leff@smu.UUCP (Laurence Leff) (07/11/89)
%C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The purpose of this paper is to broaden the scope of understanding of Network Management. The fundamental requirements for Network Management are to provide services to an entire corporation. .sp 0.4 Technological approaches to network management are as numerous and as varied as the requirement of corporate networks. Many of the present NM solutions address real operational problems however provisions for interfacing the Information Management Network department with the other corporate departments is only beginning to be understood as a requirement of NM. .sp 0.4 ISO has resolved that management [1] of a network requires disciplines that fall into five definitive categories: Fault, Configuration, Accounting, Performance, Security; (FCAPS) [2] Whereas numerous groups are addressing these specific categories in depth, this paper is intended to foster discussion of broader application for the excellent FCAPS principles already being developed. These extensions may be beyond the present charter of OSI. Figure 1 shows the familiar 7 layer stack as the logical network model within the context of the physical network and the organizational network of typical corporate departmental user. %A A. Dupuy %A J. Schwartz %A Y. Yemini %A G. Barzilai %A A. Cahana %T Network Fault Management: A User's View %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 101-109 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper has two goals. First, describe typical fault management problems arising in large-scale, heterogeneous, real-time transaction processing networks. We aim to abstract the generic technical questions, while providing real examples of issues arising in fault detection and recovery. Second, report early results of cooperative research conducted by the authors towards effective solutions of these problems. This work resulted in a network control monitor (NCM), providing comprehensive graphical display of complex network situations, currently serving as the control point of a large scale banking network. Complementary simulation testbed tools, being developed over NEST [1], support analysis of transient network behaviors and studies of incremental network changes. These efforts resulted in significant unification and simplification of critical network management tasks. %A D. Milham %A K. Willetts %T British Telecom's Communications Management Architecture %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 109-119 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper reviews some of the studies conducted by British Telecom (BT) into the development of an architectural approach to communications management. The approach has been driven by consideration of the fundamental needs in managing resource or endeavour, and applying these to Telecommunication Networks. It is also based on the OSI Management Standards emerging in CCITT and ISO, and forms part of BT's strategic initiative: ONA Management Architecture. .sp 0.4 The paper covers two main themes. The first defines the scope and requirement of the architecture based upon the business needs of both, a ``telco,'' and its customers. The second describes the role of the architecture in the specification methodology used to translate specific business objectives into the technical specifications used for developement and procurement. %A W. Collins %T OSI Management Service Elements, Protocols and Application Layer Structure (ALS) %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 119-132 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The standards under development for OSI Systems Management Service Elements and Protocol play an important role in contemporary communication networks. Correct application of these services elements and protocols results in a stable OSI System Management Environment. This article discusses some components of the Application Layer known as Application Services Elements (ASEs). The ASEs require direct user input. Specifically the article focuses on three specific sub-components of OSI Systems Management Services elements and Protocols: CMISE, ROSE, and ACSE that are necessary to access valuable services of OSI network systems. %A P. Kennedy %A C. Sluman %A P. Pranschke %T Management Requirements Arising from a NATO Study of Quality of Service %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 133-140 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper presents an analysis of Quality of Service (QoS) requirements and issues in the general open systems environment. Specifically, it considers the question of QoS with respect to its impact on the management of systems and networks by proposing an architecture within which various QoS issues may be discussed. %A L. Tarouco %T Intelligent Network Management %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 141-158 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP In order to achieve efficient network management intelligent support should be provided to help the identification, trace and solve of situations involving physical fault, software malfunction or aspects of training which affect the correct functioning of the network. In this paper desirable characteristics in network management support are discussed and the state of art and trends in network management and expert systems applied to network management are presented. %A Jeremy Tucker %T A Common Approach to Managed Objects %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 159-168 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Many different communities, including OSI, are developing network management tools. Because networks will be interconnected, and also because OSI does not encompass all management requirements, these tools will need to coexist; and systems will need to manage other systems from different management domains with different control regimes. So all possible forms of compatibility between different network management developments are to be encouraged. This paper considers the definition of managed objects and suggests that the ``core'' capability developed for OSI could with advantage also be adopted beyond the confines of OSI. However, a number of technical problems still remain to be solved. %A M. Sylor %T Guidelines for Structuring Manageable Entities %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 169-184 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Digital's Enterprise Management Model describes those things which are managed as abstract objects called entities. This paper describes a technique for designing entities. The technique has seven steps: requirements analysis, information modeling, operation modeling, events and counters, naming and relationships, iteration, and documentation. The method is based on both current theory and experience gathered from developing the entities of Phase V of Digital's Network Architecture (DNA). %A K. McCloghrie %A M. Rose %A C. Partridge %T Defining a Protocol-Independent Management Information Base %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 185-197 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP TCP/IP network management standardization has become focussed on two protocols: the Simple Network Management Protocol (SNMP) as the short-term standard; and OSI's Common Management Information running over TCP/IP (CMOT) as the (likely) standard for the future. In order to preserve short-term implementation investments and to ease the transition from short-term to longer-term, a common Management Information Base (MIB) has been defined for both operational use by SNMP and experimental use by CMOT. Experiences have resulted in practical guidelines being developed for protocol independent management frameworks. %A J. Crawford %T Graphics for Network Management: An Interactive Approach %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 197-210 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The reliability of a telecommunications network is greatly influenced by how effectively network operations staff identify and respond to problems as they occur. To assist in this task, many Network Management systems provide graphic representations of the network, but these graphics are often inflexible and diffuculties to adapt to changing network conditions. We addressed these problems in the development of NETWORKS, a new Network Management System incorporating innovative object-oriented graphic techniques. This paper describes these techniques and discusses how graphic objects can be used to help solve network management problems. %A W. Cellary %A M. Stoinski %T Performance Management Architecture for Protocol Entity Optimization %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 227-234 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %A S. Mazumdar %A A. Lazar %T Knowledge-Based Monitoring of Integrated Networks %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 235-246 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %A H.L. Sorkin %T OSI Independent Audit: the Audit Trail %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 247-256 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The ISO OSI standards provide for operational information that can be used by the security manager. Provisions are not made to provide information that satisfies criteria for an independent audit. This paper extends the ISO OSI model to provide for an independent audit trail. %A P.J. Sevcik %A L. Korn %T A Network Monitoring and Control Security Architecture %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 257-268 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper describes an architectural design for a secure Monitoring Center System (MCS) for the monitoring and control of a wide area network. Presented here is a general security architecture for a MCS to ensure protection of sensitive information from unauthorized internal uners and outside intruders. The approach is based on United States DoD requirements and security guidelines, but it is also applicable to commercial environments. %A Y. Lamdan %A B. Iyer %A G. Barzilai %A A. Cahana %T Dialog Manager for Network Management %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 269-278 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Multi-vendor communications networks are common today. Centralizing such networks managment is an ongoing difficult problem. Common network management interface standards are a long way down the road before they become a reality; therefore, communications with the various vendor-supplied network management systems becomes a complex issue. This motivated us to devise a powerful and flexible interface scheme. We developed a new language and an interpreter to carry on sessions. The new language allows for the definition and control of ``conversations'' (dialogs) between two parties. Special constructs support sending messages and analyzing replies. Parsing primitives ``hunt'' for certain responses, extract parameters, and take the appropriate action. The language supports robust and efficient handling of data communications. The language's interpreter is composed of front and back ends, designed as two independent processes, for saving unnecessary data movements and enhancing the interpreter's portability. The dialog manager was implemented and successfully tested in real-life situations. %A Lev Feldkhun %T Integrated Network Management Schemes - A Global Perspective on the Issue %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 279-304 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The inherent problem of managing a multi-vendor network is that the management tools provided by each vendor are disparate and not integrated. .sp 0.4 An enormous technical and intellectual effort is presently being committed by leading companies within the computer and telecommunications industry and by national and international standardization organizations to develop Integrated Network Management Systems (INMS). At the same time, the evolving INMS technology is still in its infant stage. An indication of this is the plurality of definitions and visions of Integrated Network Management. A stable reference base for INMS is needed, both by vendors to drive their R & D efforts, and by customers to help them strategize their product and services selection process. .sp 0.4 This paper discusses the overall INMS problem/solution spaces from a global perspective. A model is presented to describe a typical non-integrated network management environment in terms of the real network objects managed, the functional scope achieved through diverse and incompatible management tools, and the user interfaces employed. .sp 0.4 The main objectives of an Integrated Network Management System are then formulated and a generic model of an INM system is introduced. This model focuses on levels of integration (both from user and from internal system perspectives) and on segmentation of INMS systems into modular architectural components. .sp 0.4 The approaches of emerging INM systems and standards are then contrasted in terms of the INMS model. Finally, certain open issues are identified as per the scope and the overall methodology of developing standards in the area of Integrated Network Management. %A E. Gregori %A C. Menchi %T The Design of an OSI Compatible Network Management %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 305-314 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The emerging OSI management standards will ease the development of heterogeneous management application process, i.e. those processes (human and/or computer programs) required to manage an OSI compatible network. .sp 0.4 This paper defines a management architecture and related application processes. Sections 1 and 2 describe functions required for network management and organization principles respectively, section 3 defines the management software architecture and the relationship between the software modules and the OSI management Standards; section 4 contains considerations on the usefulness of the OSI management standards. %A D. Khakhar %A L. Fernebro %A S. Carlsson %T ESDC - An Expert System for Data Communication Design %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 315-322 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The application of expert systems technology to data communication design is investigated. A structure identifying key components of the data communication design process and expert systems implementation is presented. A prototype has been constructed in order to evaluate possibilities and limitations of an expert system approach to data communication design. %A H-K. Lam %A M. Fragano %A G. Maymon %A R. Yamarone %T Unified Primary-Operations-State Model for Objects in the Network-Element Database %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 323-332 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper presents a unified state model for object entities (subscriber lines, channel units, facilities, etc.) which are supported by network element (NE) databases. Previously, operations states were designated on a per-operations function basis, i.e., each operations technology are (e.g., memory administration, maintenance, and test access) had its own state model without a well understood coupling to the other operations areas. A hierarchical primary operations state model (Figure 3) is presented. The primary operations states are identified along with the key operations commands (Table 2) that can affect the primary operations states. The combination of the current state of an object entity and a received input command (or input event) will yield the next state for the object entity and the output response to the given input. These interactions are formally defined in a state table (Table 1). The primary operations state model will provide the means to resolve the contention for access to an object entity by memory administration, testing, and maintenance operations during the actual working operation of an NE. %A J. Goldman %A P. Hong %A J. Jeromnimon %A G. Louit %A J. Min %A P. Sen %T Integrated Fault Management in Interconnected Networks %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 333-344 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %A Z. Wang %T Model of Network Faults %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 345-352 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Network fault isolation and diagnosis are important issues in network management. However, the complexity of network faults makes the problems difficult to handle. This paper examines the environment in which network faults occur and presents a model of network faults. It then investigates Fault Propagation, one of the most important characteristics of network faults. Various aspects related to network faults are discussed as well. %A M. Antonellini %A L. Sebastiani %T Error Rates: A Convenient Technique for Triggering Fault Management Procedures %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 353-364 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper shows how local error rate monitoring may be conveniently used to trigger the execution of Network Management Procedures, using the event notification mechanism provided by the OSI Management. %A L. Feldkhun %A J. Erickson %T Event Management as a Common Functional Area of Open Systems Management %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 365-378 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Event Management has been identified as a Common Management Functional Area (CMFA) of the OSI Management Architecture. As yet, the OSI Management standardization efforts have not fully defined the Event Management CMFA. This paper proposes model for Event Management, and explores the related concepts and functions. %A Phillip H. Enslow, Jr. %T Management of High Speed and Large Networks %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 383-386 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The effective management of large, high-speed networks requires different approaches and different philosophies of management than those employed by small and medium sized networks. Major problems are encountered when an attempt is made to design a management system for a large network by simply ``scaling up'' the management techniques and operations employed with a small network. An example of such a change in philosophy is the emphasis that should be placed on the ``survival'' of the interconnection subnetwork -- for smaller networks, the management philosophy might be to continue to provide some form of service to all of the users of the network; on the other hand, for a large, high-speed network with a large number of users, the highest emphasis of periods of trouble must be placed on survival of the subnetwork even at the cost of sudden and arbitrary termination of the services made available to subscribers. The rational for this difference in short-term goals is the difficulty in initializing and restarting a large network once it has been allowed to totally collapse. Other differences between ``large-network'' and open ``small-network'' management lie in the type and quantity of network management data that can or should be collected and analyzed; in the degree of integration appropriate for separate management subsystems (it might be noted that the telephone companies have over twenty-five management subsystems in use throughout their system and there is little motivation to attempt to integrate them at this time); and other, more subtle, factors such as the roles of analysis of data collected and the value of prediction. %A S.A. Owen %T An End-User Network Management Model for ISDN %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 387-396 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP A communication link subsystem model is presented and related to the Integrated Services Digital Network (ISDN). The management aspects of the different segments of the ISDN are discussed in their relationship to end-to-end network management. %A K.-B. Sy %A P.N. Turcu %T Network Management of ISDN Customer Premises Equipment %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 397-408 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper describes a proposal to manage ISDN customers premises equipment (CPE). Three basic elements are described: management data required to be kept in the CPE, functions that are needed to manage the CPE, and the format and protocol based on the OSI. %A S. Pileri %A R. Saracco %T Towards the Integrated Network Management of the Italian Telecommunication Network %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 409-422 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %A M. J. Johnson %T Performance Issues in Management of the Space Station Information System %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 423-434 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The onboard segment of the Space Station Information System (SSIS), called the Data Management System (DMS), will consist of a Fiber Distributed Data Interface (FDDI) token ring network. We analyze performance of the DMS in scenarios representing two types of network management. In the first scenario we examine how the transmission of routine management messages impacts performance of the DMS. In the second scenario we examine techniques for ensuring low latency of real-time control messages in an emergency situation. [\fIEd. Note:\fP Also available as a RIACS Technical Report, see October '88 bibliography in \fICCR\fP]. %A D. Wilson %T A Proposed Approach for Integrated Network Management in the Army Tactical Command and Control System %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 435-444 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper briefly describes the emerging Army Tactical Command and Control System (ATCCS), focusing on the challenges the U.S. Army Signal Corps must face in its role as the network manager. It concludes with a description of a proposed means to provide an automated, integrated, and distributed network management system that is responsive to the requirements of ATCCS and, ultimately, the battlefield commander. %A Amatzia Ben-Artzi %T Architecture for a Multi-Vendor Network Management System %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 445-454 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Over the past few years, network management has increasingly become an important topic for networks. At the same time, multi-vendor operability, and with it the use of standards, has become more and more necessary. In spite of this growing need for standards and network management, however, a single standard has not emerged. .sp 0.4 This paper discusses the three main standards that are emerging today, and an architecture that provides a single, comprehensive network management system for the benefit of network users. %A E. Carrieri %A A. Fioretti %A C.A. Rocchini %A M. Cole %T OSI Compatible Architecture and Management for Integrated Ultrawideband Multichannel Networks %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 455-466 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The ever-growing requirement for more sophisticated services introduces the need of new networks with increasing capacity. It is expected that in a few years the need for Gbit/s networks will arise [1]. It is therefore important to study and define new architectures and the related management scheme for ultrawideband networks. A suitable way to achieve very high capacity is to use coherent optic and multichannel techniques. This paper presents an OSI compatible architecture and a management scheme for multichannel networks. A simulation model of this scheme has been realized by using the SIMULA language. The aim of the model was to validate the architectural scheme in order to achieve better network performance and full OSI compatibility. The final architecture and management scheme that have been achieved through the simulation work are discussed. %A J. Pimentel %T Fieldbus Network Management: Requirements and Architectures %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 467-478 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Fieldbus networks are intended to interconnect sensors, actuators, controllers, and similar devices in order to support message exchange involving reference, measurement, control, alarm, and other variables. Fieldbus networks typically operate in harsh environments, support highly automated industrial system, are of low cost and have limited processing and data storage capabilities, and they need to be initially configured and later reconfigured. Accordingly, network management for fieldbus networks have special requirements and architectures when compared to other manufacturing networks or telecommunication networks. Fieldbus network management requirements are discussed in terms of end user applications and underlying communication support. Two network management architectures are analyzed and a third one based on the notion of MMS transactions is suggested. It is proposed that a specific network management operation be added to MMS to meet all requirements. The proposed architecture is simple, requires no explicit network management protocol since it could work using the MMS protocol, and provides great flexibility for defining management domains in terms of network management objects. %A W. Morrison %T Ethernet LAN Management: NMCC/VAX ETHERnim, a Case Study %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 479-490 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP NMCC/VAX ETHERnim is a network maintenance application program for managing Ethernet local area networks. It provides the network manager with the ability to monitor and test communications, and to ensure that communications between nodes is functioning. This paper describes why it was developed, how it was designed, and how its functions might be assimilated into an integrated network management system. %A A. Mouttham %A M. Frontini %A J. Griffin %A S. Lewis %T LAN Management Using Expert Systems %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 491-504 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The data network management task may be made easier by the use of expert systems in the area of configuration, monitoring and diagnosis, for instance. The systems we developed interface to an object-oriented management information base which contain a hierarchical representation of the network and its components. This paper specifically focuses on centralized LAN configuration and diagnosis. %A Morris Sloman %A Johnathan Moffett %T Domain Management for Distributed Systems %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 505-518 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Large distributed systems are characterised by large numbers of diverse components, multiple autonomous managements. A means of grouping components into units to which a common management policy can apply is required in order to cope with the complexity of scale. .sp 0.4 The informal concept of domain is in common use as a means of describing groups of objects subject to a single management policy. The paper develops and refine the concept. We discuss relationships between domains and who these can be used to partition and structure the mangement of large distributed systems. The necessary operation on domains themselves are outlined. %A J.I. Davies %T An Analysis of Requirements for the Management of Distributed Systems %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 519-530 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper attempts to contribute to the advancement of Distributed Systems Management (DSM) by bringing together experience and requirements derived from current practice, recognising that distributed systems are now being built and used to serve genuine applications. It presents an analysis of management requirements beginning with general organisational characteristics and a number of examples drawn from a variety of application sectors. The objectives are to provide reference material and checklists to support both more specific analysis of particular management problems and wider general analysis. It represents a current view of how to analyse the requirements for DSM - transmitting experience and providing tools for further analysis. %A J. Hall %A D.M. Turnbull %T Accounting and Resource Management in an Internet Distributed Operating System %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 531-540 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Resource allocation and accounting are functions required by network management. When several management domains are joined together in an internet, facilities must be provided to allow these functions to be performed throughout the internet. This paper reports on a project concerned with extending management functions from a local network to an internet environment and shows how resource allocation and accounting can be scaled to the internet by using the concept of the market place and by providing an infrastructure that not only includes an accounting mechanism but also enables clients to find out about available services and how to locate them. %A S. Goldsmith %T Enterprise Network Management %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 541-554 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP Enterprise Network Management(ENM) is an extension of Integrated Network Management that serves as a framework for providing and maintaining an Enterprise network. The Enterprise network is a concept for an integrated communications network supporting voice, data, video, and messaging technologies, that is designed according to corporate priorities. ENM employs a set of tools that provides functional integration across all network management domains. The result, if managed properly, is an efficient, reliable, and productive network that has the flexibility to evolve according to the requirements and priorities of the corporation. The Enterprise Network, made possible by Enterprise Network Management, is a corporate strategic asset. It provides a competitive advantage, allowing corporate divisions to increase revenues and reduce overall costs. %A P. Homan %A B. Malizia %A E. Reisner %T Tandem's Distributed Systems Management %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 555-564 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %A K. Klemba %T OpenView's Architectural Models %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 565-572 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper introduces two models developed as part of Helwett-Packard's OpenView Network Management (NM) architecture: the Organizational Model and the Operational Model. Both models are potentially valuable tools in designing any NM product. They provide both a high-level view for initial planning, as well as a detailed view for implementation. %A G. Bonn %A U. Buegel %A F. Kaiser %A Th. Uslaender %T The Management of an Open, Distributed and Dependable Computing System - Delta-4 %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 573-584 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP As part of the European Strategic Programme for Research in Information Technology (ESPRIT), the Delta-4 project is seeking to define an open, distributed an dependable real-time computing architecture. The approach is based on the user-transparent replication of software components on distinct host computers. A kernel component is the integrated management system. Its design is primarily dictated by the interdependencies between the communications management, which is aligned with emerging standards, and the management of distributed applications. %A D. Holden %T Predictive Languages for Management %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 585-598 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The difference between a system manager and someone who maintains software is not great in small computer system; while in large system, there are often a large number of people needed to support the system. With the advent of large distributed systems created by interconnecting large numbers of samll systems, the problems of managing the large system are being forced onto the small system managers. This paper highlights the issues, and shows how extensible languages, and the use of predictive methods can overcome the problems in the field of distributed system management. %A Rodney Goodman %A John Miller %A Padhraic Smuth %A Hayes Latin %T Real Time Autonomous Expert Systems in Network Management %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 599-626 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP In this paper we describe our approach to developing real time expert system for integrated network management. Such systems are becoming urgently needed as the complexity of networks and the pace of new technology increases. Real time expert systems are very demanding and require interfaces to existing databases and sense and status systems. There are thus significant integration issues in their design. In the paper we first introduce expert system technology, and then develop the requirements for a real time system in the network management domain. We then describe the integrated approach to network management being developed by Pacific Bell, and show how expert systems fit into this model. Finally we describe a prototype real time expert system, NETREX, that is aimed at automatically maintaining Pacific Bell internal data networks. In particular, we describe a prototype system which implements one aspect of NETREX, a trouble ticket expert. The prototype scans open (active)trouble tickets in real time and then proceeds to collect sense and status information from the network environment. A diagnosis is made, annotations are made to the ticket, and a repair is attempted if possible. %A C. Joseph %A J. Kindrick %A K. Muralidhar %A C. So %A T. Toth-Fejel %T MAP Fault Management Expert System %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 627-636 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP The MAP Fault Management system (MAPFaM) is an evolving expert system for fault management of MAP and other OSI networks. This paper describes the design requirements for MAPFaM, including its architecture, data model, reasoning modalities, and knowledge base. As background, the paper also describes fundamentals of ISO network management, and provides a brief survey of other work done in this area. %A T. Yamahira %A Y. Kiriha %A S. Sakata %T Unified Fault Management Scheme for a Network Troubleshooting Expert System %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 637-646 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %A A. Patel %A G. McDermott %A C. Mulvihill %T Integrating Network Management and Artificial Intelligence %J Integrated Network Management: Proc. of the IFIP TC 6/WG 6.6 Symp. on Integrated Network Management %P 647-662 %C Boston, Massachusetts %I North-Holland %D May 14-17, 1989 %X \fBAbstract:\fP This paper discusses the application of AI techniques to Network Management. Current techniques are evaluated. Difficulties are assessed. The impact of new research in AI is considered. Future trends are indicated. %A J. Magee %A J. Kramer %A M. Sloman %T Constructing Distributed Systems in Conic %P 663-676 %J IEEE Trans. on Software Engineering %V 15 %N 6 %D June 1989 %A R.A. Finkel %A M.L. Scott %A Y. Artsy %A H.-Y. Chang %T Experience with Charlotte: Simplicity and Function in a Distributed Operating System %P 676-685 %J IEEE Trans. on Software Engineering %V 15 %N 6 %D June 1989 %A A.R. Tripathi %T An Overview of the Nexus Distributed Operating System Design %P 686-695 %J IEEE Trans. on Software Engineering %V 15 %N 6 %D June 1989 %A M.-Y. Lai %A K. Wilkinson %A V. Lanin %T On Distributing JASMIN's Optimistic Multiversioning Page Manager %P 696-704 %J IEEE Trans. on Software Engineering %V 15 %N 6 %D June 1989 %A A. Kumar %A M. Stonebraker %T Performance Considerations for an Operating System Transaction Manager %P 705-714 %J IEEE Trans. on Software Engineering %V 15 %N 6 %D June 1989 %A K.H. Kim %T An Approach to Experimental Evaluation of Real-Time Fault-Tolerant Distributed Computing Schemes %P 715-725 %J IEEE Trans. on Software Engineering %V 15 %N 6 %D June 1989 %A B. Bhargava %A J. Riedl %T The Raid Distributed Database System %P 726-736 %J IEEE Trans. on Software Engineering %V 15 %N 6 %D June 1989 %A C.T. Yu %A K.-C. Guh %A D. Brill %A A.L.P. Chen %T Partition Strategy for Distributed Query Processing in Fast Local Networks %P 780-793 %J IEEE Trans. on Software Engineering %V 15 %N 6 %D June 1989 %A Brian Frederick Hanks %T Multiprocessing Using Workstations with A Shared File System %R UCSC-CRL-88-12 %D September, 1987 %I University California, Santa Cruz %X \fBAbstract:\fP This thesis describes a method for exploiting large grain parallelism on a network of workstations. Parallelism is provided by a collection of C callable library functions and macro expansions, giving the applications programmer the ability to create, execute, and synchronize processes using synchronous message passing. A program using these library calls consists of a main process and a set of created subprocesses executing on multiple workstations in the network. Significant speedup has been obtained with this library. [\fIEd. Note\fP: Available for $4 from Technical Report Librarian Baskin Center for Computer Engineering & Information Sciences, Applied Sciences Building, University of California, Santa Cruz, CA 95064] %A Dragomir D. Dimitrijevic %A Mon-Song Chen %T Dynamic State Exploration in Quantitative Protocol Analysis %J Proc. 9th Intl. Symp. on Protocol Specification, Testing and Verification (IFIP WG 6.1) %D June 6-9 %C Enschede, The Netherlands %X \fBAbstract:\fP This paper extends the dynamic state exploration technique used in an integrated probabilistic protocol verification and evaluation procedure [7]. The extension reduces the complexity from \fIO( n \s-2\u2m\d\s+2)\fP to \fIO(n \s-2\u3\d\s0)\fP, where n and m are the numbers of generated states and explored transitions, respectively. Additional properties of the technique are also analyzed to further enhance the verification and evaluation procedure. The procedure based on this technique is unique in that (1) it evaluates the importance of states in the course of global reachability graph (GRG) generation, (2) it, based on the importance of states, explores only the most interesting subset of states, and (3) it computes important reliability and performance measures such as Mean Time To Unknown and Confidence Level, which is the probability sum of checked scenarios, and uses them as the stopping criteria. The procedure is demonstrated using the call establishment phase of X.75 protocol. .sp 0.4 [7] Dimitrijevic, D. D. and M-S. Chen, ``An Integrated Algorithm for Probabilistic Protocol Verification and Evaluation,'' Proc. INFOCOM'89 %A Gerard J. Holzmann %A Joanna Patti %T Validating SDL Specifications: an Experiment %J Proc. 9th Intl. Symp. on Protocol Specification, Testing and Verification (IFIP WG 6.1) %D June 6-9 %C Enschede, The Netherlands %X \fBAbstract:\fP This paper describes a method for validating specifications written in the CCITT language SDL. The method has been implemented as part of an experimental validation system. With the experimental system we have been able to perform exhaustive analyses of systems with over 250 million reachable composite system states. The practicality of the tool for the analysis of substantial portions of AT&T's 5ESS\(rg Switch code is now being studied. %A Marc Phalippou %T Functional specification for an ISDN switching system: an experience using Estelle %J Proc. 9th Intl. Symp. on Protocol Specification, Testing and Verification (IFIP WG 6.1) %D June 6-9 %C Enschede, The Netherlands %X \fBAbstract:\fP FDTs are considered as potential candidates for the specification of the functional architecture of an ISDN switching system. This particular application requires specific possibilities from the language: the top-down approach which is used for designing such a complex system implies the use of an abstract and flexible view of data, and needs some structuring facilities for the stepwise refinement process of the program. We show here how both problems, in spite of their apparent difficulty, can be elegantly solved in Estelle using dynamic features. %A Wendy Y. L. Chan %A Son T. Vuong %A M. Robert Ito %T On Test Sequence Generation for Protocols %J Proc. 9th Intl. Symp. on Protocol Specification, Testing and Verification (IFIP WG 6.1) %D June 6-9 %C Enschede, The Netherlands %X \fBAbstract:\fP This paper presents preliminary findings obtained from an ongoing study on the generation of conformance test sequences for communication protocols. A method for generating test sequences for the control structure of protocols is first presented, followed by the proposal of a hybrid technique to generate test sequences that separately cover both the control and data flow aspects of protocols. The control structure of a protocol is assumed to be modelled by a finite state machine (FSM). The goal is to separately provide state-identification capabilities to those input/output pairs (I/Os) that lack them. During testing, each I/O is thus capable of assuming dual roles: as a test for the transition that it occurs in and as a means of verifying the correctness of the state transition that occurred during the previous transition test. A minimal tour that traverses each I/O pair at least once is now capable of detecting both transition and transfer errors. The concept of unique input/output (UIO) sequences previously developed for another test generation procedure is applicable here. This method is combined with a data flow oriented method, called static data flow analysis, to produce test sequences which test the control and data flow aspects of protocols separately. The resulting hybrid technique is directly applicable to protocols specified in Estelle. %A D. Katz %T A Proposed Standard for the Transmission of IP Datagrams over FDDI Networks; RFC 1103 %J Internet Request for Comments %N 1103 %D June 1989 %I Network Information Center, SRI International %C Menlo Park, CA %X \fBAbstract:\fP This RFC specifies a method of encapsulating the Internet Protocol (IP) datagrams and Address Resolution Protocol (ARP) requests and replies on Fiber Distributed Data Interface (FDDI) Networks. This RFC specifies a proposed protocol standard for the Internet community. %A I. Svobodova %A P. Janson %A F. Mumprecht %T OSI in Heterogeneous Environments %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A J. Radigan %A B. Meandzija %T Y.A.N.C. (Yet Another Network Compiler) %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A T. Nakakawaji %A K. Katsuyama %A N. Miyauchi %A T. Mizuno %T Development and Evaluation of APRICOT (Tools for Abstract Syntax Notation One) %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A Y. Ohara %A T. Suganuma %A S. Senda %T ASN.1 Tools for Semi-automatic Implementation of OSI Application Layer Protocols %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A K.F. Lini %T Interoperability for Efficiency and Profit: the USDA OSI Implementation Strategy %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A G. Chanshan %A Z. Gendu %T An Example of OSI Standards Implemented on a LAN %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A S. Choi %A S. Park %A K. Hun %A Y. Chae %A Y. Choi %T An Implementation of OSI Protocols o Daeduk Science Town Network %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A S. Miyazaki %A H. Taniguchi %A Y. Hattori %A T. Kuwatobi %A S. Kohyama %T Implementation of Open Systems Interconnection in Communication Management Program %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A B.D. Carlos %A Y. Kobayashi %T OSI Systems Management and SNA/Management Services %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A N. Shimizu %A S. Keneko %T A Prototype BITNET-X.400 Gateway for Mail Exchange %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A A. Ben-Artzi %T Interoperable Network Management System %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A M. Tsukamoto %A Y. Sato %A H. Shitanda %A N. Yoshie %A N. Funato %A N. Tanaka %A S. Nakagome %A A. Tojo %T The Architecture of OZ: Object-oriented Open Distributed System %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A M. Tanaka %A Y. Ishizawa %T A Gateway with an Access Control Mechanism for Network Security %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A A. Iwabuchi %A R.J. Linn Jr. %A J.P. Favreau %T Application of Formal Specification Techniques to the Specification of the MHS Test System %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A J. Domon %A K. Endoh %T OSI Protocol Verification and Conformance Testing in the Nomura New On-Line System %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A G. Bonnes %T OSI-X.400 Inter-Operation Verification Services %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A T. Wakasugi %A Y. Takaoka %T Design and Evaluation of Test Suites for the OSI Transport Layer at the Conformance Test Center of INTAP %J Proc. 2nd Intl. Symp. on Interoperable Information Systems, ISIIS '88 %D November 10-11, 1988 %C Tokyo, Japan %I Elsevier Publishing %A Sharon Heatley %A Dan Stokesberry %T Analysis of Transport Measurements Over A Local Area Network %J IEEE Communications %D June 1989 %V 27 %N 6 %P 16-22 %A David D. Clark %A Van Jacobson %A John Romkey %A Howard Salwen %T An Analysis of TCP Processing Overhead %J IEEE Communications %D June 1989 %V 27 %N 6 %P 23-29 %A Per Gunnigberg %A Mats Bjorkman %A Erik Nordmark %A Stephen Pink %A Peter Sjodin %A Jan-Erik Stromquist %T Application Protocols and Performance Benchmarks %J IEEE Communications %D June 1989 %V 27 %N 6 %P 30-36 %A David R. Cheriton %A Carey L. Williamson %T VMTP as the Transport Layer for High-Performance Distributed Systems %J IEEE Communications %D June 1989 %V 27 %N 6 %P 37-44 %A Morten Skov %T Implementation of Physical and Media Access Protocols for High-Speed Networks %J IEEE Communications %D June 1989 %V 27 %N 6 %P 45-53 %A Eugene H. Spafford %T The Internet Worm: Crisis and Aftermath %J Communications of the ACM %D June 1989 %V 32 %N 6 %P 678-687 %A Jon A. Rochlis %A Mark W. Eichin %T With Microscope and Tweezers: The Worm from MIT's Perspective %J Communications of the ACM %D June 1989 %V 32 %N 6 %P 689-698 %A Donn Seeley %T Password Cracking: A Game of Wits %J Communications of the ACM %D June 1989 %V 32 %N 6 %P 700-704 %T The Worm Story %J Communications of the ACM %D June 1989 %V 32 %N 6 %P 677-710 %X \fBNote:\fP This issue of CACM focuses on the Internet worm. In addition to several articles (which are listed separately in this bibliography) several ethics statements by prominent networking groups are printed on the pages facing the first page of the articles. %A J.J. Garcia-Luna-Aceves %T Routing Management in Very Large-Scale Networks %J Future Generation Computing Systems %V 4 %N 2 %D September 1988 %P 81-93 %O \fIReviews\fP: \fIComputing Reviews\fP, Vol. 30, No. 6, June 1989. %A Craig Partridge %T How Slow is One Gigabit Per Second %R Report No. 7080 %I BBN Systems and Technologies Corporation %D June 1989 %X \fBAbstract:\fP At first blush, one would expect that increasing data network transfer rates by two orders of magnitude (from the ubiquitous 10 Mbit speed of today's LANs to the greater than 1 gigabit-per-second speeds we expect of networks in the early 1990s) would severely impact our choice of network protocols and architectures. This report presents the strawman argument that, in fact, moving to one-gigabit data rates presents surprisingly \fIfew\fP problems. [\fIEd. Note:\fP Available from the author \(em see address inside front cover of this issue]. %A H-W. Braun %T Models of Policy Based Routing; RFC 1104 %J Internet Request for Comments %N 1104 %D June 1989 %I Network Information Center, SRI International %C Menlo Park, CA %X \fBAbstract:\fP The purpose of this RFC is to outline a variety of models for policy based routing. The relative benefits of the different approaches are reviewed. Discussions and comments are explicitly encouraged to move toward the best policy based routing model that scales well within a large internetworking environment. %A K. Lougheed %A Y. Rekhter %T A Border Gateway Protocol (BGP); RFC 1105 %J Internet Request for Comments %N 1105 %D June 1989 %I Network Information Center, SRI International %C Menlo Park, CA %X \fBAbstract:\fP This RFC outlines a specific approach for the exchange of network reachability information between Autonomous Systems. .sp 0.4 At the time of this writing, the Border Gateway Protocol implementations exist for cisco routers as well as for the NSFNET Nodal Switching Systems. A public domain version for ``gated'' is currently being implemented. %A I. J. Hayes %A M. Mowbray %A G.A. Rose %T Signalling System No. 7: The Network Layer %J Proc. 9th Intl. Symp. on Protocol Specification, Testing and Verification (IFIP WG 6.1) %D June 6-9 %C Enschede, The Netherlands %X \fBAbstract:\fP The paper formally specifies message streams, routing and rerouting within CCITT's Signalling System No 7, using the Z specification language. The paper advocates the use of several related views (projections) to specify complex systems. Multiple projections provide a useful technique for separation of concerns. Three views are presented: .sp 0.4 (1) An External View \- a service view, concerned with the identification of individual correctly sequenced message streams from origin to destination without routing detail. .sp 0.4 (2) A Sectional Stream View \- a refinement of the external view which identifies the route of a stream as a sequence of sections, each a signalling point and an outgoing linkset. This view is the basis for rerouting. .sp 0.4 (3) A Network View \- a specification of overall network connectivity (topology) accommodating the totality of streams. Routing labels are added to messages to identify streams. .sp 0.4 The stream and network views are then related for consistency of connectivity (e.g., that stream routes are consistent with current link availability) and consistency of projection (the projection of the totality of messages in the network with respect to a particular routing label is consistent with the corresponding stream view). .sp 0.4 The paper does not discuss service-primitive or protocol-unit details, congestion control, or management or test procedures.