THE ADOPTION OF IPv6 *

THE ADOPTION OF IPv6 * STUDENT PAPER Brian Childress Southwest Texas State University BC56075@swt.edu Bryan Cathey Southwest Texas State University BC1033@swt.edu Sara Dixon Southwest Texas State University SD1009@swt.edu ABSTRACT There is no question the Internet has revolutionalized the world. Many of the things we take for granted today were not even possible a few years ago. Will the Internet continue to push the boundaries of technology? The answer lies within IPv6. IPv6 will replace IPv4 as the standard addressing scheme of the Internet. Why do we need a new addressing scheme? There are numerous amounts of limitations in IPv4, namely, the shortage of addresses. Most countries, excluding the United States, do not have nearly as many IP addresses as they need. The United States owns 70% of the available IP addresses, so it is no surprise that the United States is not pushing for a quick transition to the new technology. This is all new material for professional technologists. The first implementation was IPv4, which built the addressing scheme, but now IPv6 is the first step to revise and expand it. This will take extraneous time and perseverance to complete the Internet Protocol next generation. The European Union and Japan are leading the transition to IPv6. This is inevitable since they rely on the Internet as much as the Americans, yet own significantly fewer IP addresses. IT professionals question the theory of which to upgrade first the Domain Name Server (DNS), routers, or hosts. There is * Copyright 2003 by the Consortium for Computing in Small Colleges. Permission to copy without fee all or part of this material is granted provided that the copies are not made or distributed for direct commercial advantage, the CCSC copyright notice and the title of the publication and its date appear, and notice is given that copying is by permission of the Consortium for Computing in Small Colleges. To copy otherwise, or to republish, requires a fee and/or specific permission. 153

JCSC 18, 4 (April 2003) not an IPv6 DNS root server. Presently, the DNS implementations all run on top of IPv4 addresses and the DNS system that supports IPv6 is linked to IPv4 information. Some DNS implementations are beginning to support national IPv6 transitions, such as, bind8 with a KAME patch, or newbie and bind9 (which are still under development). Upgrading these servers is a crucial step in the conversion process. Asia is in the process of implementing 6Bone, a virtual network layer that allows IPv4 and IPv6 to coexist. This is a major requirement for a smooth transition to IPv6. Without an intermediate layer like 6Bone, the only alternative would be to change every network in the world over to IPv6 on a specified day at a specific time-not a very ideal situation. Japan's 6Bone will take 4-5 years to complete, not including additional time to debug and fix errors. 1.1 INTRODUCTION Extensive research, design, and organization has been put forth by the European Union (EU) and Japan in the quest to implement Internet Protocol version 6 (IPv6). Both efforts will be collaborated into finalizing the new generation of Internet protocol. The European Commission's IPv6 task force and the IPv6 promotion council of Japan said in a joint statement that they will cooperate "to foster promotion and deployment and garner support for the new generation Internet Protocol."[2] The European Union is completing its second phase, contracting international cooperation agreements to help set up IPv6 task forces at national and regional levels. They recommended deployment strategy by network designers is to begin at the edge and then move towards the network core reducing costs and operational impacts of integration. [3] Next is the deployment of IPv6 throughout Europe by a 2005 due date. [2] With Japan's construction of 6Bone and other useful software to implement the change this should not be a challenge. We all know the benefits from the new protocol. We do not know the problems it has in store. A major concern is the ongoing Questions "What securities will it provide?" or "Will it cause more intense types of failures?"[7] Truth is no one knows. IPv6 is being implemented and there is no telling what to expect. Experts are embarking upon new and untainted ground. They are just starting the foundation and as the past has shown there are many complications waiting to be discovered as well. And where is the United States in all of this commotion? Since the US still has quite a few Internet protocol address left there isn't a demand for the implementation of more.[3] However, the US does have its hand in developing IPv6; in fact Microsoft is one of the leading designers for the switch. Microsoft, Cisco, and others have a few of its top computer experts overseas assisting international efforts for developing the future's technology. [3] In some perspective, the US is waiting to see what problems will occur from the new protocol and research its effects. Then when all of the flaws are squared away the US will play and intense game of catching up, but this is only a theory. Another decision which has caused numerous headaches is what to implement first. The domain name servers (DNS) deal with the most transport of data so they were chosen first to upgrade. [6] The IPv4 address scheme will be compacted into the IPv6 address. With the IPv6 addressing scheme there is an excess of room to be filled. There is an added section of the IPv6 address to deal with telling the server what type of protocol to handle. This allows the servers to communicate with dual 154

CCSC: South Central Conference IPv4/IPv6 encoding of the data. [6] This process changes data from IPv4 by adding zeroes in the address or just by putting zeros in the empty spaces. The server will pick up these zeros or emptied space colons (which are used to shorten a lengthy amount of zeros) and read the address as an IPv6. [6] Then normal computations will take place. With the next generation of protocol intact, the possibilities are endless to what computers can accomplish. 2.1 DOMAIN NAME SYSTEM The domain name system makes it possible for the Internet to be convenient for the average user. Without the DNS, web users would be responsible for keeping track of their favorite websites by memorizing their numerical addresses. Imagine an advertisement; Visit our web site at 128.34.43.118!"[7] Even worse, imagine an IPv6 advertisement without the services of a DNS, "Visit our web site at 3FFE:0B00:0C18:0001:0290:27FF:FE17:FC1D!"[7] Obviously, Domain Name System servers are crucial to the continued popularity and, more importantly, to the effectiveness of the internet. [3] This is why the Domain Name System servers must be the first physical devices that transition to IPv6. [6] This is a logical assumption because in order to have IPv6 address lookup, there must be an IPv6-compatible DNS sever to answer. Coexistence between IPv4 and IPv6 is the most basic underlying assumption of the transition to IPv6. In order to upgrade DNS servers to be compatible with IPv6, a new resource record type, 'AAAA', that will handle queries by IPv6 hosts, has been created. [4] These IPv6 'AAAA' records are equivalent to IPv4 'A' records. Compatible DNS servers will be capable of handling IPV6 'AAAA' address lookup. It is important to mention that there is not a DNS root server that is accessible through IPv6.The implementations currently installed are all running on top of IPv4. [5] Once the DNS has been upgraded to support 'AAAA' records, it will then be possible for lower level routers and hosts to upgrade to IPv4/IPv6 compatibility. [5] This interoperability will also require DNS resolver libraries to be IPv6 and IPv4 compatible. This means the DNS tables can store and look up 'AAAA' and 'A' records.[5,6] The resolver library must choose from three options when returning a dual node address, return only the IPv6 address, return the only the IPv4 address, or both. [4] However, if automatic tunneling is not supported by the DNS, then it would not be able to return an IPv4 compatible IPv6 address, because this is only possible with tunneling. Most importantly, the DNS server must be the first physical implementation in the transition to IPv6. After an IPv6 compatible DNS is established, nodes running IPv6 exclusively can interact with dual nodes and IPv4 nodes. How will a dual nod host know which protocol to use? When the DNS returns a 32-bit address, it will know to use IPv4. If the DNS returns a 128-bit address, the host will know to use IPv6. [1] one of the great benefits of IPv6 in general is its ability to auto configure. Dynamic DNS updates, which link IPv6 auto configuration to IPv6 DNS, allow DNS servers to be updated when an IPv6 node is renumbered. 2.2 Current Implementations Currently, the largest implementation of IPV6 is 6Bone, once a virtual private network turned worldwide collaboration that was initially created to test standards and implementations. 155

JCSC 18, 4 (April 2003) [1] Now, the focus is on testing the transition and the operational procedures. Administrators are connection IPv6 networks to 6Bone to get a jump on the competition and learn the problems that will encounter when they actually switch their networks over to IPv6. Pseudo top level aggregation identifiers and pseudo next level aggregation identifiers (ptla's and pnla's) are being used on 6Bone. [1] The Internet Assigned Numbers Authority (IANA) allocated some of the actual IPv6 addresses for testing on 6Bone. [1] These ptla's and pnla's define the underlying backbone a system is working on. 3.1 TUNNELING There are several strategies to deploy IPv6. The five covered in this paper are deploying IPv6 over dual stack. Backbones, deploying IPv6 over IPv4 tunnels, deploying IPv6 over dedicated data links, deploying IPv6 over multiprotocol label switching (MPLS) backbones, and deploying IPv6 is using protocol translation mechanisms. [6] 3.2 Deployment IPv6 Over Dual Stack Backbone A network can choose to upgrade a portion of their routers, such as CPE routers and aggregation routers, to a have dual stack interoperability. This allows IPv4 and IPv6 to coexist in a dual IP layer routing backbone. Dual stack end systems allow applications to upgrade one at a time to IPv6. Those that are not upgraded can coexist with upgraded applications on the same system. Upgraded application can still make use of IPv4 protocol stack. An application that supports both protocols requests all available addresses for the destination host name from a DNS server. The DNS server replies with all available addresses. [4] In most cases IPv6 is the default and the computer will be connected using this protocol stack. Limitations to this approach are that the routers requires a dual addressing scheme to be defined, require dual management of the IPv4 and IPv6 routing protocols, and must be configured with enough memory for both the IPv4 and IPv6 routing tables.[6] 3.3 Deploying IPv6 Over IPv4 Tunnels These tunneling techniques include manually configured tunnels, generic routing encapsulation tunnels, and semiautomatic tunnel mechanisms such as tunnel broker services. [2] Proper tunneling requires that the endpoints must run in dual stack mode.[6] manually configured tunnels are used to provide a stable and secure connection for regular communication between two edge routers, or an end system and an edge router, or for connection to remote IPv6 networks such as 6Bone. [2] Each tunnel is independently managed, the more tunnel endpoints you have, the more tunnels you need, and the greater the cost of management overhead. The GRE tunneling technique is designed to provide the services necessary to implement any standard point-to-point encapsulation scheme. It has the same endpoint router requirements and increases the management costs as manual configuration. [6] The tunnel constructed on the destination is automatically determined by the IPv4 address. The tunnel source and destination are automatically determined by the IPv4 address. The 156

CCSC: South Central Conference requirement of the IPv4 address removes the benefits of the large IPv6 addressing space. [5] The automatic 6 to 4 tunnel allows isolated IPv6 domains to be connected over an IPv4 network and allows connections to remote IPv6 networks, such as 6Bone. [2] The 6 to 4 tunnel treats the IPv4 infrastructure as a virtual no broadcast link using an IPv4 address embedded in the IPv6 address to find the other end of the tunnel. One 6 to 4 address assignment is necessary for the external interface of the router. All sites need to run an IPv6 interior routing protocol for routing IPv6 within the site; exterior routing is handled by the relevant IPv4 exterior routing protocol. Teredo Tunneling is defined for the case where NAT devices can't be upgraded to offer native IPv6 routing or act as a 6 to 4 router. The mechanism used provides IPv6 connectivity to nodes located behind one or more IPv4 NAT by tunneling IPv6 packets over the User Datagram Protocol (UDP) through NAT devices. 3.4 Deploying IPv6 Over Dedicated Data Links Data links enable IPv6 domain to communicate by using the same Layer 2 infrastructure used for IPv4, but with IPv6 using separate Frame Relay or ATM permanent virtual circuits (PVC), separate optical links, or dense wave division multiplexing (DWDM).[5] Most WAN and MAN have been implemented using this technology. This configuration has the added benefit for the service provider of no loss in service or revenue for the IPv4 traffic. 3.5 Deploying IPv6 Over MPLS Backbone This allows isolated IPv6 domains to communicate with each other, but over and MPLS IPv4 backbone without modifying the core infrastructure. [4] Multiple techniques are available at different points in the network, but each requires little change to the backbone infrastructure because forwarding is based on labels rather than the IP header itself. [5] Three strategies are under development for this use of tunnels on the customers edge (CE) routers, over a circuit transport over MPLS, or on the provider edge (PE) routers. 4.1 PROTOCOL TRANSLATION MECHANISMS Some organization might not want to implement any of these IPv6 strategies. Under the circumstances, intercommunication between IPv6 only hosts and IPv4 only hosts requires some level of translation between the IPv6 and IPv4 protocols on the host or router, or dual stack hosts, with an application level understanding of which protocol to use. [4] Considering, and IPv6 only network might still want to be able to access IPv4 only resources, such as and IPv4 only web server. The following translation mechanisms are under consideration by IETF NGtrans working group: Network Address Translation Protocol Translation (NAT-PT, TCP-UPD relay, Bump in the stack (BIS), Dual Stack Transition Mechanism (DSTM), and SOCKS-based Gateway. [6] The two categories separate the two by those that require no changes to either the IPv4 or IPv6 host, such as TCP-UPD, and those that do require a change. 157

JCSC 18, 4 (April 2003) 5.1 SUMMARY The development of IPv6 is underway with a flexible time table for the actual upgrade. Eventually it will become more of a necessity to upgrade. Large and small companies will implement the changes to improve security, scalability, and support. The root of the DNS is currently being established but they are not global. Each site will choose a plan to implement over their network. The strategy of starting on the edge allows companies to focus on the applications, which are still being assembled. The process will require many patches to software as we learn through the testing now in place. The inevitability of this upgrade will surely change our entire lives as we move forward to having home networks to automate everyday tasks. As professionals push the borders of technology by the use of IPv6 new designs of systems, software, and networks will unmask what the future has in store. REFERENCES [1] "NTT Communications to Provide Global IPv6 Internet Access Service", http://www.ntt.com, April 2001. [2] Meller, Paul, "Europe, Japan join forces on IPv6 adoption", http://www.idg.net, September 2002. [3] Berger, Matt, "Gates, academics join on security, shared source", http://staging.infoworld.com, July 2002. [4] Buclin, Bertand, "IPv6 DNS Setup Information", http://wwwisi.edu, January 2000. [5] "ABCs of IPv6", http://www.cisco.com, August 2002. [6] "IPv6 DNS Settings", http://www.ietf.org, RFC1884. [7] Zaborav, Dev, "Hack the Fridge", http://www.itworld.com, August 2002. 158