ACM SIGCOMM computer communication review

Abstract

At some point in the future, how far out we do not exactly know, wireless access to the Internet will outstrip all other forms of access bringing the freedom of mobility to the way we access the web, communicate with each other, and conduct business. In short, the Internet is going mobile and wireless, perhaps quite soon. A number of diverse technologies are leading the charge, including, 3G cellular networks based on CDMA technology, a wide variety of what is deemed 2.5G cellular technologies (e.g., EDGE, GPRS and HDR), and IEEE 802.11 wireless local area networks (WLANs). Wireless ISPs will offer a number of these technologies to mobile users. In some case, handsets will come with software radios that simultaneously support multiple access technologies on-the-fly; for example, IEEE 802.11 for high-bandwidth access in urban areas and GPRS for wide area access in rural areas. Each technology has its pros and cons. First and second generation cellular systems offer wide area low bandwidth voice services based on analog and digital technology, respectively. The 3G cellular systems are designed to carry voice, video and data simultaneously, and offer data rates of 144 Kbps for fast-moving mobile users in vehicles, 384 Kbps for slower moving pedestrian users, and 2 Mbps from fixed locations. Note that all users within a cell share these data rates. The 3G networks offer higher capacity and increased spectral efficiency but retain a circuit-switched, hierarchical architecture. In contrast, WLAN offers even higher bandwidth and is considered IP friendly because it offers a link layer that is very similar to wired Ethernet. However, in comparison to 3G networks, WLAN only operates within the local area, only supports best effort services, and uses shared unlicensed spectrum where few quality assurances can be provided to users. Recently, there has been a considerable amount of press on the slow rollout of 3G. However, there are some signs for optimism. Japan's NTT DoCoMo started offering 3G services in October 2001 in the Tokyo area. This came after the initial postponement of the rollout of 3G services by providers in Japan and Europe. Since May 2001, 5,000 residents in the Tokyo area have been using new 3G phones that offer improved i-mode service and real-time videoconferencing. The initial video offering uses a 64 Kbps circuit that carries video and audio combined. One of the guest editors had the opportunity to use a trial handset to set up a video call to a colleague in a taxi while traveling through Tokyo. The real-time video call, which used MPEG4 technology, presented mixed service quality but the experience of setting up the call between two taxis was exciting. I-mode currently has 29 million subscribers in Japan and DoCoMo hopes to keep that figure rising with the new service offerings. The DoCoMo radio access network is based on WCDMA and the core network on ATM switching. Many carriers in the US and Europe will be keenly watching what is happening in Tokyo. Wireless providers in the United States are eager to follow suit but are rolling out service in phases with emphasis on 2.5G technologies such as GPRS, which provides an always-on connection to the Internet that allows users to toggle between surfing the web, a phone call, or text messaging without losing the connection. Carriers in Europe, which have invested more than $100 billion to buy 3G radio spectrum licenses and will need to invest another $100 billion for the build-out of the 3G networks, will be keeping a close watch on DoCoMo's successes and failures. The vast majority of WLAN deployed today is based on IEEE 802.11b operating at 2.4 Ghz and offering data rates up to 11 Mbps. Recently, a number of companies have demonstrated IEEE 802.11 a, which operates in the 5Ghz band and offers data rates up to 54 Mbps. In fact, Atheros Communications supports a "turbo-networking" mode that delivers 108 Mbps, roughly equivalent to Fast Ethernet. The cost of the 3G spectrum and the build-out of the 3G networks have been so prohibitive that many operators have been pushed to the brink of bankruptcy. As a result, many small operators in Europe are sharing the cost of the build-out by sharing core and radio access network infrastructure. In contrast, WLAN infrastructure operates in unlicensed frequency bands and is very cheap in comparison to cellular equipment. Cheap, because WLAN base-station transceivers are priced at less than $1,000, and transceiver cards are around $100 or come built into computers. Public wireless LANs can handle large volumes of data at significantly lower costs compared to leading 3G technologies. The cost benefit and bandwidth differential offered by WLAN technology makes it a disruptive technology as the cellular operators migrate from 2G to 3G. Disruptive technologies are characterized as being cheaper and of lower performance than sustainin g technologies (e.g., 2.5G or 3G solutions). Most public wireless networks and enterprise networks use WLAN, not because it is more secure, robust or spectrally efficient, but simply because it is cheap, offers high bandwidth, makes networks easy to build and configure, and, importantly, it works. Typically, customers are not initially satisfied with the performance offered by disruptive technologies when they are first introduced. For WLAN to compete in the marketplace with 2.5G and 3G solutions, public WLAN operators would need to be capable of building metropolitan area networks that provided suitable support for voice-over-IP there by enabling voice communications. Sharing unlicensed spectrum means that wireless ISPs cannot build managed networks where services are tightly controlled, in isolation from other operators, as a means of assuring performance. Historically, however, disruptive technologies have tended to resolve such performance problems as they mature and begin to capture market share. Examples of wireless extensions to Internet are all around us today. Here in New York City many companies, university campuses, coffee shops and stores offer wireless access to the web using WLAN technology. Columbia University, for example, provides students and faculty wireless access to the web as they move around campus. Companies such as MobiStar and Waypoint provide wireless connections at hotels, airports and cafes. Around Manhattan, Starbucks coffee shops offer wireless access to the Internet. At the grassroots level, community groups are putting up wireless antennas around the New York City area and in other cities offering free access to Internet. Some predict that these "freenets", which have a feel reminiscent to Napster, will ultimately succumb to a sustained corporate challenge or new wireless ISPs that offer cheap services across dense urban areas. The road to success for such fledgling operators may be littered with a number of business, regulatory and performance obstacles. There are a number of companies, standards bodies, and industrial fora vying to define future wireless extensions to the Internet. The end result is that operators are faced with a large and confusing array of choices on how best to build next generation mobile networks. 3G systems offer support for seamless mobility, paging, and service quality but are built on complex and costly connection-oriented networking infrastructure that lacks the inherent flexibility, scalability, and cost effectiveness found in IP networks. In contrast, Mobile IP represents a simple and scalable global mobility solution but lacks support for fast handoff control, real-time location tracking, and authentication and distributed policy management found in cellular networks today. There has also been considerable interest in new emerging wireless technologies such as personal area networks, mobile ad hoc networks and sensor networks. How these technologies interwork with the global Internet is an active area of research. A number of micro-mobility protocols (e.g., Cellular IP, Hawaii, Hierarchical Mobile IP) and fast handoff schemes have been discussed in the IETF Mobile IP Working Group that address some of these performance and scalability issues. These protocols are designed for environments where mobile hosts change their point of attachment to the network so frequently that the basic Mobile IP protocol tunneling mechanism introduces network overhead in terms of increased delay, packet loss and signaling. For example, many real-time wireless applications (e.g., voice-over-IP) would experience noticeable degradation of service with frequent handoff. Establishment of new tunnels can introduce additional delays in the handoff process, causing packet loss and delayed delivery of data to applications. This delay is inherent in the round-trip incurred by Mobile IP as the registration request is sent to the home agent and the response sent back to the foreign agent. Micromobility protocols aim to handle local movement (e.g., within a domain) of mobile hosts without interaction with the Mobile IP enabled Internet. This has the benefit of reducing delay and packet loss during handoff and eliminating registration between mobile hosts and possibly distant home agents when mobile hosts remain inside their local coverage areas. Eliminating registration in this manner reduces the signaling load experienced by the network in support of mobility. As the numbers of wireless users grow so will the signaling overhead associated with mobility management. In cellular networks registration and paging techniques are used to minimize the signaling overhead and optimize mobility management performance. Currently, Mobile IP supports registration but not paging. An important characteristic of micro-mobility protocols is their ability to reduce the signaling overhead related to frequent mobile migrations taking into account a mobile host's operational mode (i.e., active or idle). When wireless access to Internet becomes the norm then Mobile IP will have to provide efficient and scalable location tracking in support of idle users, and IP paging in support of active communications. Support for "passive connectivity" to the wireless Internet balances a number of important design considerations. For example, only keeping the approximate location information of idle users requires significantly less signaling and thus reduces the load over the air interface and in the network. Reducing signaling over the air interfaces in this manner also has the benefit of preserving the power reserves of mobile hosts. Currently, the IETF Seamoby Working Group is tasked with developing an IP paging protocol. The papers in this special issue address a number of the issues and challenges discussed above. We received a total of 32 excellent submissions for this special issue -- a much greater response to our call for papers than we expected. The papers came from different regions around the world and addressed many different aspects of research. Each paper was reviewed by three or more experts, who evaluated the technical content and suitability of the paper for publication in this special issue. As guest editors of the special issue we had the very difficult job of selecting only six papers from those submitted. Several deserving papers could not be accommodated in this special issue because of space. We hope to see those papers appear later in ACM SIGCOMM Computer Communication Review. The first three papers in this special issue address a number of enhancements to Mobile IP and cellular networks to provide for better support for fast handoff and context transfer, wireless Internet telephony, and IP paging. The final three papers deal with the emerging technologies of personal area networks, mobile ad hoc networks and sensor networks. In the first paper, Jonathan Lennox, Kazutaka Murakami, Mehmet Karaul and Thomas F. La Porta, Lucent Technologies, discuss internetworking Internet telephony and wireless telecommunications networks. The authors propose a number of schemes to directly interconnect the 3G UMTS and SIP Internet telephony systems. The next paper by Rajeev Koodli and Charles E. Perkins, Nokia, deals with seamless handoff and context relocation in mobile networks. Context transfer refers to state information (e.g., QOS state) associated with a particular service (e.g., VoIP) that needs to be re-established with mobility. The authors show that fast handoff with context transfer at the network layer can support uninterrupted voice-over-IP services. The paper by Pars Mutaf and Claude Castelluccia, INRIA, proposes adaptive per-host IP paging. The authors observe that many of the existing IP paging proposals found in the literature promote the use of static or manually configured paging areas. The authors argue that there is a need for dynamic and adaptive paging area management that takes into account host mobility and traffic patterns in the network. Robin Kravets, Casey Carter and Luiz Magalhaes, University of Illinois, Urbana-Champaign, discuss cooperative approaches to user mobility. The authors propose the necessary networking functionality that allows groups of mobile devices (e.g., a set of devices that collectively comprise a personal area network) to interact and be seamlessly integrated into the Internet. In the next paper, Jyoti Raju and J. J. Garcia-Luna-Aceves UC Santa Cruz, present a new mobile ad hoc network routing protocol called source tracing and compare it with dynamic source routing (DSR). Both on-demand and table-driven implementations are considered. The final paper in this special issue, by Samir Goel and Tomasz Imielinski, Rutgers University, considers the problem of monitoring data in large sensor networks. The authors propose a prediction-based monitoring scheme that can be visualized by leveraging concepts and techniques found in image processing. There are many other technical challenges before Internet goes truly wireless and mobile. For example, there is a need to minimize the impact of mobility on TCP performance, resolve security issues over-the-air, and further study how best content can be pushed toward mobile users. Finally, in the wake of the recent attack in New York City, we anticipate new advances in rapidly deployable wireless infrastructure, self-configuring networks, and sensor networks - collectively forming disaster relief networks. As guest editors it has been a great pleasure to put together this issue. We would like to thank the authors for their contributions and the reviewers for their time, energy, and comments that helped shape this special issue. We hope you enjoy it as much as we do.