Somewhere between 1993 and 1995, the Internet started entering the everyday lives of the general public. As the number of websites grew rapidly in those early years, several different "Internet Yellow Pages" books where published to catalog the hundreds of new cyber-destinations to which to travel. These books appear to modern "surfers" more like fictional parodies than the serious reference works they were then. In today's Google/Yahoo world, the idea of going to a book to look up a website or newsgroup sounds as archaic as requiring a week to cross the Atlantic.
Nevertheless, in those bygone times of more than 10 years ago, a technology still associated with the future today was being forged. In December 1995, RFC 1883 presented one of the first introductions to a new internetworking protocol: IPv6.
Perhaps the architects of the Internet suspected a revolution was coming and the infrastructure that had served the networks of research labs and universities well for some 15 years wasn't going to be up to the task. Surprisingly IPv4, the fifth version of the defining protocol upon which the modern Internet is based, has held up an additional 10-plus years, but, change is in the air. Those interested in the predecessors of IPv4 can find a list of references for where to read about IPv0 (yes, there was a version 0) through IPv3 in RFC 762. The references are historical now, but Vint Cerf's and Jon Postel's early papers on IP are required reading for anyone curious about the how the Internet evolved.
Meet IPv6
Internet Protocol, version 6, or IPv6 for short, is the next step in the evolution of internetworking protocols. These are the packet-based network-layer core protocols responsible for moving information from one place to another on all IP-based networks, the largest and most popular such network being the Internet itself. No matter whether you are navigating a web page, sending e-mail, or downloading music, IP is the protocol that's moving the data.
From roughly 1980 to the present day, the fourth version of IP, IPv4, has been performing the heavy lifting. Conceived when there were no more than a few dozen interconnected systems, which were all cooperating with each other, IPv4 has been stretched far beyond what its designers could ever have hoped. Enter IPv6.
The next chapter seeks to demystify and explain some of the basics of IPv6 so that it can be looked at for what it is without all the Fear, Uncertainty, and Doubt (FUD), just another technology with good points and bad ones. After that, the IPv6 landscape of usage, deployment rates, and future predictions, as it exists today, is described. None of these topics is the main purpose of this book. After the modest IPv6 primer, including lists of references from which to gather the details, the rest of the book addresses one of the least understood and most feared aspects of IPv6: how to make the transition to it from the current networking technologies.
Making the Transition from IPv4 to IPv6
In 1980, it was pretty easy to upgrade from IPv4's predecessor. The total number of systems to be upgraded was less than that found nowadays in a typical small company or branch office of fewer than 100 employees. The system administrators, who often also happened to be the network programmers, all knew each other, for the most part, and the network could be upgraded in unison with minimal process or synchronization. Furthermore, if a mistake were made, the consequences were, well, inconsequential. None of these systems was critical to the world economy, national defense, or any essential communications fabrics, but, it's not 1980 any more.
Anyone in network operations or management knows that all but the most minimal of changes to mission-critical systems and infrastructure requires huge amounts of planning. A priori agreements between all stakeholders as to what to do, when to do it, and how to recover to a sane state should there be a failure along the way are par for the course. It's at worst a mild exaggeration to say that, for most large modern networks, millions (if not billions) of dollars, the defense of nations, and the abilities of first-responders to save lives are dependent on whether those networks continue to function. Getting from IPv4 to IPv6 while still keeping the lights on is what this book is all about.
IPv6 is destined to become the dominant networking protocol, but this won't happen overnight. In fact, the early days of world-wide transition, that is, right now, are fraught with peril. There will be those who leap into IPv6 with both feet, not taking the necessary precautions required with the enterprise-wide adoption of any new technology. There will be those who think they are only putting their toe in the water, only to find that they're up to their necks from a lack of planning or understanding the technology's capabilities. This book seeks to identify the risks, offer mitigation strategies for them, and generally help navigate around the pitfalls that can occur during a poorly planned or executed transition.
Looking beyond IPv6
An article published in Japan in 1996 was entitled "IPv6: The Final Frontier." Yet, for the next 10 years we barely set foot in that frontier. When we do finally become fully entrenched in IPv6, it should hardly be expected to be a final frontier by any student of the history of technology. Even so, computer networks are certainly entering a brave new world with the maturing of this latest version of IP.
We should expect to take a while to move to IPv6 and we should expect to stay on it for some time, as well. After all, IPv4 has lasted more than a quarter of a century, which is something you can be pretty sure its inventors weren't expecting in 1980. Even with its long and promising future, though, IPv6 is no more the end in networking evolution than sailing ships were in trans-Atlantic travel.
Who Should Read This Book?
This book is a practical guide through the process of upgrading complex IPv4-based networks to IPv6. The primary audience consists of network managers and the MIS/IT staff responsible for taking care of an enterprise's networks. A secondary audience includes CIOs and other senior IT managers responsible for the strategic decisions in network configuration, operations, and management.
A key group within the primary audience is the set of network managers and designers employed by or contracted to the U.S. federal government. It has been mandated that the federal government make the transition to IPv6 by June 2008, but there are significant gaps in knowledge about how to accomplish that task. This book fills the need to have a single reference source on how to upgrade networks the size of those used by even the largest of government agencies to IPv6.
The ideal reader is a network manager overwhelmed with the daunting task of planning and executing a transition of his or her network to IPv6. This book is meant for such a person to get a grasp of what needs to be done and get started doing it.
This is a book on project management. It is about taking what appears to be a Herculean problem and reducing it to manageable parts, in some cases many parts, but all manageable ones. It is about assessing the current state of your networks, defining plans for how you want them to make the transition and when and how you want them to do so. Possibly most importantly, this book is about managing a transition so that you know whether you're on track, and how to fix it if you're not.
This is not a book on technology. There are numerous books on IPv6 technology that make excellent companions to this one for working out the implementation specifics unique to each network's environment, mission, and budget. Sufficient technology is presented in this book to describe the classes of tasks to be undertaken, but the details are outside its scope. The reader should have knowledge of TCP/IP networking, including routing, tunneling, firewalls, and the troubleshooting of networking problems. At the end of this chapter, further reading on these topics is offered.
Why IPv6?
A fair question to ask is that, if IPv4 has lasted so long, why the sudden need for IPv6? The federal mandates in particular are considered by some to have been put forth by the Office of Management and Budget (OMB) from out of the blue. There are, however, several good reasons for why this is the right time to move to IPv6.
Addressing Every PC, Cell Phone, and Toaster
Demand for Internet-enabled devices is accelerating rapidly and that growth requires the use of more IP addresses. In total, there are approximately 4 billion total addresses in the IPv4 address space, that is, 32-bits worth or [2.sup.32] addresses. Up until fairly recently, IP addresses have been primarily used for computers, ranging from PCs to mainframes. That's changed in the past few years and will continue to do so as huge new categories of Internet-enabled devices become available.
The cellular communications industry is tying at least one IP address to each of millions of phones, pagers, and PDAs. Internet-enabled smart automobiles and appliances such as refrigerators and washing machines are being developed that can, for example, contact their manufacturer when a fault occurs. Cable television companies are looking to use IP features known as multicast to distribute video on demand to millions of set-top cable boxes. That requires each set-top box to have at least one IP address. Given that 82 percent of all IPv4 addresses have already been allocated (per http://www.potaroo.net/tools/ipv4/, as of May 9, 2007), there's not much room left for growth in all these new areas.
Further Extending the Uses of IP Addresses
The preceding are some of the nearer-term uses of the Internet, and therefore IP addresses, above and beyond the current uses by traditional computers. There are other concepts, further out, that foresee using IP addresses as identifiers to tag mass-produced items like medicine bottles, articles of clothing, supermarket shopping carts, and so on. Clearly, such concepts, even if only a few are developed to maturity, place a great load on the available IP address space.
To offer one final perspective, according to the U.S. Census Bureau's website (http://www.census.gov/main/www/popclock.html), there are approximately 302 million people in the United States and 6.6 billion people in the world as of May 2007. The Japan Network Information Center (JPNIC) published a report in March 2006 analyzing the IPv4 space exhaustion matter (http://www.nic.ad.jp/en/research/IPv4exhaustion_trans-pub.pdf offers an English translation of that report). According to JPNIC, the United States is assigned nearly 60 percent of the available IPv4 address space while the Census Bureau reports that the United States accounts for less than 5 percent of the world population. As the underdeveloped nations of the world become more connected to the Internet, it is clear there are not enough IPv4 addresses.
Predicting the Exhaustion of IPv4 Addresses
There has been a continuing debate for decades on how long the IPv4 address space will last. This debate and the ongoing fears of address-space exhaustion looming in the near future have been the primary reasons that IPv6 has endured so many years despite no significant adoption by commercial networks. Per the JPNIC report mentioned earlier, it has been generally agreed to by experts in the field that IPv4 addresses will run out some time between 2009 and 2016. When weighed against transition projects that themselves can take years to plan, fund, and execute, 2009 is right around the corner and even 2016 is starting to get uncomfortably close.
Easing Network Load
A side-effect of increased IP address space allocation is increased network load. This is beyond just increased network traffic. It also applies to the increased amount of overhead in the routing infrastructure that determines how to get IP packets from one place to another.
The world-wide allocation of IP addresses today has evolved over decades of organic growth and, therefore, huge lists representing hundreds of thousands of network routes must be maintained in all Internet routers in order for those routers to know how to forward a packet to a particular IP address. The IP address ranges in a given region, especially in Europe and Asia, are far from contiguous, requiring more route entries to represent all of the distinct non-connected blocks.
Defining a World-Wide Networking Hierarchy
In contrast to the organic deployment that IPv4 saw, the IPv6 address allocation scheme implements a more hierarchical structure that greatly reduces the size of the network route lists. Regions (for example, the Americas, Europe, Asia, Africa, and so on) are allocated IPv6 addresses in huge blocks, which is only possible due to the vast address space provided by IPv6. These huge allocations reduce the number of times each region must return to the world-wide central authority to request more addresses. They also reduce the number of network routes that each router needs to remember in order to forward traffic to another region.
Though it's a bit of an oversimplification, one can imagine the IPv6 allocation scheme as requiring a regional router to know only roughly as many network routes as there are continents or macroscopic geo-political entities. This is far different from an IPv4 regional router needing to know about what looks like every country, company, town, and hamlet.
The concepts used in the regional hierarchy are replicated at the national and, if necessary, the local level for individual countries in regions, and for Internet Service Providers (ISPs) in those countries. In this way, the complexities of underlying or far away routing infrastructures are hidden from any given router and that router needs only worry about forwarding any given packet either further down into the region/locale they service or to a peer region/locale. Routers below the regional level must also know how to forward traffic upward to the next larger level of hierarchy, but presumably there are not many of those upward routes, either. Ignoring system redundancy for a moment, one would be enough. It's more likely, however, that there will be more like two or three, still a far cry from hundreds.
Routing Each Packet Faster
By reducing the size of the network route lists, IP packets are forwarded faster. Another improvement in routing efficiency over IPv4 that IPv6 offers is the use of a constant-size packet header. Instead of combining seldom-used options as part of a single variable-length IPv4 header, IPv6's core capabilities are augmented using optional extensions providing security, custom routing features, and even vendor or application-specific options defined only at the packet's ultimate destination.
The vast majority of IPv6 packets require no extension headers and, therefore, the amount of processing to forward those packets is minimal. As an example, one computation that must be performed by every router on an IPv4 packet is determination of the IP header length. This computation is eliminated in IPv6, which is especially valuable in implementing the protocol in firmware.
Enabling the Future
A great many of the lessons learned from IPv4's shortcomings went into the design of IPv6. The increase in address space size and the extensibility of the core protocol via additional sub-headers was introduced earlier. The next chapter presents a few more details of each. More important than the implementation details, however, are the new classes of network architectures that such features enable.
(Continues...)
Excerpted from IPv6 Mandates by Karl A. Siil Copyright © 2008 by Karl A. Siil. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.