Wordtrade.com
Mathematics
Review Essays of Academic, Professional & Technical Books in the Humanities & Sciences
Mesh-based Survivable Transport Networks: Options and Strategies for Optical,
MPLS, SONET and ATM Networking by Wayne D. Grover
(Prentice Hall PTR )"Internet hindered by severed cable." "Internet chaos
continues today... " "Cable break disrupts Internet in several countries. "
"Calling and major businesses down from cable cut. "
These are
headlines arising in just one month from fiber optic cable disruptions. Despite
the enormous advantages of fiber optics and wave-division multiplexing, the
truth is that the information economy-fueled by fiber-optic capacity-is based on
a surprisingly vulnerable physical medium. Every effort can be made to protect
the relatively few thumb-sized cables on which our information society is built,
but the cable-cuts and other disruptions just don't stop. From deep-sea shark
bites to the fabled "backhoe fade," serious transmission outages are common and
of increasing impact. Some form of fast rerouting at the network level has
become essential to achieve the "always on" information networks that we depend
upon.
One way to
survive optical network damage is to duplicate every transmission path. In the
form of rings and diverse-routed protection switching schemes, this is actually
the most widespread solution in use today. Our view has long been that this is
an inefficient expedient-the "get a bigger hammer" approach to solving a
problem. Admittedly, rings filled the void when survivability issues reached
crisis proportions in the 1990s. But now network operators want options that are
just as survivable but more flexible, more growth-tolerant, able to accommodate
service differentiation, and far more efficient in the use of capacity. This is
where "sharedmesh" or "mesh-restorable" networks take the stage.
The
author's love affair with the mesh-survivability approach began in 1987-with the
naive certainty back then that sheer elegance and efficiency would suffice to
see it adopted in SONET by 1990! The actual journey has been much longer and
complicated than that. The Internet and Optical Networking had to happen first.
And all the ideas had much more development to undergo before they would be
really ready for use. Today, we understand the important values of mesh-based
survivability go beyond just efficiency (and of course you can rarely make money
off of elegance alone). Flexibility, automated provisioning, differentiated
service capabilities, and the advent of Internet-style signaling and control
are all important in making this a truly viable option. But its full
exploitation still requires new concepts and ideas about network
operation-letting the network self-organize its own logical configuration, for
example, and letting it do its own preplanning and self-audit of its current
survivability potential. New planning and design models are also required. These
are the central topics of this book-designed to stimulate and facilitate the
further evolution toward highly efficient, flexible and autonomous meshbased
survivable networks.
The book is
written with two main communities in mind. One is my colleagues in industry;
the system engineers, research scientists, technology planners, network
planners, product line managers and corporate technology strategists in the
telcos, in the vendor companies, and in corporate research labs. The are the key
people who are continually assessing the economics of new architectural options
and guiding technology and standards developments. Today these assessments of
network strategy and technology selection put as much emphasis on operational
expense reduction as on capital cost-capacity efficiency and flexibility are
both important in future networks. Network operators are in an intensely
competitive environment with prices dropping and volumes rising and success is
dependent on all forms of corporate productivity enhancement. Mesh networking
can provide fundamental productivity enhancements through greater network
efficiencies and flexibility. The book aids the operating companies in finding
these new efficiencies by giving many new options and ideas accompanied with the
"how to" information to assess and compare the benefits on their own networks.
Examples of the new directions and capabilities the book provides are in
topology evolution, ring-to-mesh conversion by "ring-mining," multiple
Quality-of-Protection design, tailoring restoration-induced packet congestion
effects in a controlled manner, simplifying dynamic demand provisioning, and so
on. An important plus is that the book also contains the first complete
treatment of the intriguing and promising new concept called "p-cycles"-offering
solutions with ring-speed and meshefficiency.
Providers
of the networking equipment, the vendors, are-as the saying goes-"fascinated by
anything that interests their boss." That means their network operating
customers. Vendors must not just keep in step with the problems, opportunities,
and thinking of their customers, but also aspire to bring their own unique
equipment design strategies to the market and to provide leadership in
development of advantageous new networking concepts. The vendor community will
therefore be especially interested in the techniques for split-second mesh
restoration and self-organizing traffic-adaptation as features for their
intelligent optical cross-connects and Gigabit routers, for instance-as well
ideas for new transport equipment such as the straddling span interface unit
that converts an existing add/drop multiplexer into a p-cycle node. All of the
other topics covered are of interest to vendors too because they enhance their
ability to assist customers with network planning studies as part of the
customer engagement and sales process.
Developers
of network modeling, simulation and planning software will also be interested in
many ideas in the book. By incorporating capabilities to design all types of
architecture alternatives or, for example, to simulate dynamic provisioning
operations in a protected working capacity envelope, or to model the incremental
evolution of a survivable capacity design in the face of uncertain demand, or to
support ring-mining evolutionary strategies-these suppliers enable their
customers to pursue a host of interesting new "what if' planning studies.
The second
main community for whom the book is intended is that of graduate-level teaching
and research and new transport networking engineers who want a self-contained
volume to get bootstrapped into the world of transport networking planning or
to pursue thesis-oriented research work. A principle throughout has been to
draw directly on my experience since 1992 at the University of Alberta of
teaching graduate students about survivable transport networking. This allowed
me to apply the test: "What needs to be included so that my graduate students
would be empowered both to do advanced investigations in the area, but also to
be knowledgeable in general about transport networking?" On one hand, I want
these students to be able to defend a Ph.D. dissertation, but on the other hand
also to have enough general awareness of the technology and the field to engage
in discussion with working engineers in the field. This is really the reason the
book has two parts.
The test of
needed background has guided the definition of Chapters 1 to 4 which are called
the "Preparatory" chapters. These chapters cover a lot of generally useful
ground on IP and optical technology, routing algorithms, graph theory, network
flow problems and optimization. Their aim is to provide a student or new
engineer with tools to use, and an introductory understanding of issues, trends,
and concepts that are unique to transport networking. In my experience, students
may have done good theoretical research, but at their dissertations a committee
member may still stump them with a down-to-earth question like "How often do
cables actually get cut?" "Is this just for SONET or does it work for DWDM
too?", "How does restorability affect the availability of the network?" or
(perennially it seems) "...but I don't see where are you rerouting each phone
call or packet." These few examples are meant just to convey my philosophy that
as engineers we should know not only the theory, the mathematical methods, and
so on, to pursue our "neat ideas" but we also need to know about the technology
and the real-world backdrop to the research or planning context. This makes for
the best-prepared graduate students and it transfers to the training of new
engineers in a company so that they are prepared to participate and contribute
right away in all discussions within the network planning group he or she joins.
An engineer who can, for example, link the mathematics of availability analysis
to a contentious, costly, and nitty-gritty issue such as how deep do cables have
to be buried, is exactly the kind of valuable person this book aims to help
prepare. Someone who can optimize a survivable capacity envelope for mixed
dynamic services, but is also savvy enough to stay out of the fruitless "50 ms
debate" is another conceptual example of the complimentary forms of training and
knowledge the book aspires to provide.
The book is
ultimately a network planners or technology strategists view of the networking
ideas that are treated. It employs well-grounded theoretical and mathematical
methods, but those are not the end in itself. The book is also not filled with
theorems and proofs. The emphasis is on the network architectures, strategies
and ideas and the benefits they may provide, not primarily on the computational
theory of solving the related problems in the fastest possible way. Our
philosophy is that if the networking ideas or science look promising, then the
efforts on computational enhancement are justified and can follow. Fundamental
questions and ideas about networks, and network architecture, (which is the main
priority in my group) stand on their own, not to be confused with questions and
ideas about algorithms and solution techniques to solve the related problems as
fast as possible (others are stronger in that task). Obviously work in this area
involves us in both networking science and computation, but the logical
distinction is important-and often seems lost in the academic literature. The
book is also not a compendium or survey of previously published papers. While
suitably referenced, its content is unabashedly dominated by the author's own
explorations and contains a large amount of previously unpublished material.
Although setting the context in terms of modem transport technologies (WDM,
SONET, ATM, IP, MPLS) our basic treatment of the networking ideas and related
planning problems is in a generic logical framework. The generic models can be
easily adapted for to any specific technologies, capacities, costs, or signaling
protocols, etc. The book thus provides a working engineer or a new researcher
with a comprehensive, theoretically based, reference book of basic
architectural concepts, design methods and network strategy options to be
applied on mesh-survivable networks now and in the future.
A few words
about the flow of the book. The Introduction gives a much fuller roadmap of the
content and novelty in each chapter. Briefly, however, Chapter 1 is an
orientation to transport networking. Chapter 2 is devoted to background on IP
and DWDM optical networking developments, as the technological backdrop. Part of
Chapter 3 is partly just "interesting reading" on the effects of failures and
the range of known schemes and techniques to counteract or avoid failures. The
rest of Chapter 3 includes a more technical "sorting out" of the "-ilities":
availability, reliability, network reliability, restorability, and other
measures. Chapter 4 is then devoted to graph theory, routing algorithms and
optimization theory and techniques but only as these topics specifically relate
to transport network problems. Chapter 5 starts the second part of the book on
more advanced studies and applications with an in-depth treatment of span
protection and restoration. It has its counterpart devoted to path restoration
in Chapter 6. Chapter 5 considerably "updates" the thinking about span-oriented
survivability in optical networks with dynamic traffic.
If the book
was a musical score, Chapters 7 through 11 would be the "variations." Each
chapter treats a more advanced topic or idea selected by the author because of
its perceived usefulness or possible influence on the direction of further
research and development. These are some of the author's "shiny pebbles" (in the
earnestly humble sense of Newton). Chapter 7 recognizes an important
difference-and opportunity-in cell or packet-based transport: that of controlled
oversubscription of capacity upon restoration. This is a unique advantage for
MPLS/ IP-based transport survivability. Chapter 8 is devoted to all aspects of
dual-failure considerations in mesh restorable networks. An especially
interesting finding is that with a "first-failure protection, second-failure
restoration" concept, higher than I+ 1 availability can be achieved for premium
service paths at essentially no extra cost. Chapter 9 treats the challenging,
and so far almost unaddressed, problem of optimizing or evolving the basic
facility-route (physical layer) topology for a mesh-restorable network. Chapter
10 explains the new (and to us, very exciting) concept of p-cycles, which are
rooted in the idea of pre-configuration of mesh spare capacity.
p-Cycles
are, in a sense, so simple, and yet they combine the fast switching of ring
networks with the capacity-efficiency of mesh-based networks. We include
p-cycles as a meshbased survivable architecture because they exhibit extremely
low mesh-like capacity redundancy and because demands are routed via shortest
paths over the entire facilities graph. They are admittedly, however, a rather
unique form of protection scheme in their own right in that lies in many regards
in-between rings and mesh. Candidly, I venture that many colleagues who went
through the decade-long ring-versus-mesh "religious wars" of the 1990s would
understand when I say that p-cycles call for that forehead-bumping gesture of
sudden realization-this solution (which combines ring and mesh) was unseen for
the whole decade-long duration of this debate! As of this writing the author
knows several research groups that are shifting direction to work on p-cycles as
well as a half-dozen key industry players looking closely at the concept.
Chapter 11
on ring-mesh hybrids and ring to mesh evolution is placed at the end. The logic
is that if we assume success of the prior chapters in motivating the mesh-based
option then the "problem" this creates is that many current networks are
ring-based. Its like "Ok, we believe you-but how do we get there now?" The
closing chapter therefore devotes itself to bridging the gulf between existing
ring-based networks and future mesh or p-cycle based networks by considering
the design of intermediate ring-mesh hybrid networks and "ring-mining" as a
strategy to get to a mesh future from a ring starting-point today.
The
Appendices, and other resources such as chapter supplements, a glossary, student
problems, research project ideas, network models, and more are all web-based-so
they can be continually updated and expanded in scope and usefulness. Many
directly usable tools and resources are provided for work in the area of
mesh-survivable networking. This includes AMPL models and programs to permit
independent further study of most of the planning strategies presented, plus
Powerpoint lectures on a selection of topics, technical reports, and additional
references and discussions. The aim has been to create a highly useful and
hopefully interesting book that is laden with new options, ideas, insights and
methods for industry and academia to enjoy and benefit from.
Third Generation CDMA Systems for Enhanced Data Services by Giridhar D.
Mandyam, Jersey Lai (Communications, Networking and Multimedia: Academic
Press) First book to cover both of the leading CDMA standards and it provides an
authoritative, current review of the newest third generation technologies. Ideal
for engineers developing wireless devices, as well as for those in the service
sectors.
Until now, most wireless devices have been used for voice transmission, but the
new third generation of wireless devices promise greatly enhanced delivery of
both voice and data communications. These devices are currently under
development and are being built using the two major CDMA (code division multiple
access) systems, cdma2000 and wideband CDMA. These telecommunication systems use
signal codes to receive voice and data information. This authoritative new book
reviews both of these systems and deals with the challenges engineers face in
bringing these next generation devices to market. The authors work at Nokia, one
of the world's leading companies involved with the design, development, and
manufacture of wireless telephones and devices. They are closely involved within
Nokia and the industry in working with CDMA standardization to bring this
technology to the consumer. This is the first book to cover both of the leading
CDMA standards (cdma2000 and wideband CDMA), and it provides an authoritative,
current review of the newest third generation technologies. This book is ideal
for the engineers developing wireless devices (at Nokia, Ericsson, Qualcomm,
Motorola, etc.), as well as for those in the service sectors (i.e. AT&T,
Verizon, Voicestream, Quest, etc.).
Headline 3
insert content here