One document matched: draft-ietf-iab-case-for-ipv6-01.txt
Differences from draft-ietf-iab-case-for-ipv6-00.txt
Internet Architecture Board Steve King, Bay Networks
INTERNET DRAFT Ruth Fax, Bay Networks
Dimitry Haskin, Bay Networks
Wenken Ling, Bay Networks
Tom Meehan, Bay Networks
Robert Fink, LBNL
13 March 1998 Charles E. Perkins, Sun Microsystems
The Case for IPv6
draft-iab-case-for-ipv6-01.txt
Status of this Memo
This document is an Internet-Draft. Internet-Drafts are working
documents of the Internet Engineering Task Force (IETF), its areas,
and its working groups. Note that other groups may also distribute
working documents as Internet-Drafts.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at
any time. It is inappropriate to use Internet- Drafts as reference
material or to cite them other than as "work in progress."
To view the entire list of current Internet-Drafts, please check
the "1id-abstracts.txt" listing contained in the Internet-Drafts
Shadow Directories on ftp.is.co.za (Africa), ftp.nordu.net (Northern
Europe), ftp.nis.garr.it (Southern Europe), munnari.oz.au (Pacific
Rim), ftp.ietf.org (US East Coast), or ftp.isi.edu (US West Coast).
Distribution of this memo is unlimited.
Abstract
This document outlines the business and technical case for IPv6. It
is intended to acquaint both the existing IPv4 community with IPv6,
to encourage its support for change, and to attract potential future
users of Internet technology.
King, et.al. Expires 13 September 1998 [Page 1]
Internet Draft The Case for IPv6 13 March 1998
1. Introduction
This document was produced at the request of the IAB, based on an
existing original. Since IPv6 development is still in progress
at the time of this writing, the technical details are subject to
change, and the references cited may become obsolete.
The Internet Protocol (IP) has its roots in early research networks
of the 1970s, but within the past decade has become the leading
network-layer protocol. This means that IP is a primary vehicle for
a vast array of client/server and peer-to-peer communications, and
the current scale of deployment is straining many aspects of its
twenty-year old design.
The Internet Engineering Task Force (IETF) has produced
specifications (RFC 1752, 1883, 1886, 1971, 1993, etc.) that define
the next-generation IP protocol known as "IPng," or "IPv6." IPv6
is both a near-term and long-range concern for network owners and
service providers. IPv6 products have already come to market; on the
other hand, IPv6 development work will likely continue well into the
next decade. Though it is based on much-needed enhancements to IPv4
standards, IPv6 should be viewed as a new protocol that will provide
a firmer base for the continued growth of today's internetworks.
Because it is intended to replace IP (hereafter called IPv4) IPv6
is of considerable importance to businesses, consumers, and network
access providers of all sizes. IPv6 is expected to improve upon
IPv4's scalability, security, ease-of-configuration, and network
management; these issues are central to the ongoing competitiveness
and bottom-line performance of all types of network- dependent
businesses. On the other hand IPv6 aims to preserve existing
investment as much as possible. End users, industry executives,
network administrators, protocol engineers, and many others will
benefit from understanding the ways that IPv6 will affect future
internetworking and distributed computing applications.
By early 1998 a worldwide IPv6 testing and pre-production deployment
network, called the 6BONE, had already reached over 200 sites
and networks in over 30 countries. There are over 50 IPv6
implementations completed or underway worldwide, and over 25 in test
or production use on the 6BONE. This existing deployment has been
built by an active population of protocol inventors, designers and
implementers. They have worked together to solve the questions and
problems that might be expected to arise during such a huge project.
Their experience has served to validate the expectations of the
protocol designers.
This document presents IPv6 issues in two parts:
King, et.al. Expires 13 September 1998 [Page 2]
Internet Draft The Case for IPv6 13 March 1998
- The Business Case for IPv6, giving a high-level view of business
issues, protocol basics, and industry realities, and
- The Technical Case for IPv6, which delves deeper into the
functional and technical aspects of IPv6.
2. Part I: The Business Case for IPv6
Given the remarkable growth of the Internet, and business opportunity
represented by the Internet, IPv6 is of major interest to business
interests, enterprise internetworks, and the global Internet.
2.1. IPv6: A Protocol Overview
IPv6, the Next-Generation Internet Protocol, was approved by the
Internet Engineering Steering Group on November 17, 1994 as a
Proposed Standard. Since that time a large number of end-user
organizations, standards groups, and network vendors have been
working together on the specification and testing of early IPv6
implementations. A number of IETF working groups have produced IPv6
specifications that are well underway, including the basic IPv6
protocol specification, address architectures, Domain Name Servers
(DNS), security, transition mechanisms, mobility support, DHCP, and
Internet Control Message Protocol (ICMP).
Standards work on IPv6 and related components is far enough along
that vendors have already committed to a considerable number of
development and testing projects. All of the major router vendors
have committed to adding IPv6 to their products.
Vendors such as Apple, Digital Equipment, Hewlett-Packard, IBM,
Microsoft, Novell, Silicon Graphics and Sun have likewise begun
the task of delivering IPv6 on desktop machines and servers. In
addition, many organizations are working on IPv6 drivers for the
popular UNIX BSD and Linux operating environments. Network software
vendors have announced a wide range of support for IPv6 in network
applications and communication software products.
2.2. IPv6 Design Goals
IPv6 has been designed to enable high-performance, scalable
internetworks that should operate as needed well into the next
century. Much of this design process involved correcting the
inadequacies of IPv4. Some of the qualities of IPv6 are found in
enhanced features, such as the larger address space and streamlined
packet design. Other qualities relate to the fresh start that IPv6
gives to those who build and administer networks. For instance, a
King, et.al. Expires 13 September 1998 [Page 3]
Internet Draft The Case for IPv6 13 March 1998
new, well-structured, efficient routing hierarchy will be possible.
The following sections give an overview of the improvements that IPv6
brings to enterprise networking and the global Internet.
2.2.1. Addressing and Routing
IPv6 provides a framework for solving a number of problems that
currently exist inside and between enterprises. On the global
scale, IPv6 will allow Internet backbone designers to create a
flexible and expandable global routing hierarchy. At the level of
the Internet backbone where major enterprises and Internet Service
Provider (ISP) networks come together, it is necessary to maintain
a hierarchical addressing system, much like that of the national
and international telephone systems. Large central-office phone
switches, for instance, only need a three-digit national area code
prefix to route a long-distance telephone call to the correct local
exchange. The current IPv4 system uses address hierarchy to sort
traffic towards networks attached to the Internet backbone.
Without an address hierarchy, backbone routers would be forced
to store routing table information on the reachability of every
network in the world. Given the current number of IP subnets in the
world and the growth of the Internet, it is not feasible to manage
routing tables and updates for so many routes. With a hierarchy,
backbone routers can use IP address prefixes to determine how traffic
should be routed through the backbone. IPv4 uses a technique called
Classless InterDomain Routing (CIDR [RFC1219]), which specifies
variable-length network prefixes. CIDR permits "route aggregation"
at various levels of the Internet hierarchy, whereby backbone routers
can store a single routing table entry that provides reachability to
many lower- level networks.
But CIDR does not guarantee an efficient and scalable hierarchy.
In order to avoid maintaining a separate entry for each route
individually, it is important for routes at lower levels of the
routing hierarchy, that naturally have longer prefixes, to be
aggregated (or "summarized") into less specific routes at higher
levels of the routing hierarchy, that naturally have shorter
prefixes.
Legacy IPv4 address assignments that originated before CIDR and
the current access provider hierarchy often do not facilitate
summarization. The lack of uniformity of the current hierarchical
system, coupled with the rationing of IPv4 addresses, means that
Internet addressing and routing are complicated at all levels. These
issues affect high-level service providers and individual end users
in all types of businesses. Renumbering IPv4 sites when changing
from one ISP to another, to maintain address/route aggregation, is
King, et.al. Expires 13 September 1998 [Page 4]
Internet Draft The Case for IPv6 13 March 1998
unnecessarily complicated compared to IPv6's ease, and thus low cost,
of site renumbering.
2.2.2. Regularity at all levels
Many of the same problems that exist today in the Internet backbone
are also being felt at the level of the enterprise and the individual
business user. When an enterprise can't summarize its addresses by
using a small set of prefixes, it becomes more difficult to support
the expansion of backbone routing tables. If an enterprise can't
present unique addresses to the Internet, it may be forced to deploy
private, isolated address space that isn't visible to the Internet.
Users in private address spaces with non-unique addresses typically
require gateways and Network Address Translators (NATs) to manage
their connectivity to the outside world. In such situations, some
services are simply not available. A NAT is meant to allow an
enterprise to have whatever internal address structure it desires,
without concern for integrating internal addresses with the global
Internet. The NAT device sits on the border between the enterprise
and the Internet, converting private internal addresses to a smaller
pool of globally unique addresses that are passed to the backbone and
vice versa (see Figure 1).
|
|
|
|
--------------- | ----------
| | ------- | the |
| Enterprise | | | | |
| |----| NAT |-----| Internet |
| Network | | | | |
| | ------- ----------
--------------- |
NAT = Network Address Translator
|
|
Private address space | Unique global addresses
|
|
|
Figure 1: Network Address Translator
King, et.al. Expires 13 September 1998 [Page 5]
Internet Draft The Case for IPv6 13 March 1998
NAT may be appropriate in some organizations, particularly if
full connectivity with the outside world is not desired. But for
enterprises that require robust interaction with the Internet, NAT
devices often get in the way. The NAT technique of substituting
address fields in each and every packet that leaves and enters the
enterprise is very demanding, and can lead to a bottleneck between
the enterprise and the Internet. A NAT may keep up with address
conversion in a small network, but as Internet access increases, the
NAT's performance must increase in parallel. The bottleneck effect
is exacerbated by the difficulty of integrating and synchronizing
multiple NAT devices within a single enterprise. Enterprises
with NAT are less likely to achieve the reliable high- performance
Internet connectivity that is common today with multiple routers
attached to an ISP backbone in an arbitrary mesh fashion.
NAT translators also run into trouble when applications embed their
IP addresses in the packet payload, above the Network Layer. This
is the case for a number of applications, including certain File
Transfer Protocol (FTP) programs, Mobile IP, and the Windows Internet
Name Service (WINS) registration process of Windows 95 and Windows
NT. Unless a NAT parses every packet all the way to the application
level, it may fail to translate some embedded addresses, which
can lead to application failures. NAT can also break Domain Name
Servers, because they work above the Network Layer. NATs prevent
the use of IP-level security all the way from one end to another of
a transaction. NAT services are a valuable tool for certain limited
scenarios, but are not generally advantageous for the long-term
health of the Internet.
2.2.3. Mobility
IPv4 has difficulties managing mobile computers, for several reasons:
- A mobile computers needs to make use of a forwarding address at
each new point of attachment to the Internet, and it's not so
easy to get such an address with IPv4
- Informing any agent in the routing infrastructure about
the mobile node's new location requires good authentication
facilities which are not commonly deployed in IPv4 nodes.
- In IPv4, it is often more difficult for mobile nodes to find out
if they are still attached to the same network or not.
- It is unlikely in IPv4 that mobile nodes would be able to inform
their communication partners about any change in location.
King, et.al. Expires 13 September 1998 [Page 6]
Internet Draft The Case for IPv6 13 March 1998
Each of these problems is solved in a natural way by using features
designed into the IPv6 protocol suite. The benefits for mobile
computing are also apparent in quite a number of smaller details of
the IPv6 protocol design [5].
2.2.4. Renumbering
Another effect of IPv4's limitations relates to the ongoing need in
many organizations to renumber network devices -- i.e., assign new IP
addresses to them. When an enterprise changes ISPs, it may have to
either renumber all addresses to match the new ISP-assigned prefix,
or implement address translation devices (NATs). Renumbering may be
indicated when a corporation undergoes a merger or an acquisition
with consequent network consolidation.
Address shortages and routing hierarchy problems threaten the network
operations of larger enterprises, but they also affect small sites
-- even the home worker who dials in to the office via the Internet.
Smaller networks can be completely dropped from Internet backbone
routing tables if they do not adhere to the address hierarchy,
while larger networks may refuse to renumber and cause a larger
routing problem for the backbone providers of the Internet. With
today's IPv4 address registries, ISPs with individual dial-in clients
cannot allocate IP numbers as freely as they wish. Consequently,
many dial-in users must use an address allocated from a pool on a
temporary basis. In other cases, small dial-in sites are forced to
share a single IP address among multiple end systems.
A unique IP address enables users to gain direct connectivity
to other users on the Internet, and simplifies a wide range of
productive interactive applications, of which telecommuting is only
a single example. Today's hierarchy of limited and poorly allocated
addresses has already caused problems, and it will continue to
degrade as more and more devices of varying capabilities are added to
ISP rosters.
2.3. IPv6 -- the solution
IPv6, with its larger address space, enables the definition of a
flexible, hierarchical global routing architecture with several
levels. Using CIDR-style flexible prefixes, the IPv6 address space
can be allocated in a way that facilitates route summarization,
and controls expansion of route tables in backbone routers. IPv6
addressing means that large enterprises can avoid private address
spaces. ISPs will have enough addresses to allocate to smaller
businesses and dial-in users that need globally unique addresses
to fully exploit the Internet. Using an example from crowded
King, et.al. Expires 13 September 1998 [Page 7]
Internet Draft The Case for IPv6 13 March 1998
telephone networks, one might say that IPv6 eliminates the need for
"extensions", so that all offices have direct communication lines and
do not need operators (automatic or otherwise) to redirect calls.
2.3.1. Reducing Address Administration Workloads
Each IPv6 node initially creates a local IPv6 address for itself
using "stateless" address autoconfiguration, not requiring a manually
configured server. Stateless autoconfiguration further makes it
possible for stations to configure their own globally routable
addresses with the help of a local IPv6 router. Typically, the
station combines its 48 or 64 bit MAC address (assigned by the
equipment manufacturer) with a network prefix it learns from a
neighboring router. This keeps end user costs down by not requiring
knowledgeable staff to properly configure each workstation before it
can be deployed. These costs are currently part of the equipment
rollout for almost all IPv4 computing platforms.
IPv4 networks often employ the Dynamic Host Configuration Protocol
(DHCP) to reduce the effort associated with manually assigning
addresses to endstations. DHCP is termed a "stateful" address
configuration tool because it maintains static tables that determine
which addresses are assigned to new or moved stations. A new version
of DHCP is being developed for IPv6 to provide similar stateful
address assignment as may be desired by many network administrators.
DHCPv6 also assists with efficient reconfiguration in addition to
initial address configuration, by using multicast from the DHCP
server to any desired population of clients.
The robust autoconfiguration capabilities of IPv6 will be a boon to
internetwork users at many levels. When an enterprise is forced to
renumber because of an ISP change, IPv6 autoconfiguration will allow
hosts to be given new prefixes without manual reconfiguration of
workstations or DHCP clients. This function also assists enterprises
in keeping up with dynamic end-user populations. Autoconfiguration
allows mobile computers to receive valid forwarding addresses
automatically, no matter where they connect to the network.
2.4. IPv6's Streamlined Format
IPv6 streamlines and enhances the basic header layout of the IP
packet (see Figure 2). In IPv6, some of the IPv4 header information
was dropped or made optional. The simplified packet structure is
expected to offset the bandwidth cost of the longer IPv6 address
fields. The 16- byte IPv6 addresses are four times longer than the
4-byte IPv4 addresses, but as a result of the retooling, the total
King, et.al. Expires 13 September 1998 [Page 8]
Internet Draft The Case for IPv6 13 March 1998
IPv6 header size is only twice as large; many processing aspects are
substantially more efficient.
+-------+--------------+----------------------------------------+
|4 bits | 8 bits | 20 bits |
|Version| Traffic Class| Flow Label |
+-------+--------------+--------+---------------+---------------+
| 16 bits | 8 bits | 8 bits |
| Payload Length | Next Header | Hop Limit |
+-------------------------------+---------------+---------------+
| |
+ 128 bits +
| |
+ Source Address +
| |
+ +
| |
+---------------------------------------------------------------+
| |
+ 128 bits +
| |
+ Destination Address +
| |
+ +
| |
+---------------------------------------------------------------+
Figure 2: IPv6 Packet Layout
IPv6 encodes IP header options differently to streamline the
forwarding process. Optional IPv6 header information is conveyed
in independent "extension headers" located after the IPv6 header
and before the transport-layer header in each packet. Most IPv6
extension headers are not examined or processed by intermediate nodes
(in contrast with IPv4). IPv6 header extensions are variable in
length and can be longer than before. Network protocol designers can
introduce new header options in a straightforward manner.
So far, option fields have been defined for carrying explicit
routing information created by the source node, as well as for
authentication, encryption, and fragmentation control. At the
application level, header extensions are available for specialized
end-to-end network applications that require their own header fields
within the IP packet.
King, et.al. Expires 13 September 1998 [Page 9]
Internet Draft The Case for IPv6 13 March 1998
2.5. Native Security
Encryption, authentication, and data integrity safeguards are needed
for enterprise internetworking. IPv4 network-layer extensions for
this have been defined, compatible with those for IPv6, but are not
yet in wide use.
To correct this situation, IPv6 builds in security capabilities that
are based on its flexible header extensions. The authentication
header extension to IPv6 allows a receiver to know for certain
whether a packet does indeed come from the host indicated in its
source address. This authentication safeguards against the use
of forged source addresses by attackers. Such source-address
masquerading (spoofing) might otherwise fool a server into granting
access to valuable data, passwords, or network control utilities.
According to recent studies, IP spoofing is statistically one of
the most common forms of deliberate intrusion; with IPv4 it is
typically impossible for a server to determine whether packets are
being received from the legitimate endstation. Some enterprises
have responded by putting proprietary firewalls in place, but these
devices introduce a number of new problems, including performance
bottlenecks, restrictive network policies, and limited connectivity
to the Internet or even between divisions of the same company.
IPv6 gives the industry a standard method to determine the
authenticity of packets received at the Network Layer. Because the
authentication headers in IPv6 are defined in IETF standards, network
products from different vendors can use interoperable authentication
services. IPv6 implementations are required to support the MD5
algorithm for authentication and integrity checking, but since the
specification is algorithm-independent, other techniques may be used
as well. IPv6 authentication will be particularly valuable where
autoconfiguration is deployed. Without Network Layer authentication,
malicious users may take advantage of DHCP and similar services to
gain unassisted entry to a network. IPv6 authentication is expected
to prevent such illicit autoconfiguration.
2.6. Confidentiality and Privacy
Along with packet spoofing, another major hole in Internet security
is the widespread deployment of traffic analyzers and network
"sniffers," which can surreptitiously eavesdrop on network traffic.
These generally helpful diagnostic devices can be misused by those
seeking access to credit card and bank account numbers, passwords,
trade secrets, and other valuable data. IPv4 provides no native data
encryption scheme, resulting in reliance on other methods less likely
to be interoperable.
King, et.al. Expires 13 September 1998 [Page 10]
Internet Draft The Case for IPv6 13 March 1998
Privacy (data confidentiality) is provided by a standard header
extension for end-to-end encryption at the Network Layer. IPv6
encryption headers indicate which encryption keys to use and carry
other handshaking information, enabling interoperable encryption of
the payloads in IP packets. Both IPv6 security headers can be used
directly between hosts or in conjunction with a specialized security
gateway that adds an additional level of security with its own packet
signing and encryption methods.
2.7. Multicast and Anycast
Modern internetworks need to transmit streams of video, audio, news,
financial, or other timely data to groups of functionally related
but dispersed endstations. This is best achieved by Network Layer
multicasting techniques. Typically, a server sends out a stream
of multimedia or time-sensitive data that needs to be received
by subscribers. A multicast-capable network can automatically
replicate the server's packets and route them to each subscriber in
the multicast group using an efficient path (see Figure 3). Routers
use multicast protocols such as DVMRP (Distance Vector Multicast
Routing Protocol) and MOSPF (Multicast Open Shortest Path First) to
dynamically construct a packet distribution "tree" that connects all
members of a group with the multicast server.
King, et.al. Expires 13 September 1998 [Page 11]
Internet Draft The Case for IPv6 13 March 1998
Multicast Source
---
| |
Multicast transmissions | |
only received by ---
Multicast Group Members |
| Video, audio, animated graphics.
| real-time financial data, news feeds...
-----------------------------------------------------------------
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| --- | | --- | | | |
| | | | | | | | --- | |
| | | | | | | | | | | ---
--- --- | --- --- | | | | | |
| | | | | | --- | | |
| | --- | | --- Multicast | ---
--- | | --- | | Group |
Multicast | | Multicast | | Member ---
Group --- Group --- | |
Member Member | |
---
Multicast
Group
Member
Figure 3: Multicast in Action
A new member becomes part of a multicast group by sending a "join"
message to a nearby router. The distribution tree is then adjusted
to include the new route. Servers can then multicast a single
packet, and it will be replicated as needed and forwarded through the
internetwork to the multicast group. This conserves both server and
network resources and, hence, is superior to unicast and broadcast
solutions. Multicast applications have been developed for IPv4, but
IPv6 extends IP multicasting capabilities by defining a much larger
multicast address space as well as a scope identifier that defines
the propagation limits of the multicast information. In IPv6,
broadcast is viewed as a special case of multicasting.
King, et.al. Expires 13 September 1998 [Page 12]
Internet Draft The Case for IPv6 13 March 1998
2.8. Anycast
Anycast services, supported in the IPv6 specification, are not
defined architecturally in IPv4. Conceptually, anycast is a cross
between unicast and multicast: an arbitrary number of nodes are
designated as an anycast group. A packet addressed to the group's
anycast address is delivered to only one of the nodes in the group,
typically the node with the "nearest" interface in the group,
according to current routing protocol metrics. This is in contrast
with multicast services, which deliver packets to all members of the
multicast group. Nodes in an anycast group are specially configured
to recognize anycast addresses, which are drawn from the unicast
address space.
Anycasting is a new service, and its applications have not been fully
developed. Using anycast, an enterprise could forward packets to
exactly one of the routers on its ISP's backbone (see Figure 4). If
all of a provider's routers have the same anycast address, traffic
from the enterprise will have several redundant access points to the
Internet. And if one of the backbone routers goes down, the next
nearest device automatically will receive the traffic.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 4: Anycast
As anycast matures, it may become an important method for allowing
endstations to efficiently access well-known services, mirrored
databases, Web sites, and message servers. It provides a flexible
and cost-effective model for enabling application robustness and load
balancing. For instance, anycast could provide enterprise robustness
by assigning all the DNS servers in an enterprise the same anycast
address.
King, et.al. Expires 13 September 1998 [Page 13]
Internet Draft The Case for IPv6 13 March 1998
2.9. Quality of Service
IPv4 carries a "differentiated services" byte and IPv6 carries an
equivalent "traffic class" byte, intended for support of simple
differentiated services. Both IPv4 and IPv6 can support the RSVP
protocol for more complex quality of service implementations.
Additionally, the IPv6 packet format contains a new 20-bit
traffic-flow identification field that will be of great value
to vendors who implement quality-of- service network functions.
Network-layer Quality of Service (QoS) products are still in the
planning stage, but IPv6 lays the foundation so that a wide range
of QoS functions may be made available in a open and interoperable
manner.
2.10. The Transition to IPv6
The transition from IPv4 to IPv6 could take one of several paths.
Some are lobbying for a wholesale, rapid adoption of IPv6 in the
very near future. Others prefer to let the IPv6 project wait until
address-space exhaustion and other issues leave no other choice.
Given the magnitude of the migration, IPv4 and IPv6 will coexist for
an extended period of time.
Therefore, IETF protocol designers have expended a substantial amount
of effort to ensure that hosts and routers can be upgraded to IPv6
in a graceful, incremental manner. The transition will prevent
isolation of IPv4 nodes, and "fork-lift" upgrades for entire user
populations. Transition mechanisms have been engineered to allow
network administrators flexibility in how and when they upgrade hosts
and intermediate nodes. IPv6 can be deployed in hosts first, in
routers first, or, alternatively, in a limited number of adjacent or
remote hosts and routers. The nodes that are upgraded initially do
not have to be colocated in the same local area network or campus.
IPv6 transition designers note that many upgraded hosts and routers
will need to retain downward compatibility with IPv4 devices for
an extended time period (possibly years or even indefinitely). It
was also assumed that upgraded devices should have the option of
retaining their IPv4 addressing. To accomplish these goals, IPv6
transition relies on several special functions that have been built
into the IPv6 standards work, including dual-stack hosts and routers
and tunnelling IPv6 via IPv4.
2.11. The Dual-Stack Transition Method
Initial users of IPv6 machines will require continued interaction
with existing IPv4 nodes. This is accomplished with the dual-stack
King, et.al. Expires 13 September 1998 [Page 14]
Internet Draft The Case for IPv6 13 March 1998
IPv4/IPv6 approach. A great many hosts and routers in today's
multivendor, multiplatform networking environment already support
multiple network stack components. For instance, the majority of
routers in enterprise networks are multiprotocol routers. Many
workstations run some combination of IPv4, IPX, AppleTalk, NetBIOS,
SNA, DECnet, or other protocols. The inclusion of one additional
protocol (IPv6) on an endstation or router is a well-understood
problem. When running a dual IPv4/IPv6 stack, a host has access to
both IPv4 and IPv6 resources. Routers running both protocols can
forward traffic for both IPv4 and IPv6 end nodes.
Dual-stack machines can use totally independent IPv4 and IPv6
addresses, or they can be configured with an IPv6 address that
is IPv4-compatible. Dual-stack nodes can use conventional IPv4
autoconfiguration services (DHCP) to obtain their IPv4 addresses.
IPv6 addresses can be manually configured in the 128-bit local
host tables, or preferably obtained via IPv6 stateless or stateful
autoconfiguration mechanisms. Major servers will run in dual-stack
mode until all active nodes are converted to IPv6.
2.12. IPv6 DNS
Domain Name Service is something that administrators must consider
before deploying IPv6 or dual-stack hosts. The current 32-bit name
servers cannot handle name-resolution requests for 128-bit addresses
used by IPv6 devices. In response to this issue, IETF designers have
defined an IPv6 DNS standard (RFC 1886, DNS Extensions to Support IP
Version 6). This specification creates a new 128-bit DNS record type
named "AAAA" (quad A) that will map domain names to an IPv6 address.
Domain name lookups (reverse lookups) based on 128-bit addresses also
are defined. Once an IPv6-capable DNS is in place, dual-stack hosts
can interact interchangeably with IPv6 nodes. If a dual-stack host
queries a DNS and receives back a 32-bit address, IPv4 is used; if a
128-bit address is received, then IPv6 is used. Where the DNS has
not been upgraded to IPv6, hosts can resolve name-to-address mappings
through the use of manually configured local name tables.
IPv6 autoconfiguration and IPv6 DNS can be linked by using dynamic
DNS updates, coupled with secure DNS. By these means DNS servers can
be securely and automatically updated whenever an IPv6 node acquires
a new address, enabling an additional measure of convenience compared
with renumbering in IPv4 today.
2.13. Application Modification for IPv6
Applications that do not directly access the network stack (i.e.
do not call a socket or DNS API and do not handle numeric IP
King, et.al. Expires 13 September 1998 [Page 15]
Internet Draft The Case for IPv6 13 March 1998
addresses in any way) need no modifications to run in the dual-stack
environment. Network applications that directly interface with IP
and related components will require updating if they are to use the
IPv6 protocol. For example, applications that access DNS or use
sockets must be enhanced with the capability to handle AAAA records
and 128-bit addresses. Applications using IPv6 security, quality of
service, and other features will need more extensive updating.
2.14. Routing in IPv6/IPv4 Networks
Routers running both IPv6 and IPv4 can be administered in much the
same fashion that IPv4-only networks are currently administered.
IPv6 versions of popular routing protocols, such as Open Shortest
Path First (OSPF) and Routing Information Protocol (RIP), are already
under development. In many cases, administrators will choose to
keep the IPv6 topology logically separate from the IPv4 network,
even though both run on the same physical infrastructure. This will
allow the two to be administered separately. In other cases, it may
be advantageous to align the two architectures by using the same
domain boundaries, areas, and subnet organization. Both approaches
have their advantages. A separate IPv6 architecture can be used
to replace the inefficient IPv4 addressing systems burdening many
of today's enterprises. An independent IPv6 architecture presents
the opportunity to build a fresh, hierarchical network address
plan that will facilitate connection to one or more ISPs. This
lays a foundation for efficient renumbering, route aggregation
(summarization), and the other goals of a routing hierarchy.
In most organizations where IPv6 is deployed incrementally, IPv6
hosts will likely have direct connectivity to each other only via
IPv4 routers. Such hosts will exist in islands of IPv6 topology
surrounded by an ocean of IPv4. Fortunately, IPv6 designers have
fashioned transition mechanisms that allow IPv6 hosts to communicate
over intervening IPv4 networks. The essential technique of these
mechanisms is IPv6 over IPv4 tunnelling, which carries IPv6 packets
within IPv4 packets (see Figure 5).
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
King, et.al. Expires 13 September 1998 [Page 16]
Internet Draft The Case for IPv6 13 March 1998
Figure 5: IPv6 over IPv4 Tunnelling
Tunnelling allows early IPv6 implementations to take advantage of
existing IPv4 infrastructure without any change to IPv4 components.
A dual-stack router or host on the "edge" of the IPv6 topology simply
inserts an IPv4 header in front of ("encapsulates") each IPv6 packet
and sends it as native IPv4 traffic through existing links. IPv4
routers forward this traffic without knowledge that IPv6 is involved.
On the other side of the tunnel, another dual-stack router or host
"decapsulates" (removes the extra IP header from) the IPv6 packet and
routes it to the ultimate destination using standard IPv6.
To accommodate different administrative needs, IPv6 transition
mechanisms include two types of tunnelling: automatic and
configured. To build configured tunnels, administrators manually
define IPv6-to- IPv4 address mappings at tunnel endpoints. On
either side of the tunnel, traffic is forwarded with full 128-bit
addresses. At the tunnel entry point, a router table entry is
defined manually to dictate which IPv4 address is used to traverse
the tunnel. This requires a certain amount of manual administration
at the tunnel endpoints, but traffic is routed through the IPv4
topology dynamically, without the knowledge of IPv4 routers. The
128-bit addresses do not have to align with 32-bit addresses in any
way.
2.15. Automatic Tunnelling
Automatic tunnels use "IPv4-compatible" addresses, which are hybrid
IPv4/IPv6 addresses. A compatible address is created by adding
leading zeros to a 32-bit IPv4 address to pad it out to 128 bits.
When traffic is forwarded with a compatible address, the device at
the tunnel entry point can automatically address encapsulated traffic
by simply converting the IPv4-compatible 128-bit address to a 32-bit
IPv4 address. On the other side of the tunnel, the IPv4 header is
removed to reveal the original IPv6 address. Automatic tunnelling
allows IPv6 hosts to dynamically exploit IPv4 networks, but it does
require the use of IPv4-compatible addresses, which do not bring the
benefits of the 128-bit address space.
IPv6 nodes using IPv4-compatible addresses cannot take advantage
of the extended address space, but they can exploit the other IPv6
enhancements, including flow labels, authentication, encryption,
multicast, and anycast. Once a node is migrated to IPv6 with IPv4-
compatible addressing, the door is open for a fairly painless move to
King, et.al. Expires 13 September 1998 [Page 17]
Internet Draft The Case for IPv6 13 March 1998
the full IPv6 address space (hopefully with the help of an IPv6-based
autoconfiguration service). IPv4-compatible addressing means that
administrators can add IPv6 nodes while initially preserving their
basic addressing and subnet architecture. Automatic tunnels are
available when needed, but they may not be necessary in cases where
major backbone routers are upgraded all at once to include the
IPv6 stack. This is something that can be achieved quickly and
efficiently when backbone routers support full remote configuration
and upgrade capabilities .
2.16. Other mechanisms and summary
Additional mechanisms for transition or for IPv4/IPv6 coexistence
are also in discussion. For example there is a proposal to use IPv4
multicast to support neighbour discovery by isolated IPv6 nodes.
There are several proposals on how to support transactions between
IPv4-only nodes and IPv6 nodes that do not have IPv4-compatible
addresses.
It could be argued that IETF members are putting as much effort
into transition as they are the basic IPv6 protocol specification.
The combination of tunnels, compatible addresses, and dual-stack
nodes gives network administrators the range of flexibility and
interoperability they need when they deploy IPv6. Transition
services allow organizations depending upon current IPv4 networks to
take advantage of the more technical IPv6 features, many of which are
discussed in Part II of this document.
3. Part II: The Technical Case for IPv6
3.1. IPv6 Headers vs. IPv4 Headers
A good way to start an in-depth investigation of IPv6 is to compare
the new streamlined IPv6 header with the current IPv4 header. Both
headers carry version numbers and source/destination addresses,
but as Figure 6 shows, the IPv6 header is considerably simplified,
which makes for more efficient processing by routing nodes. Whereas
the IPv4 headers are variable in length, all IPv6 headers have a
fixed length of 40 bytes. This allows router software designers
to optimize the parsing of IPv6 headers along fixed boundaries.
Additional processing efficiencies have been realized by reducing the
number of required header fields in IPv6. The classic IPv4 header
contains 14 fields, whereas IPv6 only uses 8 fields.
One of the first IPv4 components to be discarded was the header
length field, which is clearly no longer required due to the fixed
header length of all IPv6 packets. The total length field of IPv4
King, et.al. Expires 13 September 1998 [Page 18]
Internet Draft The Case for IPv6 13 March 1998
has been retained in the guise of the IPv6 payload length field. But
this field does not include the length of the IPv6 header, which is
always assumed to be 40 bytes. The new payload length field can
accommodate packets up to 64 KB in length. Even larger packets,
called "jumbograms", can be passed between IPv6 nodes if the payload
length field is set to zero and a special extension header is added,
as discussed below.
The time-to-live field of IPv4 has been renamed the IPv6 hop limit
field, to describe more accurately its actual function. The field is
used by routers to detect and break loops, by decrementing a maximum
hop value by 1 for each hop of the end-to-end route. The hop-limit
field is set to the appropriate value by the source node. When the
value in the hop limit field is decremented to zero, the packet is
discarded. The IPv6 hop-count field allows up to 255 hops, which
exceeds the needs for even the largest of networks, as best we can
calculate today.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 6: IPv4 and IPv6 Header Formats
In addition to the header length field, a number of basic IPv4
fields were eliminated from the IPv6 header: fragment offset,
identification, flags, checksum, and header length. The IPv4
type-of-service field is replaced by the IPv4 traffic class field,
plus the all-new flow label field. The IPv4 fragmentation fields
(offset, identification, and flags) have been moved to optional
headers in IPv6, as is discussed below. Finally, the IPv4 checksum
field has been abandoned in IPv6, since error checking typically is
duplicated at other levels of the protocol stack. It is assumed that
bad packets will be detected below, at the link layer, or above,
at transport or higher layers. Requiring routers to perform error
checking has caused reduced performance in today's Internet.
King, et.al. Expires 13 September 1998 [Page 19]
Internet Draft The Case for IPv6 13 March 1998
3.2. Specialized Extension Headers
To allow IPv4 packet headers the flexibility to carry optional
information relevant to the routing process or host applications,
IPv4 headers included an options field. This little-used field is
meant to convey information about security, source routing, and
other optional parameters. The IPv4 options field has been replaced
in IPv6 by flexible extension headers that are located after the
primary IPv6 header and before the transport header and application
payload. IPv6 extension headers are optional and provide support
security, fragmentation, source routing, network management, and
many other functions. An IPv6 packet can carry virtually any number
of extension headers between the initial header and the higher
layer payload. Figure 7 shows encryption and fragmentation headers
travelling after the primary IPv6 header and before the Transmission
Control Protocol (TCP) header.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 7: IPv6 Extension Headers
The IPv6 extension header architecture replaces the IPv4 options
field and also impacts the protocol type field, which is currently
used to indicate the type of protocol within the datagram's payload,
e.g., TCP or User Datagram Protocol (UDP). IPv6 replaces the protocol
type field with a next header field that indicates the protocol
carried in the next extension or payload header (e.g., a TCP/UDP
header or a IPv6 optional header).
The IPv6 standards groups have already defined a number of extension
headers and have also created a suggested (but not mandatory)
guideline for the order of header insertion.
The suggested order for extension headers is as follows:
- (Primary IPv6 header)
- Hop-by-Hop options header
- Destination options header-1
King, et.al. Expires 13 September 1998 [Page 20]
Internet Draft The Case for IPv6 13 March 1998
- Source Routing header
- Fragmentation header
- Authentication header
- IPv6 Encryption header
- Destination options header-2
- (Upper-layer headers)
- (Payload)
Each extension header typically occurs only once within a given
packet, except for the destination header, as explained on the
following page.
3.3. Hop-by-Hop Options Header
When present, this header carries options that are examined by
intermediate nodes along the forwarding path. It must be the first
extension header after the initial IPv6 header. Since this header
is read by all routers along the path, it is useful for transmitting
management information or debugging commands to routers. One
currently defined application of the hop-by-hop extension header
is the Router Alert option, which informs routers that the packet
should be processed completely by a router before it is forwarded
to the next hop. An example of such a packet is an RSVP resource
reservation message.
3.4. Destination Options Headers
There are two variations of this header, each with a different
position in the packet. The first incidence of this field is
for carrying information to the first destination listed in the
IPv6 address field. This header can also be read by a subsequent
destination listed in the source routing header address fields. The
second incidence of this header is used for optional information that
is only to be read by the final destination. For efficiency, the
first variation is typically located towards the front of the header
chain, directly after the hop-by-hop header (if any). The second
variation is relegated to a position at the end of the extension
header chain, which is typically the last IPv6 optional header before
transport and payload.
3.5. Source Routing Header
The IPv6 routing extension header is an incarnation of the source
routing function supported currently by IPv4. This optional header
allows a source node to specify a list of IP addresses that dictate
what path a packet will traverse. IETF RFC 1883 defines a version
King, et.al. Expires 13 September 1998 [Page 21]
Internet Draft The Case for IPv6 13 March 1998
of this routing header called "Type 0," which gives a sending node
a great deal of control over each packet's route. Type 0 routing
headers contain a 24-bit field that indicates how intermediate nodes
may forward a packet to the next address in the routing header.
IPv6 source routing corresponds to IPv4's "loose source routing"
(LSR) option (see Figure 8). In "loose" forwarding, unlisted routers
can be visited by a packet. The source routing feature works in
conjunction with another routing header field that contains a value
equal to the total number of segments remaining in the source route.
Each time a hop is made, this "segments left" field is decremented.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 8: Source Routing Extension Header
When Type 0 routing headers are used, the initial IPv6 header
contains the destination addresses of the first router in the
source route, not the final destination address. At each hop,
the intermediate node replaces this destination address with the
address of the next routing node, and the "segments left" field is
decremented.
3.6. Fragmentation Header
IPv4 has the ability to fragment packets at any point in the path,
depending on the transmission capabilities of the links involved.
This feature has been dropped in IPv6 in favor of end-to-end
fragmentation/reassembly, which is executed only by IPv6 source
and destination nodes. Packet fragmentation is not permitted in
intermediate IPv6 nodes. The elimination of the fragmentation field
allows a more streamlined packet and better router performance for
the majority of cases where fragmentation is not required. Today's
networks generally support frame sizes that are large enough to
carry typical IP packets without fragmentation. In the event that
fragmentation is required, IPv6 provides an optional extension header
King, et.al. Expires 13 September 1998 [Page 22]
Internet Draft The Case for IPv6 13 March 1998
that is used by source nodes to divide packets into an arbitrary
number of smaller units.
The IPv6 fragmentation header contains fields that identify a
group of fragments as a packet and assigns them sequence numbers.
Because IPv6 routers do not fragment packets between end nodes, the
responsibility for sending the correct size packet is with the source
node, which needs to determine the Maximum Transmission Unit (MTU) of
the links in the end-to-end path. For instance, if two FDDI networks
with 4500-byte MTUs are connected by an Ethernet with an MTU of 1500,
then the source station must send packets that are no larger than
1500.
If higher level applications are using larger payloads, the source
node can make use of the IPv6 fragmentation extension header to
divide large packets into 1500-byte units for network transmission.
The IPv6 destination node will reassemble these fragments in a manner
that is transparent to upper layer protocols and applications. End
nodes performing fragmentation can determine the smallest MTU of a
path with the MTU path discovery process (e.g., RFC1191; see Figure
9). Typically, with this technique, the source node sends out a
packet with an MTU as large as the local interface can handle. If
this MTU is too large for some link along the path, an ICMP "Datagram
too big" message will be sent back to the source. This message
will contain a packet-too-big indicator and the MTU of the affected
link. The source can then adjust the packet size downward (fragment)
and retransmit another packet. This process is repeated until a
packet gets all the way to the destination node. The discovered
MTU is then used for fragmentation purposes. Although source-based
fragmentation is fully supported in IPv6, it is recommended that
network applications adjust packet size to accommodate the smallest
MTU of the path. This will avoid the overhead associated with
fragmentation/reassembly on source and destination nodes.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 9: MTU Discovery Process
King, et.al. Expires 13 September 1998 [Page 23]
Internet Draft The Case for IPv6 13 March 1998
3.7. Authentication Header
IPv6 has two security extension headers, one that enables the
authentication of IP traffic for security purposes, and another that
fully or partially encrypts IP packets. Implementation of security
at the IP level can benefit "security aware" applications, as well as
"security ignorant" applications that don't take explicit advantage
of security features.
The IPv6 authentication extension header gives network applications a
guarantee that the packet did in fact come from an authentic source.
This prevents hackers from configuring an IP host to impersonate
another, to gain access to secure resources. Such spoofing can be
used to obtain valuable financial and corporate data and can give
persons outside the enterprise control of servers for malicious
purposes. With IPv6 authentication headers, hosts establish a
standards-based security association that is based on the exchange of
algorithm-independent secret keys (e.g., MD5).
In a client/server session, for instance, both the client and
the server need to have knowledge of the key. Before each packet
is sent, IPv6 authentication creates a checksum based on the key
combined with the entire contents of the packet. This checksum
is then re-run on the receiving side and compared. This approach
provides authentication of the sender and guarantees that data
within the packet has not been modified by an intervening party.
Authentication can take place between clients and servers or client
and clients on the corporate backbone. It can also be deployed
between remote stations and corporate dial-in servers to ensure that
the perimeter of the corporate security is not breached.
3.8. IPv6 Encryption Header
Authentication headers eliminate a number of host spoofing and
packet modification hacks, but they do not prevent the nondisruptive
reading (sniffing, snooping) of the content of packets as they
traverse the Internet and corporate backbone networks. This is
the area addressed by the Encapsulating Security Payload (ESP)
service of IPv6 -- another optional extension header. Packets
protected by the ESP encryption techniques can have very high levels
of privacy and integrity -- something that is not widely available
with the current Internet, except with certain secure applications
(e.g., private electronic mail and secure HTTP Web servers). ESP
provides encryption at the network layer, making it available to all
applications in a standardized fashion.
IPv6 ESP is used to encrypt the transport-layer header and payload
(e.g., TCP, UDP), or the entire IP datagram. Both these methods are
King, et.al. Expires 13 September 1998 [Page 24]
Internet Draft The Case for IPv6 13 March 1998
accomplished with an ESP extension header that carries encryption
parameters and keys end-to-end. When just the transport payload
is to be encrypted, the ESP header is inserted in the packet
directly before the TCP or other transport header. In this case,
the headers before the ESP header are not encrypted and the headers
and payload after the ESP header are encrypted. This is referred
to as "transport-mode" encryption. If it is desirable to encrypt
the entire IP datagram, a new IPv6 and an ESP header are wrapped
around all the fields (including the initial address fields) of the
packet. Full datagram encryption is sometimes called "tunnel-mode"
encryption because the contents of the datagram are only visible at
the endpoints of the security tunnel (see Figure 10).
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 10: Tunnel Mode and Transport Mode of IPv6
Fully encrypted datagrams are somewhat more secure than transport
mode encryption because the headers of the fully encrypted packet are
not available for traffic analysis.
For instance, full tunnel-mode encryption allows the addresses
contained in IPv6 source routing headers to be hidden from packet
sniffing devices for the public portion of a path. There is a
considerable performance penalty for full encryption, due to the
overhead and processing cost of adding an additional IPv6 header
to each datagram. In spite of its cost, full ESP encryption is
particularly valuable to create a security tunnel (steel pipe)
between the firewalls of two remote sites (see Figure 11). The
full datagram encryption in the tunnel ensures that the various
headers and address fields of encrypted packets will not be visible
as traffic traverses the public Internet. Within the tunnel, only
the temporary encapsulating address header is visible. Once through
the tunnel and safely within a firewall, the leading ESP headers are
stripped off and the packet is again visible, including any source
routing headers required to finish the path.
King, et.al. Expires 13 September 1998 [Page 25]
Internet Draft The Case for IPv6 13 March 1998
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 11: Firewalls and Steel Pipe
The encryption and authentication services of IPv6 together create
the security solution typically needed by business and military
applications. In some cases an authentication header will be carried
inside an encrypted datagram, providing an additional layer of data
integrity and verification of the sender's identification. In
other cases, the authentication header may be placed in front of
the encrypted transport-mode portion of the packet. This approach
is desirable when the authentication takes place before decryption
on the receiving end, which is the logical order in many cases.
Taken together, the authentication and encryption services of IPv6
provide a robust, standards-based security mechanism that will play a
decisive role in the continuing expansion of commerce and corporate
operations onto IP-based network fabrics.
3.9. The IPv6 Address Architecture
Much of the discussion of IPv4 versus IPv6 focuses on the relative
size of the address fields of the two protocols (32 bits versus 128
bits). But an equally important difference is the relative abilities
of IPv6 and IPv4 to provide an advanced hierarchical address space
that facilitates efficient routing architectures. IPv4 was initially
designed with a class-based address scheme (see Figure 12), which
divided address bits between network and host but did not create a
hierarchy that would allow a single high-level address to represent
many lower-level addresses. Hierarchical addressing systems work in
much the same way as telephony country codes or area codes, which
allow long-haul phone switches to route calls efficiently to the
correct country or region using only a portion of the full phone
number.
King, et.al. Expires 13 September 1998 [Page 26]
Internet Draft The Case for IPv6 13 March 1998
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 12: IPv4 Address Classes
As the Internet grows, the non-hierarchical nature of the original
IPv4 address space is proving inadequate.
The limitations of IPv4 addressing are currently hampering both
the local and global levels of internetworking. To combat IPv4
deficiencies at the local area network level, the subnetting
technique has been developed to create a more manageable division of
large networks. With subnet addressing, a single network address
can stand for a number of physical networks, which conserves address
space considerably (e.g., a single Class B address can be used to
access hundreds of physical networks, each of which itself could have
dozens or hundreds of individual hosts).
At the level of large internet backbones and global routing, IPv4
addresses can be more efficiently aggregated with supernetting, a
form of hierarchical addressing. With supernetting, backbone routers
store a single address that represents the path to a number of lower
level networks. This can considerably reduce the size of routing
tables in backbone routers, which increases backbone performance
and lowers the amount of memory and number of processing routers
required. Subnetting and supernetting have been particularly useful
in extending the viability of the IPv4 Class C addresses. Both of
these techniques are made possible by pairing addresses stored in
routers with bit masks that indicate which bits in an address are
valid at the various levels of the hierarchy.
The process of creating an IPv4 routing hierarchy was formalized
in Classless Interdomain Routing (CIDR) [3] which uses bit masks
to allocate a variable portion of the 32-bit IPv4 address to
network, subnet, or host. For instance, CIDR allows a number of
(plentiful) Class C addresses to be summarized by a single prefix
address, allowing Class C addresses to function in a similar way
to hard-to-get Class A and Class B addresses. CIDR has extended
King, et.al. Expires 13 September 1998 [Page 27]
Internet Draft The Case for IPv6 13 March 1998
the life of IPv4 and helped the Internet scale to its current size,
but it has not been implemented in a consistent way across the
Internet and enterprise networks. Consequently, the routing table
efficiencies and address space conservation advantages of CIDR are
not today fully realized, nor will they ever be fully realized,
due to the legacy nature of IPv4 networks and the difficulty
of restructuring them. IPv4 will continue to waste its already
inadequate address space as it continues to burden routers with
inefficient routes and excessively large routing tables.
Yet another downside of IPv4 is found at the departmental and
workgroup level of internetworking, in the high administrative
workload associated with maintaining subnet bit masks and host
addresses within the subnet structure, particularly where there
are large, dynamic populations of end users. When an end user is
moved in the subnetting environment, careful attention must be
paid to ensure that the host renumbering process does not disrupt
connectivity at any level of the stack. The complexities and
pitfalls of current subnetting methods can eventually make IPv4 less
than viable in large organizations that experience ongoing growth of
internetwork user populations.
3.10. The IPv6 Address Hierarchy
In a direct response to the experience gained from IPv4, IPv6 has
been designed from the ground up to provide a scalable address space
that can be partitioned into a flexible and efficient global routing
hierarchy. At the top of this hierarchy, several international
registries assign blocks of addresses to top level aggregators (TLA).
TLAs allocate blocks of addresses to Next Level Aggregators (NLA),
which represent large providers and global corporate networks. When
an NLA is a provider, it further allocates its addresses to its
subscribers. Routing is efficient because NLAs that are under the
same TLA will have addresses with a common TLA prefix. Subscribers
with the same provider have IP addresses with an NLA common prefix.
See Figure 13 for an example of Aggregation-based Allocation
Structures.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
King, et.al. Expires 13 September 1998 [Page 28]
Internet Draft The Case for IPv6 13 March 1998
Figure 13: Aggregation-based Allocation Structures
Although a number of allocation schemes are possible within IPv6's
huge address space, an aggregation-based hierarchy is favored by
IETF designers because it can combine the advantages of provider
and geographic allocation approaches. Provider allocation divides
the hierarchy along lines of large service providers, regardless of
their location. Geographic allocation divides the hierarchy strictly
on the basis of the location of providers/subscribers (as does the
telephony system of country and area codes). But both of these
approaches have their drawbacks because large backbone networks often
don't conform strictly to geographic or provider boundaries. Some
large networks, for instance, may connect to several ISPs. And many
large networks span numerous countries and geographical regions.
Aggregation-based allocation is based on the existence today of a
limited number of high-level exchange points, where large long-haul
service providers and telco networks interconnect. The use of
these exchange points to divide the IPv6 address hierarchy has a
geographical component because exchanges are distributed around
the globe. It also has a provider orientation because all large
providers are represented at one or more exchange points.
As shown in Figure 14, the first 3 address bits indicate what type
of address follows (unicast, multicast, etc.). The next 13 bits
are allocated to the various TLAs around the world. Eight bits are
reserved for future use, and the following 24 bits are allocated to
the next lower level of providers and subscribers.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 14: Aggregation-based IPv6 Addresses
King, et.al. Expires 13 September 1998 [Page 29]
Internet Draft The Case for IPv6 13 March 1998
Next level aggregators can divide the NLA address field to create
their own hierarchy, one that maps well to the current ISP industry,
in which smaller ISPs subscribe to higher level ISPs, and so on.
This is accomplished by the further subdivision of the 32-bit NLA
field (see Figure 15).
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 15: Subdividing the NLA Address Space
Following the NLA ID are fields for subscriber site networking
information: Site Level Aggregator (SLA) and Interface ID.
Typically, service providers supply subscribers with blocks of
contiguous addresses, which are then used by individual organizations
to create their own local addressing hierarchy and identify subnets
and hosts. The 16-bit SLA field supports up to 65,535 individual
subnets. The 64-bit Interface ID, which is used to identify an IPv6
interface on a network link, will typically be derived from the
installed IEEE LAN adapter address.
Today's Internet backbone routers must maintain up to 40,000 or
more routes. As the Internet continues to scale, IPv6's uniform
application of hierarchical routing will likely be the only viable
method for keeping the size of backbone router tables under control.
With an aggregator-based address hierarchy, all of a subscriber's
internal network segments can be reached through one or more high-
level aggregation points. This allows backbone routers around the
globe to efficiently summarize the routes to a customer's networks
with high-level TLA address prefixes. Forwarding routes in highest
level backbones can be quickly calculated by looking only at the TLA
portion of the address. IPv6's large hierarchical address space
also allows a more decentralized approach to IP address allocation.
Service providers can allocate addresses independently from central
authorities, encouraging global network growth and eliminating
bureaucratic bottlenecks in the growth process.
Aggregation-based addresses are just part of the total address
space that has been defined for IPv6. Other address ranges have
King, et.al. Expires 13 September 1998 [Page 30]
Internet Draft The Case for IPv6 13 March 1998
been assigned to multicasting and to nodes that only require
unique addressing within a limited area (site-local and link-local
addresses).
Site- and link-local addresses are available for private, internal
use by all enterprises, and are not allocated by public registry
authorities. Site-local addresses enable networks to use non-unique
local addresses that are later made globally unique by adding a
prefix. This has an advantage: if an ISP changes, site local
addressing can remain the same because it is not directly interfaced
to the outside world. Link local addresses can be used for
applications that are limited in scope to a single link, and also for
temporary "bootstrapping" of stations before they receive a globally
unique address (more on this in the section below).
3.11. Host Address Configuration
IPv6 has a large enough address architecture [4] to accommodate
Internet expansion for many decades to come. But the usefulness of
IPv6 addresses would be limited if not matched with equally advanced
configuration and management services. There is a great deal of
work underway to ensure that IPv6 hosts can have their addresses
automatically configured and reconfigured in a cost-effective and
manageable way. Automatic address configuration is necessary in
hierarchical routing fabrics because it supports scalable (and thus
cost-effective) numbering and renumbering of large populations of
IP hosts. Even a small renumbering cost, if incurred millions of
times for every change of ISP, adds up to a major administrative
headache. Conversely, scalable renumbering techniques will enable
business enterprises to shop for the best connectivity solutions
without worrying about the renumbering costs of reconnection to a new
provider.
Autoconfiguration capabilities are important whether provider-based
or geographic address allocation is in effect. Occasionally, it
may be necessary to renumber every host within an organization, as
would be the case with a company that relocated its operations (with
geographic addressing) or changed to another service provider (with
provider-based addressing). Configuration of IP addresses is a
constant fact of life at the workgroup and department levels of large
networked organizations. IP addresses need to be configured for
new hosts, for hosts that change location, and for hosts connected
to physical networks that receive address modification (e.g., a
new prefix). In addition to these traditional requirements for
configuration, new requirements are emerging as large numbers of
hosts become mobile.
King, et.al. Expires 13 September 1998 [Page 31]
Internet Draft The Case for IPv6 13 March 1998
The process of autoconfiguration under IPv6 starts with the Neighbor
Discovery (ND) protocol [6]. ND combines and refines the services
provided in the IPv4 environment by Address Resolution Protocol
(ARP) [8] and Internet Control Message Protocol (ICMP) [9]. Although
it has a new name, ND is actually just a set of complementary
ICMPv6 [2] messages that allow IPv6 nodes on the same link to
discover link layer addresses and to obtain and advertise various
network parameters and reachability information. In a typical
scenario, a host starts the process of autoconfiguration by
self-configuring a link-local address to use temporarily [10]. This
address can be formed by adding a generic local address prefix to
a unique token (typically the host's IEEE LAN interface address).
Once this address is formed, the host sends out an ND message to the
address, to ensure that it is unique. If no ICMP message comes back,
the address is unique. If a message comes back indicating that the
link-local address is already in use, then a different token is used
(e.g., an administrative token or a randomly generated token).
Using the new link local address as a source address, the host then
sends out an ND router solicitation request. The solicitation is
sent out using the IPv6 multicast service. Unlike the broadcasted
ARPs of IPv4, IPv6 ND multicast solicitations are not necessarily
processed by all nodes on the link, which can conserve processing
resources in hosts. (IPv6 currently defines several permanent
multicast groups for finding resources on the local node or link,
including an all-routers group, an all-hosts group, and a DHCP server
group). Routers respond to the solicitation messages from hosts with
a unicast router advertisement that contains, among other things,
prefix information that indicates a valid range of addresses for the
subnet. Routers also send these advertisements out periodically to
local multicast groups, whether or not they receive solicitations.
ND message exchange is shown in Figure 16.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 16: ND Message Exchange
King, et.al. Expires 13 September 1998 [Page 32]
Internet Draft The Case for IPv6 13 March 1998
Using the router advertisement message, the router can control
whether hosts use stateless or stateful autoconfiguration methods.
In the case of stateful autoconfiguration, the host will contact a
DHCP or similar address server, which will assign an address from a
manually administered list. DHCP citerfc2131,rfc2132 is becoming
popular for autoconfiguration in IPv4 networks and the standard is
being extended to the IPv6 environment [1, 7].
With the stateless approach, a host can automatically configure its
own IPv6 address without the help of a stateful address server or
any human intervention. The host uses the globally valid address
prefix information in the router advertisement message to create
its own IPv6 address. This process involves the concatenation of a
valid prefix with the host's link layer address or a similar unique
token. As long as the token is unique and the prefix received from
the router is correct, the newly configured IP address should provide
reachability for the host that extends to the entire enterprise and
the Internet at large.
The advantages of stateless autoconfiguration are many. For
instance, if an enterprise changes service providers, the prefix
information from the new provider can be propagated to routers
throughout the enterprise, and hence to all stateless autoconfiguring
hosts. Hypothetically, if all hosts in the enterprise use IPv6
stateless autoconfiguration, the entire enterprise could be
renumbered without the manual configuration of a single host. At a
more modest level, workgroups with substantial move/change activity
also benefit from stateless autoconfiguration because hosts can
receive a freshly configured and valid IP number each time they
connect and reconnect to the network.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 17: Forwarded IP Traffic
To facilitate dynamic host renumbering, IPv6 has a built-in
mechanism to create a graceful transition from old to new addresses.
Fundamental to this mechanism is the ability of IPv6 nodes to support
King, et.al. Expires 13 September 1998 [Page 33]
Internet Draft The Case for IPv6 13 March 1998
multiple addresses per interface. IPv6 addresses assigned to an
interface can be identified as valid, depreciated, or invalid.
In the renumbering process, an interface's address would become
depreciated when a new address was automatically assigned (e.g., in
the case of network renumbering). For a period of time after the new
(valid) address is configured, the depreciated address continues to
send and receive traffic. This allows sessions and communications
based on the older address to be finished gracefully. Eventually
the depreciated address becomes invalid and the valid address is
used exclusively. Multiple IP addresses allow renumbering to occur
dynamically and transparently to end users and applications.
The above described stateless autoconfiguration process is
particularly suited to conventional IP/LAN environments with
48-bit addressing and native multicast services. Other network
environments with different link characteristics may require modified
or alternative configuration techniques. For instance, current ATM
networks do not inherently support multicast services or 48-bit IEEE
addressing, due to the use of virtual circuits and telephony-style
calling numbers. Multicasting solutions for ATM are seen in the
emerging Multicast Address Resolution Server (MARS) that is being
developed for IPv4 multicast over ATM. Plans are being devised to
use MARS-style functionality to extend the IPv6 Neighbor Discovery
protocol across ATM networks. This would allow network renumbering
and stateless autoconfiguration to take place seamlessly in hybrid
ATM/IPv6 fabrics.
3.12. Other Protocols and Services
The preceding discussion focuses on some of the more innovative
and radical changes that IPv6 brings to internetworking. In many
other areas, protocols and services will operate much the same as
they do in the current IPv4 regime. As the industry moves to IPv6,
PPP, DHCP and DNS servers are being modified to accommodate 128-bit
addresses, but in terms of basic functionality, there will be little
change. This is also generally true for interior and exterior
routing protocols.
For example, OSPF is being updated with full support for IPv6,
allowing routers to be addressed with 128-bit addresses. The 32-bit
link-state records of current OSFP will be replaced by 128-bit
records. In general, the OSPF IPv6 link-state database of backbone
routers will run in parallel with the database for IPv4 topologies.
In this sense, the two versions of OSPF will operate as "ships in the
night," just as the routing engines for IPv4, OSI and proprietary
protocols may coexist in the same router without major interaction.
Given the limited nature of the OSPF IPv6 upgrade, those engineers
and administrators who are proficient in OSPF for IPv4 should have no
King, et.al. Expires 13 September 1998 [Page 34]
Internet Draft The Case for IPv6 13 March 1998
problems adapting to the new version. An updated version of RIP is
also available, referred to as RIPng.
As with the interior gateway protocols, work is underway to create
IPv6-compatible versions of the exterior gateway protocols that
are used by routers to establish reachability across the Internet
backbone between large enterprises, providers, and other autonomous
systems. Today's backbone routers use the Border Gateway Protocol
(BGP) to distribute CIDR-based routing information throughout the
Internet. BGP is known by providers and enterprises and has a large
installed base. Currently, work is underway to define BGP extensions
that will allow it to be used to exchange reachability information
based on the new IPv6 hierarchical address space.
4. Part III: Transition Scenarios
Part I of this paper provided an overview of the major transition
mechanisms that are integral to the IPv6 design effort. These
techniques include dual-stack IPv4 /IPv6 hosts and routers,
tunnelling of IPv6 via IPv4, and a number of IPv6 services, including
IPv6 DNS, DHCP, MIBs, and so on. The flexibility and usefulness of
the IPv6 transition mechanisms are best gauged through scenarios that
address real-world networking requirements.
4.1. First Scenario: No Need to NAT
Take, for instance, the case of two large, network-dependent
organizations that must interface operations due to a merger and
acquisition (M&A), or a new business partnership. Both of the
enterprises in this scenario have large IPv4-based networks that have
grown from small beginnings. Both of the original enterprises have a
substantial number of private IPv4 addresses that are not necessarily
unique within the current global IPv4 address space. Combining these
two non-unique address spaces could require costly renumbering and
restructuring of routers, host addresses, domains, areas, exterior
routing protocols, and so on. This scenario is quite common in the
current business climate, not only for M&A projects, but also for
large outsourcing and customer/supplier networking relationships,
where many hosts from the parent, outsourcer, supplier, or partner
must be integrated into an existing enterprise address structure.
For these situations, IPv6 offers a convenient solution.
The task of logically merging two enterprise networks into a single
autonomous domain can be expensive and disruptive. To avoid the
cost and disruption of comprehensive renumbering, enterprises may
be tempted to opt for the stopgap solution of a network address
translator (NAT). In the M&A scenario, a NAT could allow the two
King, et.al. Expires 13 September 1998 [Page 35]
Internet Draft The Case for IPv6 13 March 1998
enterprises to maintain their private addresses in a more or less
status quo fashion. To accomplish this, a NAT must conduct address
translation in real time for all packets that move between the
two organizations. Unfortunately, this solution introduces all
the problems associated with NATs that were discussed in Part I,
including performance bottlenecks, lack of scalability, lack of
standards, and lack of universal connectivity among all the nodes in
the new enterprise and the Internet.
In contrast with NAT, IPv6 seamlessly integrates the two physical
networks (see Figure 18). Suppose the two originally independent
enterprises are known as Enterprise A and Enterprise B. The first
step is to determine which hosts need access to both sides of the
new organization. These hosts are outfitted with dual IPv4/IPv6
stacks, which allow them to maintain connectivity to their original
IPv4 network while also participating in a new IPv6 logical
network that will be created "on top" of the existing IPv4 physical
infrastructure.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 18: IPv6 Unites Private Address Spaces
It's likely that the accounting department of the combined enterprise
will have financial applications on servers that will need to be
accessed by accounting employees in both Enterprise A and Enterprise
B. Both servers and clients will run IPv6, but they will also retain
their IPv4 stacks. The IPv6 sessions of the accounting department
will travel over the existing local and remote links as "just another
protocol," requiring no changes to the physical network. The only
requirement for IPv6 connectivity is that routers that are adjacent
to accounting department users must be upgraded to IPv6 capabilities.
Where end-to-end IPv6 connectivity can't be achieved, one of the
IPv4/IPv6 tunnelling techniques can be employed.
As integration continues, other departments in the newly merged
enterprises will also be given IPv4/IPv6 hosts. As new departments
and workgroups are added, they may be given dual-stack hosts, or in
King, et.al. Expires 13 September 1998 [Page 36]
Internet Draft The Case for IPv6 13 March 1998
some cases, IPv6-only hosts. Hosts that require communications to
the outside world via the Internet will likely receive dual stacks to
maintain compatibility with IPv4 nodes exterior to the enterprise.
But in some cases, hosts that only require access to internal servers
and specific outside partners may be able to achieve connectivity
with IPv6-only hosts. A migration to IPv6 presents the opportunity
for a fresh start in terms of address allocation and routing protocol
structure. IPv6 hosts and routers can immediately take advantage
of IPv6 features such as stateless autoconfiguration, encryption,
authentication, and so on.
4.2. Second Scenario: IPv6 from the Edges to the Core
For corporate users, connectivity requirements typically focus
primarily on access to local e-mail, WWW, database, and applications
servers. In this case, it may be best to initially upgrade only
isolated workgroups and departments to IPv6, with backbone router
upgrades implemented at a slower rate. IPv6 protocol development
is more complete for "edge" routing than for high-level backbone
routing, so this is an excellent way for enterprises to gracefully
transition into IPv6. As shown in Figure 19, independent workgroups
can upgrade their clients and servers to dual-stack IPv4/IPv6 hosts
or IPv6-only hosts. This creates "islands" of IPv6 functionality.
Please refer to the following web page for the figures referenced in
this Internet-Draft.
http://www.6bone.net/case-for-ipv6-figs.GIF
These figures will be replaced by text figures prior to the final
publication of this paper as an Informational RFC.
Figure 19: Islands of IPv6
As enterprise-scale routing protocols such as OSPF and BGP for IPv6
mature, the core backbone IPv6 connections can be deployed. After
the first few IPv6 routers are in place, it may be desirable to
connect IPv6 islands together with router-to-router tunnels. In
this case, one or more routers in each island would be configured
as tunnel endpoints. As described in Part I, when hosts use full
IPv6 128-bit addressing, tunnels are manually configured so that the
routers participating in tunnels know the address of the endpoints
King, et.al. Expires 13 September 1998 [Page 37]
Internet Draft The Case for IPv6 13 March 1998
of the tunnel. With IPv4-compatible IPv6 addresses, automatic,
nonconfigured tunnelling is possible.
From a routing protocol standpoint, tunnels appear as a single IPv6
hop, even if the tunnel is comprised of many IPv4 hops across a
number of different media. IPv6 routers running OSPF can propagate
link-state reachability advertisements through tunnels, just as
they would across conventional point-to-point links. In the IPv6
environment, OSPF will have the advantage of flexible metrics for
tunnel routes, to ensure that each tunnel is given its proper weight
within the topology. In general, routers make packet-forwarding
decisions in the tunnelling environment in the same way that they
make decisions in the IPv6-only network. The underlying IPv4
connections are essentially transparent to IPv6 routing protocols.
5. Security Considerations
A major part of this paper emphasizes the reasons that security plays
into the business and technical reasons for expediting the deployment
for IPv6. By adopting IPv6, the Internet and the enterprise-specific
applications will be much better able to satisfy their security needs
by making use of standardized network features.
6. Full Copyright Statement
Copyright (C) The Internet Society (1998). All Rights Reserved.
This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implmentation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph
are included on all such copies and derivative works. However,
this document itself may not be modified in any way, such as by
removing the copyright notice or references to the Internet Society
or other Internet organizations, except as needed for the purpose
of developing Internet standards in which case the procedures
for copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other than
English.
The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assigns.
This document and the information contained herein is provided on an
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
King, et.al. Expires 13 September 1998 [Page 38]
Internet Draft The Case for IPv6 13 March 1998
BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE."
7. Acknowledgments
This work derives from a Bay Networks white paper on IPv6 (published
in 1997) that was co-authored by Steve King, Ruth Fax, Haskin, Wenken
Ling, and Tom Meehan of Bay Networks.
A. Myths
Because of its importance and the vast number of detailed technical
choices that had to be made, the birth of IPv6 has been attended
by some controversy, and by a number of somewhat misleading myths
that can easily distract network owners who are in the process of
crafting their forward-looking network strategy. Confusion is to
be expected, considering the mammoth implications of migrating our
global internetwork infrastructure to an updated protocol. But if
the IPv6 myths are perpetuated indefinitely, there's a risk that the
Internet will not be able to progress beyond a patched-up version of
IPv4.
Myth #1: The only driving force behind IPv6 is address space
depletion.
Many of the discussions about a new Internet protocol focus on the
fact that we will sooner or later run out of globally unique Network
Layer addresses, due to IPv4's fixed 32-bit address space. The
various address registries that assign blocks of IP addresses to
large network service providers and network operators have become
cautious about the way these addresses are handed out, though most
predictions for IPv4 address exhaustion target a time frame that
starts well into the next decade.
With the long-haul in mind, IPv6 has been outfitted with an enormous
128-bit address space that should guarantee globally unique addresses
for every conceivable variety of network device for the foreseeable
future (i.e., decades). IPv6 has 16 bytes of addressing, or
340,282,366,920,938,463,463,374,607,431,768,211,456
addresses (almost half a duodecillion of them, in fact). The
addressing gets a lot of attention but it is only one of many
important issues that IPv6 designers have tackled. Other IPv6
capabilities have been developed in direct response to current
business requirements for more scalable network architectures,
King, et.al. Expires 13 September 1998 [Page 39]
Internet Draft The Case for IPv6 13 March 1998
mandatory security and data integrity, an additional field for
quality-of-service (QoS), autoconfiguration, and more efficient
network route aggregation at the global backbone level.
Myth #2: Extensions to IPv4 can replicate IPv6 functionality.
The many benefits of IPv6 will not come without a transition effort,
which has led to multiple efforts at extending the life of IPv4
incrementally with evolutionary changes to the protocol standards and
various proprietary techniques. One example of IPv4 extension is
found in network address translators (NAT) that preserve IPv4 address
space by intercepting traffic and converting private intra-enterprise
addresses into one or a few globally unique Internet addresses.
Other examples include the various quality-of-service and security
enhancements to IPv4, which are in general scaled-back or identical
to the equivalent mechanisms in IPv6.
We do not know how long IPv4's life can be extended by these
techniques. What is certain is that the widespread introduction
of NAT devices has extremely negative effects on the end-to-end
behaviour of emerging Internet applications; in practice only a
limited set of well-known applications can be correctly handled
by NAT devices or by application level gateways associated with
them. In particular NAT devices prevent the deployment of end-to-end
IPv4 security. Furthermore, the development of new and innovative
Internet applications is burdened with the design constraints posed
by NATs. Since NAT is strictly unnecessary for IPv6, standard
end-to-end IPv6 security can be deployed, and a future enlivened by
lighter weight and more fully functional new applications can be
envisioned. Network address translation is also known to create
great difficulty in the construction of complex Extranets and Virtual
Private Networks, since it turns address space administration into a
nightmare.
NAT also only works in a "flat universe" for a site accessing the
"world" - any realistic size enterprise is not flat even internally,
and its relationships are also nested, and multi-party. Realistic
NAT deployment solutions would have to include routing via multiple
ingress/egress NATs for load balanacing, multi-NAT-hop routes and
so on - all this would create in miniature the v4 (or in fact v6)
architecture, since it is solving the same problem, but piecewise and
badly.
It is hard to compare the costs of converting to IPv6 with those of
remaining with IPv4 and its ongoing fixes. Every network manager
will have to make this comparison; but staying with IPv4 has been
likened to the situation of a lobster in a pot of water, as the
temperature slowly increases - at first, it feels comfortable.
King, et.al. Expires 13 September 1998 [Page 40]
Internet Draft The Case for IPv6 13 March 1998
Myth #3: IPv6 support for a large diversity of network devices is
not an end-user or business concern.
Over the next few years, conventional computers on the Internet will
be joined by a myriad of new devices, including palmtop personal data
assistants (PDA), hybrid mobile phone technology with data processing
capabilities, smart set-top boxes with integrated Web browsers, and
embedded network components in equipment ranging from office copy
machines to kitchen appliances. Some of the new devices requiring
IP addresses and connectivity will be consumer-oriented, but many
will also become integral to the information management functions of
corporations and institutions of all sizes.
IPv6's 128-bit address space will allow businesses to deploy a huge
array of new desktop, mobile, and embedded network devices in a
cost-effective, manageable manner. Further, IPv6's autoconfiguration
features will make it feasible for large numbers of devices to attach
dynamically to the network, without incurring unsupportable costs
for the administration for an ever-increasing number of adds, moves,
and changes. The business requirement for IPv6 will be driven by
end-user applications. To remain competitive in the coming era of
high-density networking, businesses should exploit IPv6 to create a
highly scalable address space and robust autoconfiguration services
that will remain viable in the face of an explosion of end-user
networking needs.
Myth #4: IPv6 is primarily relevant to backbone routers, not
end-user applications.
It is true that IPv6 address aggregation allows efficient multitiered
routing hierarchies that limit the uncontrolled growth of backbone
router tables. But many of the advanced features of IPv6 also
bring direct benefits to end-user applications at the workgroup
and departmental levels. For instance, applications will have
available the mandatory IPv6 encryption and authentication services
as an integral part of the IP stack. For mobile business users and
changing organizations, the automatic configuration components of
IPv6 will allow the efficient assignment of IP addresses without the
delays and cost associated with manual address administration or even
traditional DHCP, which takes place in many current IP networks.
IPv6 is very much both an end-user concern and a business concern.
Myth #5: Asynchronous Transfer Mode (ATM) cell switching will negate
the need for IPv6.
ATM and other switching methods offer interesting technology for
present and future internetworks, but ATM is, by itself, not a
replacement for today's packet routing, Internet architecture.
Fortunately, network owners do not have to make a choice between ATM
King, et.al. Expires 13 September 1998 [Page 41]
Internet Draft The Case for IPv6 13 March 1998
or IPv6 because the two protocols will continue to serve different
and complementary roles in corporate networking. Large networks will
make use of both protocols. For many network designers, ATM is a
useful transmission medium for high-speed IPv6 backbone networks.
A great deal of standards and development work is being devoted to
integrating ATM and IPv6 environments. IPv6, like its predecessor
IPv4, provides Network Layer services over all major link types,
including ATM, Ethernet, Token Ring, ISDN, Frame Relay, and T1.
Myth #6: IPv6 is something that only large telco companies or the
government should worry about.
Some in the industry have characterized IPv6 as a concern that's
outside the corporate network and outside the current time frame.
In reality, IPv6 is a standards track and mainstream solution
for the operation and continued efficiency of day-to-day business
activities. But the only way that IPv6 will take hold and succeed is
if businesses and institutions of all types come to terms with the
inadequacies of IPv4 and begin to lay plans for migration. In the
past few years, Internet protocols have enabled a whole new style of
distributed commerce that brings people together inside enterprises
and gives enterprises access to the entire world. In fact, the
sustained and impressive growth of the Internet, which has inspired
the current engineering efforts for IPv6, is in large measure due to
the penetration of the World Wide Web to business and consumer end
users. Offering services to such end users is of interest to many
more institutions than merely governments and telephone companies.
King, et.al. Expires 13 September 1998 [Page 42]
Internet Draft The Case for IPv6 13 March 1998
References
Please refer to the following web page for the IPv6 technical
specifications referred to in this Internet-Draft.
http://playground.sun.com/pub/ipng/html/specs/specifications.html
These references will be replaced by specific references prior to the
final publication of this paper as an Informational RFC.
References
[1] J. Bound and C. Perkins. Dynamic Host Configuration Protocol
for IPv6. draft-ietf-dhc-dhcpv6-11.txt, May 1997. (work in
progress).
[2] A. Conta and S. Deering. Internet Control Message Protocol
(ICMPv6) for the Internet Protocol Version 6 (IPv6). RFC 1885,
December 1995.
[3] V. Fuller, T. Li, J. Yu, and K. Varadhan. Classless
Inter-Domain Routing (CIDR): an Address Assignment and
Aggregation Strategy. RFC 1518, September 1993.
[4] R. Hinden and S. Deering. IP Version 6 Addressing Architecture.
RFC 1884, December 1995.
[5] D. Johnson and C. Perkins. Mobility Support in IPv6.
draft-ietf-mobileip-ipv6-04.txt, November 1997. (work in
progress).
[6] T. Narten, E. Nordmark, and W. Simpson. Neighbor Discovery for
IP version 6 (IPv6). RFC 1970, August 1996.
[7] C. Perkins. Extensions for the Dynamic Host Configuration
Protocol for IPv6. draft-ietf-dhc-dhcpv6ext-09.txt, October
1997. (work in progress).
[8] David C. Plummer. An Ethernet Address Resolution Protocol:
Or Converting Network Protocol Addresses to 48.bit Ethernet
Addresses for Transmission on Ethernet Hardware. RFC 826,
November 1982.
[9] J. B. Postel, Editor. Internet Control Message Protocol. RFC
792, September 1981.
[10] S. Thomson and T. Narten. IPv6 Stateless Address
Autoconfiguration. RFC 1971, August 1996.
King, et.al. Expires 13 September 1998 [Page 43]
Internet Draft The Case for IPv6 13 March 1998
Authors' and Editors' Addresses
Original Authors
Steve King, Bay Networks
Ruth Fax, Bay Networks
Dimitry Haskin, Bay Networks
Wenken Ling, Bay Networks
Tom Meehan, Bay Networks
Questions about this memo can be directed to the editors:
Robert Fink
Esnet R&D
Lawrence Berkeley National Laboratory
1 Cyclotron Road
Bldg. 50A, Room 3139
Mail-Stop 50A-3111
Berkeley, CA 94720
USA
+1 510 486-5692 office
+1 510 486-4790 fax
rlfink@lbl.gov
Charles E. Perkins
Mail Stop MPK15-214
Room 2682
Sun Microsystems, Inc.
15 Network Circle
Menlo Park, CA 94025
USA
ph# +1-650-786-6464
fax# +1-650-786-6445
email: charles.perkins@Sun.COM
charles.perkins@Eng.sun.com
cperkins@Eng.sun.com
Web: http://www.svrloc.org/~charliep
King, et.al. Expires 13 September 1998 [Page 44]
| PAFTECH AB 2003-2026 | 2026-04-23 12:27:26 |