modeling that applies knowledge representation techniques from artificial intelligence is proposed as the basis. In this framework, service behavior is defined in terms of the operations associated with the domain model objects. This can accommodate basic service elements (BSEs) mandated by opennetwork architecture (ONA) plans. Further, service definitions themselves are treated as reusable objects and organized in a taxonomic hierarchy. The goal is to define new services via extensive reuse that incorporates existing, well-tested service definitions.< >
A concept is described in which a subset of network elements (NEs) is used to collect and transport operations data from other NEs. This can reduce the packet-switched network (PSN) access cost for the NEs that are lo...
详细信息
A concept is described in which a subset of network elements (NEs) is used to collect and transport operations data from other NEs. This can reduce the packet-switched network (PSN) access cost for the NEs that are located in the loop plant, on customer premises, and in central offices where PSN capabilities do not exist. Generic NE functions and protocol architectures for operations data networking are presented that provide communications among NEs, operations systems, and technicians to support multisupplier products in openarchitectures and to reduce the cost of network operations.
In the past three years the US Federal Communications Commission has conducted three computer inquiries in an attempt to resolve the complex regulatory issues resulting from the convergence of computer and communicati...
详细信息
In the past three years the US Federal Communications Commission has conducted three computer inquiries in an attempt to resolve the complex regulatory issues resulting from the convergence of computer and communication technologies. The author summarizes computer inquiries I, II, and III, describes the comparatively efficient interconnection opennetwork architecture (ONA) requirements set forth in computer III, and presents some basic service elements (BSEs). The status of ONA as of June 1987 and the issues affecting the defense switched network are discussed.
Summary form only given, as follows. Novel architectures are introduced which dramatically reduce the hardware and programming complexities of many of the best known neural networks. The Little/Hopfield associative me...
详细信息
Summary form only given, as follows. Novel architectures are introduced which dramatically reduce the hardware and programming complexities of many of the best known neural networks. The Little/Hopfield associative memory model and the p**t**h-order correlation models of Y. C. Lee, et al. can be faithfully implemented using only 2MN binary connections (M is the number of memories). Since the memory bit patterns are stored directly, no advance calculation of outer products is required. Dramatic reductions in complexity also occur for the Traveling Salesman and Hitchcock networks, the quadratic network, the ill-posed inversion network, and the simple p-flop.
This paper summarizes Computer Inquiries I, II, and III, describes comparably efficient interconnection (CEI) and opennetwork architecture (ONA) requirements set forth in Computer III, and presents some basic service...
详细信息
This paper summarizes Computer Inquiries I, II, and III, describes comparably efficient interconnection (CEI) and opennetwork architecture (ONA) requirements set forth in Computer III, and presents some basic service elements (BSEs). The development of ONA plans involves a complex interplay between a number of independent interests--the carriers, the equipment manufacturers, new entrepreneurs, regulators, and users, including the Department of Defense. This complex process, being conducted in regional and national forums, is a challenging and important one. Ultimately, it will affect the entire telecommunications environrnent. Participation in the ONA process by the Department of Defense is encouraged.
The following topics are dealt with: network performance evaluation;file allocation and consistency;network fault tolerance;concurrency control;distributed algorithms;performance evaluation of database systems;distrib...
详细信息
ISBN:
(纸本)0818606975
The following topics are dealt with: network performance evaluation;file allocation and consistency;network fault tolerance;concurrency control;distributed algorithms;performance evaluation of database systems;distributed architectures;operating system structures and data;performance evaluation of computer networks;distributed operating systems;tools for performance evaluation;resource allocation and scheduling;distributed systems;control techniques for distributed systems;protocol design issues;protocol analysis and verification;programming support;distributed name management;fault recovery;design issues;and, recovery support. 75 papers were presented, of which 74 are published in full in the present proceedings. Abstracts of individual papers can be found under the relevant classification codes in this or other issues.
The authors consider an architecture that is based on definition of processing elements, and characterized by node operation and a control unit directly controlling operations and system timing. Information paths conn...
详细信息
The authors consider an architecture that is based on definition of processing elements, and characterized by node operation and a control unit directly controlling operations and system timing. Information paths connecting the elements of the architecture are the control network, allowing both activation and control of node operations, and the data network, allowing data exchanges among different processing elements. MIMD-like structures may be implemented by suitably defining the information path representing the control network. Synchronization and timing are achieved by 'programming' the system at architectural level. An implementation using single-chip microcontrollers, physically interconnected via a global serial link, is presented.
The following topics are dealt with: Parallel algorithms;vector/array processing;problem mapping and scheduling;parallel architectures;systolic systems;logic programming;operating system problems;parallel algorithms/s...
详细信息
ISBN:
(纸本)0818606371
The following topics are dealt with: Parallel algorithms;vector/array processing;problem mapping and scheduling;parallel architectures;systolic systems;logic programming;operating system problems;parallel algorithms/simulation;parallel computation;languages for parallel processing;memory management;numeric processing;interconnection networks;data flow;numeric computing;parallel programming;network performance;aspects of parallel systems;performance measurement;logic programming/production systems;expressing parallelism;multiprocessor systems;parallel systems;mesh-structured systems;problem mapping techniques;Systolic systems;The IBM research parallel processor;Fault tolerance and reliability;Functional/numeric programming;Sorting. 133 papers were presented, of which 131 are published in full in the present proceedings, and 2 as abstracts only. Abstracts of individual papers can be found under the relevant classification codes in this or other issues.
This conference proceedings contains 107 papers. 104 papers are indexed separately. Topics covered include: management issues;user interface to distributed processing systems;networkarchitectures;computer-based messa...
详细信息
This conference proceedings contains 107 papers. 104 papers are indexed separately. Topics covered include: management issues;user interface to distributed processing systems;networkarchitectures;computer-based message systems;network architecture for distributed computers;human factors in distributed processing systems;design approaches to human communication systems;computer networks;distributed data design methodology;switching techniques;communications systems;multicomputer architecture;distributed operating system kernels;and, programming languages.
暂无评论