Introduction

When communication resources were precious, it was natural to design a special method for better utilization of these resources. Thus, for a long time, many information network infrastructure providers developed their own network designs and protocols. The recent wide deployment of Internet Protocol (IP) technology provides a simple communication framework for any kind of information infrastructure, and it is integrating all information infrastructures into one protocol—IP.

The evolution first occurred for the wired infrastructure because the wired networks had a faster communication property than the wireless networks. Whereas many wired network infrastructures changed their dedicated network designs and protocols to the generic IP-based system, the wireless infrastructures kept their own designs. The wireless infrastructure, which had a slower communication property, could not accept the overhead of the generic protocol, even though having a common protocol had many benefits, such as interoperability, simplicity, and cost performance.

Recently, however, advances in wireless communication technology have resulted in much wider broadband infrastructures for the wireless environment than in the past. The IEEE 802.11- based technology will soon provide 600 Mbps communication speed, IEEE 802.16 (WiMAX) technology provides more than 70 Mbps with an approximately 50-km communication range, and IEEE 802.16e (Mobile WiMAX) provides more than 20 Mbps communication speed for moving nodes. There is no doubt that future wireless technology will provide much faster communication properties. They are still narrower than those of wired communication devices; however, the overhead of using IP over them is no longer a major issue.

Although the history of mobility technology research and development is quite long, the technology is still not widely deployed. We now have many mobile devices, such as laptop computers, PDAs, and mobile phones, but none of them currently use IP mobility technology. One of the reasons is that the wireless communication technology has not provided the required bandwidth and quality as described previously. Another reason is that we have not had a mobile-ready environment to apply IP mobility technology.

The situation is now drastically changing. In the past, we could not utilize the full advantages of IP mobility technology, even though we had the mechanism to do so. The goal of an unwired mobile Internet world will be achieved with the combination of recent advanced communication technology and the long-researched and -developed IP mobility framework.

1.1 History of IP Mobility

IP mobility protocol is not a special feature for Internet Protocol version 6 (IPv6). The mobility support protocol for IPv4 (Mobile IP) [RFC3344] also has a long history. The initial proposal of the mobility support protocol for IPv4 was presented in 1993. At that time, there were no real “mobile” computers. There were some small computers called laptops, but they were still relatively large and they were very expensive compared to desktop computers. Mobile phones were in use, but they were large and had poor computing resources. Even with this level of technology, IP engineers were trying to provide mobility support for computer devices as if they were foreseeing the future. Mobile IPv4 was finally standardized as RFC2002 (the latest revision is RFC3344) in 1996. In the late 1990s, the Internet era began. Some pioneers started commercial services to provide Internet connectivity. Many companies and universities started providing their information and services over the Internet. Individual users soon followed, and the Internet became the largest information network in the world. Unfortunately for Mobile IPv4, the communication technology, especially the wireless communication technology, was still poor at that time. Although the protocol could support handover from one network to another, we could not use networks in this manner. Mobile IPv4 is now used in the backend system of some mobile telephone service networks. In that sense, it is deployed in the real service network, but we still do not see Mobile IP devices near us and the benefit we receive is limited.

The discussion of Mobile IPv6 [RFC3775] started in 1996. The initial action of the standardization process of IPv6 mobility was very quick. Considering that the first draft of the IPv6 protocol specification was submitted in 1995, the discussion of IPv6 mobility was started almost at the same time as that of IPv6. However, the standardization of Mobile IPv6 was a thorny path. The final specification of Mobile IPv6 was published as RFC3775 in 2004. By contrast, from the first draft to publication as request for comment (RFC), Mobile IPv4 required only 3 years. Recently, the period required to publish a specification as an RFC has become increasingly longer, but 9 years is a surprisingly long time. The draft specification was revised 24 times before it was published as an RFC.

The first turning point of the Mobile IPv6 standardization process was its 13th draft in 2000. The Mobile IP working group reached a consensus on the specification and the 13th draft was submitted to the Internet Engineering Steering Group (IESG) for final review and publication as an RFC. However, the IESG rejected the specification.

Mobile IPv6 was trying to solve one major problem of Mobile IPv4—the path optimization mechanism between a mobile node and its communicating node. Mobile IP is a kind of automatic tunnel establishment protocol. The moving node registers its current location to the proxy node called the home agent. All the packets are forwarded once to the home agent and then sent to the final destination. Apparently, if the mobile node and its communicating node reside nearby and the home agent is located far away, the communication path becomes long and redundant. The Mobile IPv4 base protocol does not mention the optimization mechanism for this case. The Mobile IPv6 specification includes the optimization mechanism from the first draft of the specification. In the mechanism, the mobile node sends its current location to its communicating node. The problem concerns how the communicating node verifies the message sent from the mobile node. If there is no authorization mechanism of the message, any node can send a bogus message to the communicating node. If a malicious node sends such a request using the identifier of the mobile node, then all the data sent from the communicating node to the mobile node are stolen by the malicious node. The Mobile IPv6 specification before the 14th draft was using the IPsec mechanism to protect the message. However, it is usually difficult to set up IPsec parameters between two random nodes. IESG pointed out the difficulty of the IPsec setup process and judged that the specification was not feasible.

After receiving the rejection message from IESG, the working group invented a new mechanism to protect the message. The 14th and 15th drafts proposed a shared secret-based authorization mechanism. It was simple and easy to understand; however, the problem of how the two nodes share the secret remained. In 2002, the 16th draft introduced a completely new mechanism called the return routability mechanism to authorize the message. By using the mechanism, a mobile node and its communicating node can generate secret information with several messages exchanged between them before sending the notification message from the mobile node to register its current location. The detailed procedure of the return routability mechanism is explained later.

Finalizing the specification required only 2 years after the 16th draft. The final draft was published in 2003, and it became RFC3775 in 2004.

1.2 Benefit of IP Mobility

Mobile IP provides a mobility function to IP devices. But what is mobility? When we say “mobility,” it implies that there are many different levels of mobility. For example, a cellular network can provide a mobility function to cellular phones. We can use our cellular phone almost everywhere with the same communication identifier—phone numbers. We can even establish an IP connection over the cellular network by using the dial-up connection function. Isn’t this mobility? Another example is the e-mail system. We send an e-mail using a fixed identifier, such as bob@example.com. Wherever Bob is, the message will be delivered to the mailbox associated with the mail address bob@example.com, and Bob can retrieve the message independent of his location. It is a kind of mobility.

Figure 1-1 shows various levels of mobility support. SIP (Session Initiation Protocol) [RFC3261] is a session-layer protocol that establishes an application session between two application entities. Because it is independent of the actual location of the terminal on which the application is running, it can be considered as a mobile protocol in the session layer. SCTP (Stream Control Transport Protocol) [RFC2960] is a new transport protocol aiming to replace TCP (Transmission Control Protocol). It is defined on top of the IP layer and supports the IP address migration function while keeping the transport connectivity. HIP (Host Identity Protocol) [RFC4423] is another IP-layer mobility protocol. Unlike Mobile IP-based protocols, HIP is a completely new protocol to pursue the ideal mobility support in the IP layer. The design is cleaner than Mobile IP-based protocols; however, it does not have compatibility with the existing IPv4/IPv6 stacks, whereas Mobile IP-based protocols do. As demonstrated in Figure 1-1, the more we focus on the lower-layer technology, the more device and...