The Status Quo and Future of Mobile Internet Technologies

With the rapid development of information network technologies, especially with new network technologies emerging, the demand for information has changed in its contents and acquisition means[1]. People are no longer content with connecting to the Internet via stationary terminals or single mobile terminals, but hope to acquire information from the Internet while in motion, with certain moving sub-networks or mobile terminals (e.g. the sub-networks for marching troops, aircrafts in space flight, ships in voyage and vehicles and trains in motion) and in a means that is relatively stable and reliable. Hence the research on mobile IP, mobile networks and mobile/wireless routers has become a hot issue at present. On the other hand, it is a likely trend that IPv6[2], as the standard of next generation Internet, might replace existing IPv4[3]. In such a situation, researches on next generation IPv6 mobile/wireless networks, routers and mobile IPv6 technologies become especially necessary and pressing.

1 The Status Quo of Researches
From their meanings perspective, early mobile internetworking theories and technologies had two main achievements: one was proposed by John Ioannidis, et al. (from University of Columbia, USA) in 1991, and called the Columbia Mobil IP[4]. It adopts virtual mobile sub-networks and IP in IP tunnel encapsulation. Later, Johen Ioannidis further consummated the design strategy and method of Columbia Mobile IP[5]. The other was the Mobile Node Protocol, also called Virtual IP (VIP )[6-8], which was issued by Fumio Terqoka, et al. from SONY. It uses special routers to remember issues about mobile nodes, and defines a new IP header option to transfer data. Later, C. Perkins and Y. Rekhter from IBM also designed a mobile node protocol that makes use of the existing IP loose source routing[9,10]. In 1994, A. Myles and C. Perkins integrated the above three mobile node protocols to create a new mobile node protocol, which was later developed to the current RFC 3344 protocol for mobile IP[11] by the Internet Engineering Task Force (IETF). In 1996, IETF consecutively issued mobile host support protocols in IPv4, including RFC2002 (IP Mobility Support)[12], RFC2003 (IP Encapsulation within IP)[13], RFC2004 (Minimal Encapsulation within IP)[14], RFC2005 (Applicability Statement for IP Mobility Support)[15] and RFC2006 (the Definitions of Managed Objects for IP Mobility Support Using SMIv2)[16]. These documents briefly summarized previous research results on mobile IP, and established a foundation for related research work. In 2003, IETF released a new version of IPv4, RFC3344, and superseding RFC2002.

  In the meantime, IETF set up 4 working groups in charge of related theoretical research and protocol standardization in 2003 according to research trends of IP network mobile routing. These 4 groups were the IPv4 Mobility Working Group, IPv6 Mobility Working Group, MIPv6 Signaling and Handoff Optimization and Network Mobility, which greatly accelerated the development in this domain.

  With the adoption of IPv6 as the next generation network protocol, applying mobile IPv4 research achievements into mobile IPv6 protocol design and performance enhancement became an important aspect of the research. In 1996, IETF released the first mobile IPv6 standard draft, and till the beginning of 2004, the IPv6 host mobility protocol had already reached version 24, and was released in June 2004 as RFC3775[17], the first mobile IPv6 Standard. Mobile IPv6 utilized IPv6 auto-configuration[18] and optimized packet header and extension options[19], simplified the design of host mobility protocol, solved the problems of IPv4 entry filtering and triangle routing, lowered network overhead, and enhanced working performance.

2 Problems Encountered
  (1) Current Internetworking topology theories and protocols of stationary networks cannot meet the demands of brand-new mobile Internet.

  Traditional Internet was originally designed for data communication, and its network theories and protocols were only fit for networks with relatively fixed topologies. Mobile Internet raised further requirements in adaptability, robustness, reliability and QoS aspects of the routing theories and protocols, which might be used to carry various services such as data, voice and video. The traditional routing theories and protocols cannot fit well with this for the following reasons. Firstly, the routing protocols executed by core routers of traditional Internet are mainly routing IP packets based on the maximum prefix matching principle. After the moving body changes its location, this routing forwarding method may cause false delivery of IP packets. When the moving body moves to a new location, the IP packets it sends cannot be transferred by the network due to issues like source address filtering if the IP addresses within the permitted address range of the current network topology are not used as the source addresses. Even if the moving body solves the source address problem using new network access address, communication interruption is still possible because the network identifies communication connection using the IP address and port number.

  Secondly, although the widely used dynamic routing protocols like the Routing Information Protocol (RIP)[20], Open Shortest Path First (OSPF) protocol[21] and  Border Gateway Protocol (BGP)[22] can adapt to local network topological changes, but the problems caused thereafter, such as management overhead and routing convergence delay, mean that those methods are not acceptable or even not applicable for mobile networks, owing to frequent massive topology changes, wide availability of wireless access and processing capability restrictions of the moving body.

  (2) Current mobile Internet theories and protocols mainly apply to mobile terminals, lacking support for dynamic connection of moving networks (mobile routers), not to mention the problem of fast and frequent topology changes.
To provide the Internet with mobility support, IETF brought out the mobile IP protocol RFC3344 based on IPv4 networks. In order to overcome the limitations of IPv4 in address resources, security and routing efficiency, IETF designed mobile IPv6 based on the IPv6 protocol, and released a series of drafts and standards. These two versions of the mobile IP protocol both uses proxy and tunneling techniques, provides the capability to transceive IP packets in motion by setting binding entries for the visiting address and home address of the mobile IP terminal. It is a typical terminal motion oriented solution, which is apparently not suitable for dynamic network changes. Setting mobile IP routes for all hosts within a moving sub-network will ignore the fact that the moving body is a relatively stable aggregation. It will cause waste of resources by sending mass location management packets and setting up independent wireless links for every terminal. In some locations where communication security and electromagnetism compatibility are of a crucial consideration, it might even be prohibited to do so. Furthermore, mobile IP is merely an extension to traditional Internet. Its topology is relatively stable, and will not affect currently widely used routing protocols such as RIP, OSPF and BGP. But mobile Internet is essentially a dynamically changing network, which requires radical changes to the above routing theories and protocols because its sub-networks or terminal collections dynamically change their connection to the network. Current mobile IP cannot deal with it at all. Besides, although Mobile Ad Hoc[23] aims at supporting the moving of individual hosts in an environment without any wireless infrastructure, it cannot meet the routing demands of dynamic changing networks.

  (3) The characteristics of mobile Internet require that brand-new routing theories and protocols be created.
Mobile Internet is a brand-new network theory and technology, differing from stationary networks in link forms, access methods, topology, communication and application patterns. It is characterized by non-stable changing topologies, multimedia services bearing and a mass of wireless links. Its moving objects might be either some sub-networks in the network prefix context or relatively stable terminal collections formed in a period of time owing to such restrictions as physical location. Hence, we might not copy current theories and research achievements, but should make adjustments according to the characteristics of its network topology and applications, or even find new replacement theories and protocols. Research on the routing theories and protocols of dynamically changing networks shouldn’t be constrained to addressing and IP forwarding. Deep research into this brand-new network is necessary.

  In particular, an effective description theory and method of system architecture should be created; addressing and routing problems should be solved; research into reliability, robustness, compatibility and working efficiency should be done according to the characteristics such as lack of link bandwidth, unstable performance and intense traffic distribution changes within the networks; new routing protocols are requested to provide dynamically changing networks with routing support of seamlessness, low latency, low packet drop rate and low overhead, which helps avoid the processing bottleneck caused by operation of the new protocols. This is a very meaningful and forward-looking systematic work. A set of original theories and protocols will be created instead of simply extending existing routing protocols, which will have a significant influence on the development of this discipline.

  (4) The continuous progress of Internet technologies makes it possible to do research on mobile Internet theories and protocols.

  In designing the mobile Internet, three elements must usually be taken into account: wired and wireless transmission techniques for application host collections or mobile sub-networks; wireless techniques for connecting the host collections and mobile sub-networks to the Internet; theories and mechanism for packet routing between the host collections or mobile sub-networks. In fact, Wireless Local Area Network (WLAN) local wireless transmission techniques like the IEEE 802.11x series and Bluetooth might implement the first element, while Wireless Wide Area Network (WWAN) technologies like the first generation cellular wireless transmission technology, IEEE 802.16x and the proposed IEEE 802.20 might implement the second. However, on the basis of related technologies and with considering feasibility, the third element still has much room for research.

  (5) The development of Internet applications requires research on multicast theories and algorithm for dynamic changing networks.

  Recently, one-to-many and many-to-many communication has become more and more important with the development of network technologies, especially the successive emergence of network applications such as video IP phone conference, network audio/video broadcasting, multimedia distance learning, IP access and forwarding of mobile users. IP multicast[24-29] is just the technology to implement such communication, which means that IP multicast may have massive market demand. Pursuing and creating new mobile multicast theories, protocols and technologies is a new research arena. It is obvious that current multicast theories and protocols for fixed networks can no longer meet the demands of these new networks with great changes. How to meet these demands and how to provide new Internet services through mobile devices have become global hotspots and difficult tasks during the research on the Internet. Hence, it is significant to do research into and create advanced theories and protocols that are adaptive to mobile networks.

3 Development Trends
  (1) Establishing theories and methods to describe mobile Internet architecture
  In the network environment with dynamic motion that is large-scale, heterogeneous, distributed and with frequent and high-rate topology changes, it is the basis for development of the mobile Internet technologies to do research on behavior patterns of moving bodies, address management and distribution policies and system architecture according to the complexity of the composition and behavior pattern of the moving body.

  Theories and methods of mobile Internet topology and architecture will at least meet the following requirements. Firstly, they can present the characteristic of frequent topology change of the network with dynamic motion, and be adaptive to the complex relationship of topology changes between moving bodies. Secondly, considering the characteristics of the network with dynamic motion, they should be able to support heuristic or on-demand topology establishment for mobile network routing protocols. Thirdly, they should somewhat be compatible with traditional topology description methods using network prefix, capable of presenting a relative stable topology outside the domain of the dynamically changing network and helping implement intra-domain and inter-domain route addressing. Fourthly, considering the enormous address resource of IPv6 and structured, functional address policies, it is important to research address allocation and management theories with new versions of Internet protocols.

  (2) Creating and perfecting theories and algorithm of mobile Internet routing protocols
  High-performance routing protocol theories and algorithm for mobile Internet should be created and perfected on the basis of research complexity of the arrangement and behavior of moving bodies on the mobile Internet, requirements of network extensibility for routing protocols, routing theories and techniques for improving network performance and dynamically changing networks with wired/wireless basic supporting networks.

  The moving body in dynamically changing networks under research can be hosts as well as a collection of hosts with relatively stable locations and a certain prefix-restricted sub-network. Massive, frequent and high-rate topology changes will notably influence traffic distribution and routing policies deployment, and will consume lots of processing and storage resources of servers. The complexity of the arrangement and motion behavior of the moving body also requires the diversification of the processing capability of network routing protocols. Besides the dynamic topology routing management problem in the case of sub-networks, the problems of host motion within sub-networks and of routing when hosts are roaming between sub-networks should also be solved. Therefore, in order to design a set of valid routing protocols, it is necessary to summarize the weaknesses of existing routing protocol theories and technologies when they are applied into the new environments, and to raise design demands for routing theories and protocols under the dynamic changing network environments on the basis of research on the motion body arrangement and motion behavior pattern research, according to the new characteristics of dynamically changing networks.

  (3) Developing multicast theories, algorithms and protocols for mobile Internet 
  Efforts should be made on the research of multicast theories, algorithms and techniques that may adapt to both current IP networks and the future dynamically changing networks. Multicast dynamic routing protocol theories should be created for dynamic networks (mobile Internet), in order to solve the problems and shortages of the multicast mechanisms of existing fixed networks and mobile IP networks. For example, to raise a layered multicast mechanism for the dynamically changing network will help the multicast routing management and maintenance of the dynamic network to adapt to the limited mobile node resource via the layered mechanism, and provide reliable IP multicast with combination of collaboration mechanisms between multicast proxies of the dynamic network. On the basis of research on layered mechanism optimization, solutions for the multicast group, multicast tree and multicast address management should be raised, especially the design of rapid turn-over mechanism, multicast routing, multicast tree structure and management protocols for dynamic nodes. Multicast routing protocols aim at implementing effective multicast data packet transfer from the source multicast sub-network to the destination multicast sub-network. The structure of multicast routing is usually a tree. Currently two types of the multicast trees are available: the source-based shortest-path tree and the shared tree. One important aspect in this research direction is to separate node motion from main multicast transmission tree through layered motion management, so as to reduce multicast tree updates caused by node motion, and to make multicast management mechanism feasible in the dynamically changing network.

  (4) Developing new routers, switching theories and protocols for dynamic (mobile or changing) networks
Since the traditional IP (Internet) network was originally designed for stationary (or fixed) networks, and routing/switching techniques and protocols were merely applicable to the stationary (or fixed) networks, next generation network technologies urgently require new network theories and protocols based on changing (mobile or dynamic) networks, so as to further develop new-type router (e.g. specific cell, wavelength and optical routers) and switching (e.g. specific cell and wavelength switching) theories and protocols based on IPv6 (or newer standards).

  (5) A stay on wireless IP network theories and technologies based on IPv6 (or newer standards)"High-speed and large-scale wireless and mobile communication" (e.g. in aviation and astronautics uses) and in particular, wireless (or mobile) IP networks, will develop greatly in the future. For example, IP WLAN (IEEE 802.11x) is developing very quickly in China and overseas; IP wireless MAN (IEEE 802.16x) is developing rapidly abroad, and its products are expected to be launched next year. The academic research on IP wideband mobile network (IEEE 802.20x) has just begun abroad, and its products will be possibly launched in several years when the related theories turn mature. Technologies for dynamic (or mobile IP) networks and intelligent networks adaptive to new changes have quick development. All these new technologies will possibly impact and even replace existing mobile (or wireless) communication networks (e.g. 2G and 3G mobile communication networks). Therefore, high/low speed IPv6-based wireless router/switch devices must be developed properly, followed by research on newer high-speed large-scale wireless and mobile communication theories, systems and devices (such as on the basis of special information cells and wavelength) adaptive to new-type dynamic (mobile) intelligent networks.

4 Conclusion
Beginning with an introduction to the status quo in mobile Internet research, this paper discusses the problems existing in the current research, and then analyzes the development trends in mobile Internet technology and the research areas that require further and deeper research. Current Internet theories and protocols only meet the demands of fixed network topology. Besides, currently, the fact is that network topology for both a moving terminal and a moving network always require frequent changes. Therefore, the paper concludes that it is necessary to do research on new Internet routing theories and protocols, structural characteristics of changing networks, multicast theories in mobile Internet environment and IPv6-based wireless IP network theories.

References
[1] Pahlavan K, Krishnamurthy P, Hatami A, et al. Handoff in Hybrid Mobile Data Networks[J]. IEEE Personal Communications, 2000, 7(2).
[2] RFC 2460 Internet Protocol Version 6(IPv6) Specification[S].
[3] RFC 791 Internet Protocol[S].
[4] Ioannidis J, Duchamp D, Maguire G. IP-based Protocols for Mobile Internetworking[A]. In: Proc. of the Conference on Communications Architecture & Protocols[C]. New York: ACM Press, 1991.
[5] Ionnisidis J. Protocols for Mobile Internetworking[D]. [Ph.D. thesis]. New York: Columbia University, 1993.
[6] Teraoka Fumio, et al. Host Migration Transparency in IP Networks[J]. ACM SIGCOMM Computer Communications Review, 1993, 23(1).
[7] Teraoka Fumio, Yokote Yasuhiko, Tokoro Mario. A Network Architecture Providing Host Migration Transparency[A]. In: Proceedings of the ACM Sigcomm Symposium on Communications, Architectures and Protocols[C]. New York: ACM Press, 1991.
[8] Teraoka Frumio, Uehara Keisuke, Sunahara Hideki, et al. VIP: A Protocol Providing Host Mobility[J]. Communications of the ACM, 1994, 37(8).
[9] Perkings C E. Providing Continuous Network Access to Mobile Hosts Using TCP/IP[J]. Computer Networks and ISDN Systems, 1993, 26(3).
[10] Rckhter Y, Perkings C. IETF Draft-work in Progress, Optimal Routing for Mobile Hosts Using IP’s Loose Source Route Option[Z]. 1992.
[11] RFC 3344 IP Mobility Support for IPv4[S].
[12] RFC 2002 IP Mobility Support[S].
[13] RFC 2003 IP Encapsulation within IP [S].
[14] RFC 2004 Minimal Encapsulation within IP[S].
[15] RFC 2005 Applicability Statement for IP Mobility Support[S].
[16] RFC 2006 The Definitions of Managed Objects for IP Mobility Support Using SMIv2[S].
[17] RFC 3775 Mobility Support in IPv6[S].
[18] RFC 2462 IPv6 Stateless Address Autoconfiguration[S].
[19] RFC 1881 IPv6 Address Allocation Management[S].
[20] RFC 1058 Routing Information Protocol[S].
[21] RFC 2328 OSPF Version 2[S].
[22] RFC 1771 A Border Gateway Protocol 4 (BGP-4)[S].
[23] Chelius G, Fleury E. Ananas: A Local Area Ad Hoc Network Architectural Scheme[A]. In: 4th International Workshop on Mobile and Wireless Communications Network (MWCN)[C]. Los Alamitos: IEEE, 2002.
[24] RFC 1075 Distance Vector Multicast Routing Protocol (DVMRP)[S].
[25] RFC 1584 Multicast Extensions to OSPF[S].
[26] Deering S, Estrin D, Farinacci D, et al. The PIM Architecture for Wide-Area Multicast Routing[J]. IEEE/ACM Transactions on Networking, 1996,4(2).
[27] RFC 2362 Protocol Independent Multicast-Sparse Mode (PIM-SM): Protocol Specification[S].
[28] Adms A, Nicholas J, Siadak W. Protocol Independent Multicast-dense Mode (PIM-DM): Protocol Specification, Internet Draft[DB/OL]. http://www.ietf.org/draft-ietf-pim-dm-new-v2-03.txt, 2003.
[29] RFC 2201 Core Based Trees (CBT) Multicast Routing Architecture[S].

Manuscript received: 2004-10-11