Data processing in the 1990s and beyond will be established around workgroup computing, whether the workers are in a building, on a campus, around town, around the nation, or around the world. To achieve the desired worker productivity enhancements sought through communication technology, workgroups will increasingly be interconnected over an enterprise-wide high-speed backbone network. Asynchronous Transfer Mode (ATM) technology and cell relay service are likely to play a major role in enterprise networks of the mid and late 1990s. ATM technology not only supports cell relay service, but also other fastpacket services, such as frame relay service, Switched Multimegabit Data Service, circuit emulation service, LAN emulation, and video/multimedia services. It is anticipated that ATM services will transform the enterprise network_from a data-only network to an integrated data, voice, video, image, and multimedia corporate infrastructure.
Specifically, many large corporations now have a growing need to extend high-speed communications beyond key sites, to support applications such as distributed cooperative computing, business/scientific imaging, video conferencing, video distribution, multimedia, (corporate) distance learning, etc. ATM presents itself as a viable, perhaps optimal, approach to meet these evolving corporate needs. However, because of the expense of high-speed networks, and the fact that in an ATM environment, most, if not all, of the corporate information flow may be aggregated over one or a few integrated broadband links, effective network management becomes very critical.
Therefore, such transformation of the enterprise network to an ATM-based infrastructure will result in an ever-increasing need for optimal
management of information processing and communication resources and services, whether these services are locally based or wide-area-based. Users expect rapid reconfiguration of the services they are using, to help them meet market demands on the products they produce, since many of these products utilize, in one way or another, communication and computers. Corporate financial monitors expect cost-effective communication, low outage time, and few work disruptions. Communication and computing expenditures can represent as much as 10% (5% on the average) of the total revenue of a corporation. This can equate to tens of millions of dollars a year. Communications managers want tools to facilitate planning, monitoring, operation, maintenance, administration, and reconfiguration of the enterprise network, to satisfy their customer base.
The purpose of this book is to familiarize ATM and network planners in the user community with the network management challenges, opportunities, and capabilities that ATM presents. The technology is just beginning to be deployed, but as soon as the equipment is taken out of the box and powered up, the question immediately presents itself to the user: "How do we manage _ in a cost-effective manner _ this thing?" A network technology without a mechanism for the user (and the provider, for that matter) to manage it, is like a car without a steering wheel _ of limited or no value.
Detrimentally, there has been a feast of hyperbolae about ATM during the past couple of years. Some in the trade press imply that ATM is "magical" in its support of bandwidth on demand, scalability, and other features. These people attribute to ATM features that have never existed and never will exist. The definition used by many neophytes_that ATM will fix whatever significant networking problems you have today, be they staffing; complexity; inability to reach remote exurban, rural, or third-world locations; expenditures; the need to keep up with technology, the need for analytical design; the need for tight operational discipline; or security, is absolutely positively wrong. This book seeks to expunge these misleading accretions. They have no place in a rational discourse of what can be done to next-generation corporate networks.
So what is good about ATM? It is not a panacea. it is only the best technology to come along so far. It can be put to effective use in meeting high-speed connectivity requirements, particularly at the backbone/WAN level, where the tributary bandwidth requirements have been aggregated well beyond the bandwidth needed by a single desktop user. For example, on the assumption that 1.5 Mbps dedicated is adequate to support (e.g., by using MPEG-1) video to each desktop, a company with 100 desktops may see an aggregate requirement of 155 Mbps (if all users must access a remote video server containing some information of interest). As a straight connectivity technology at the WAN level, ATM presents itself as a good candidate, in this example, to support this bandwidth need.
Some also look to ATM to support legacy LAN requirements at the premises level (we don't mean here new ATM services to the desktop, but a hybrid of a premises-based ATM-made-to-look-like-Ethernet suite of transitional equipment). This may or may not occur. In the past, new technologies have failed to win over the legacy support market. For example, when satellite links for data applications first appeared, they were not able to support mainframe links; a whole new application (VSAT) had to emerge. When ISDN standards were published (e.g., Q.931), three-quarters of the protocol state machine (as measured by the thickness of the standard) was dedicated to the interworking of traditional packet services over the D-channel (only one-quarter of the standard related to supporting circuit-switched services). This legacy support never occured in ISDN. Similarly, no frame relay service now available is offered in conjunction with ISDN as originally conceived. Numerous other examples can be cited. Naturally, it is the market (but not the vendors) who will decide.
After debunking the hype, users will deploy ATM services for those (new) applications that really need it. Several studies have shown that network management can take as much as 40% of the total communication budget. To ensure that the network management methodology, and the tools that may be selected by an organization, provide for efficient network operations, several issues need to be addressed:
These and other issues are treated in this text. We hope it will assist corporations early on, in making the necessary decisions about ATM. The text takes a pragmatic approach. It is aimed directly at communication professionals and corporate network managers and planners.
After an overview of the field (Chapter 1), an assessment of the importance of ATM in enterprise networks is provided (Chapter 2). This is followed by a primer on key ATM technologies that will play significant roles in the near future and which corporations expect their communication planners to manage (Chapter 3).
Chapter 4 provides a detailed technical analysis of ATM network management capabilities. Chapter 5 discusses the management of the protocol stack at the physical and ATM level. Chapter 6 looks at models to support Customer Network Management of ATM services. Chapters 7,8, and 9 cover the Open Systems Interconnection Network Management Specific Management Functional Areas (fault management, accounting management, configuration management, performance management, and security management) in the ATM context. Chapter 10 covers some of the network management issues faced by carriers offering ATM services, as they contemplate managing their own ATM networks.
The chapters that follow look at specific user environments, and assess the issue of managing ATM overlays to these embedded networking architectures _ which clearly organizations cannot obsolete overnight in favor of an all-ATM enterprise network. Chapter 11 looks at emerging computing environments and the specific management requirements these impose on ATM. Chapter 12 examines the issue of establishing corporate network management goals. Chapter 13 addresses the predecision issues related to technology assessment for the deployment of ATM. Chapter 14 assesses management issues in Virtual LANs. People-related issues are covered in Chapter 15 along with a description of other collateral tools for effective network management. Finally, Chapter 16, provides an assessment of the industry in terms of commercial tools that have emerged and are expected to emerge in the next couple of years. Typical features and functions are discussed.
In addition to its professional market target, we believe that the book can be used for an undergraduate or graduate course, particularly from an enterprise network perspective. This material has, in whole or in parts, been used for teaching ATM technology at Stevens Institute of Technology to students already familiar with fastpacket communications.
The authors hope that this book will be of value to early adopters of ATM in selecting an appropriate network management strategy that provides a path of cost-effectiveness that goes beyond technical innovation. In the experience of the authors, corporations seek technology for productivity enhancement rather than pure academic pursuit of abstract technical advancement and sophistication.