Adun Wireless Optical

Computer Networks 52 (2008) 1864–1872

Contents lists available at ScienceDirect

Computer Networks journal homepage: www.elsevier.com/locate/comnet

A Ubiquitous wireless network architecture and its impact on optical networks Haruhisa Ichikawa a,*, Masashi Shimizu b, Kazunori Akabane b, Osamu Ishida b, Mitsuo Teramoto b a b

The University of Electro-Communications: JST CREST, 1-5-1, Choufugaoka, Choufu-shi, Tokyo 182-8585, Japan NTT Network Innovation Laboratories, 1-1 Hikari-no-oka, Yokosuka, Kanagawa 239-0847, Japan

a r t i c l e

i n f o

Article history: Received 10 July 2007 Accepted 8 December 2007 Available online 18 March 2008

Keywords: Optical network Ubiquitous wireless network Internet Terabit LAN Grid RFID DROF SDR Disruptive innovation

a b s t r a c t Optical networks will change greatly over the next 10 years. This is because, if the current growth rate is maintained, the Internet will have expanded 100–1000 times. Networked wireless appliances, such as radio frequency identification (RFID) tags and wireless sensors, are expected to greatly outnumber PCs. Such exponential changes in network capacity and terminals may lead to the emergence of post-IP networks. This paper introduces a candidate for a post-IP network called the ‘‘appliance defined ubiquitous network (ADUN)”, which supports niche ubiquitous network applications for affordable implementation. The ADUN will demand optical networks that can transport 10–100 Gbps streams, each of which requires almost the full transmission capacity of one wavelength or a wavelength group. This paper discusses directions for the functional enhancement of optical network architecture, dynamically using wavelengths for grid computing, so as to support the ADUN. Ó 2008 Elsevier B.V. All rights reserved.

1. Introduction The Internet has continued to grow rapidly. A task force organized by the Japanese Ministry of Internal Affairs and Communications estimated that Japanese Internet traffic had more than doubled in a year, by measuring the traffic at the Internet Exchange Point in Tokyo in April 2004 [1]. Their most recent report, in February 2007, shows that the Japanese Internet is still growing at a rate of 150% per year. At this rate, the Japanese Internet will be 100– 1000 times larger after 10 years. This is true for both the core and access networks. The number of fibers to the

* Corresponding author. Tel.: +81 46 240 5000; fax: +81 46 2222. E-mail addresses: [email protected] (H. Ichikawa), [email protected] (M. Shimizu), [email protected]. co.jp (K. Akabane), [email protected] (O. Ishida), teramoto. mitsuo@ lab.ntt.co.jp (M. Teramoto). 1389-1286/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.comnet.2007.12.013

home (FTTH) service subscribers will soon exceed 10 million. This Internet growth rate will eventually be matched throughout the world. Such exponential growth will inevitably lead to a huge qualitative change in the network. Optical networks should be able to support both exponential capacity growth and qualitative changes, since they form one of the most basic infrastructures of information and communication technology. Network architectures have previously been determined by the major terminals for the networks. Telephony networks are designed to provide stable voice communication channels for 3.4 kbps voice streams between telephones by ensuring that the quality of the service meets certain standards. It should be noted that packets are routed between host computers on the Internet, while telephony networks are switched networks. This difference between the characteristics of the two types of network traffic leads to a difference between the two network

H. Ichikawa et al. / Computer Networks 52 (2008) 1864–1872

1865

architectures. Therefore, the essential architectural change of the Internet is considered to depend on whether computers remain as the major terminals in the future. Passive or active radio frequency identification (RFID) tags and wireless sensors are expected to be employed for networking real objects of the real world. If this approach is implemented, such devices will become the major terminals, since the number of real objects is a hundred or a thousand times greater than the number of PCs. Most of the devices are much weaker than PCs in terms of their functions. They will be power-sensitive and will not always operate at high speed. They provide functions for limited purposes. Once they constitute the majority of terminals, the networks, hereafter ‘‘ubiquitous wireless networks”, will undergo essential changes in terms of architectures. This paper introduces a ubiquitous wireless network architecture leading to the post-IP network, and discusses its impacts on the optical network architecture supporting both the IP and the post-IP networks with much greater bandwidths than those presently in use. It is often useful for future network design to study extremely high-end applications. As one of such network designs, Section 2 introduces an optical local area network that can transmit Terabit-class bulk data with low latency in a dynamic manner for grid applications with super-computers [2]. Section 3 reviews the history of disruptive network innovations. The review points out that network architectures are determined by terminals or appliances to be networked, and that optical networks should be designed accounting for the possible shift of appliances to be networked in the future. In Sections 4 and 5, active RFID appliances are introduced as the entry and the ultimate appliances for the ubiquitous wireless networks, and the grand challenges for ubiquitous wireless network deployment are discussed and summarized. In Section 6, we introduce an approach to ubiquitous wireless network deployment and show that this approach may lead to a post-IP network. The approach introduces a new architecture, called the appliance defined ubiquitous network (ADUN) [3], where wireless radio spaces are virtually extended over broadband networks to enable affordable implementation of ubiquitous network applications. Broadband streams of 10–100 Gbps are transmitted over the ADUN. In order to employ advanced digital radio signal-processing technologies as far as possible, virtually extended wireless spaces should maintain the relationship between their corresponding real spaces. This paper discusses the new requirements that this will impose on optical networks. These requirements are different from those imposed by grid applications. They indicate the progress direction of the optical network as the basic infrastructure for both the Internet and post-IP networks.

tions typically have terabit to petabit data size requirements [4], and each data set should be transmitted within several seconds or several tens of seconds. This section introduces the approach proposed in [2] that is designed to satisfy these requirements by using wavelength groups dynamically. We call optical networks that employ groups of wavelengths dynamically from end to end ‘‘lambda access networks.” An optical local area network architecture has been proposed [2], where terabit-class bulk data can be transmitted with low latency in a dynamic manner. Fig. 1 shows an example network configuration. The network consists of parallel optical interfaces, each of which is attached to three components: a super-computer or a visualization server; wavelength-banks, each of which provides required wavelengths to each interface; and an optical switch, which connects an optical interface via optical fibers in the star configuration. It should be noted that the concept must include multiple switching nodes and not be restricted to the star topology. The main characteristics of the network architecture, known as optical virtual concatenation, are (1) the assignment of an arbitrary number of wavelengths to a bulk-data set for transmission according to the latency requirement, and (2) the transmission of a bulk-data set by dividing it into an arbitrary number of blocks and passing the blocks across parallel WDM signals as virtually contiguous data. When a bulk data set arrives at an optical interface, the transmitter requests the establishment of optical channels on multiple wavelengths through the control plane, where the number of wavelengths is derived from the latency requirement. The transmitter divides the bulk-data set into parallel streams, and stores them in electronic buffer-memories until the reservation is complete. If the optical switch finds a group of available wavelengths, it sends a positive command and information about the wavelengths, either to the next optical switch (hop) or to the receiver, through the control plane. If the receiver accepts the positive command and information, it sends back an acknowledgement signal with wavelength group information. The optical switch receives the acknowledgement, transfers this to the transmitter, and then sets up the switch configuration. The key physical technologies in the above network have been verified in a field demonstration [5]. A successful optical virtual concatenation of 32 km was achieved, where the wavelength group switching was implemented by an 8  8 planar lightwave circuit (PLC) optical switch matrix, and the wavelength group conversion was implemented by a polarization independent waveband converter based on the quasi-phase – matched LiNBO3 (QPM-LN) waveguide.

2. Optical network evolution

3.1. Network architecture dependability on terminals or appliances

When developing innovative technologies for network capacity expansion, it is often useful to study the requirements of high-end applications. For example, grid computing for super-computers has the potential to generate extremely large data transmissions. Recent grid applica-

3. Possible disruptive innovation of ubiquitous wireless networks

We should pay attention to the possibility that the extraordinary expansion of network capacity may lead to a shift in network architectures or, more generally speaking, the ICT platforms. Network architectures have histori-

1866

H. Ichikawa et al. / Computer Networks 52 (2008) 1864–1872

Storage Server

Super computer

Wavelength-bank

Optical frequencies Delay non-critical -> needs 2 waves Visualizing Server

Delay critical da -> needs 6 waves

x

Fast optical switch

Parallel optical I/F with Virtual Concatenation Fig. 1. Network configuration of Terabit LAN.

cally been determined by terminals or appliances. In the era of plain old telephone services, the terminals were telephone terminals, and the role of the network was to provide communication channels for voice communications, where a stable quality of 3.4 KHz bandwidth stream transmission between the caller and the receiver was demanded for each voice communication. As a result, the circuit switching architecture was adopted for telephony networks. On the other hand, the end systems for the Internet are computers. Computers transmit bulk-data, so the packet switching architecture was adopted. We can find the impact of the terminal change from telephones to computers in the following architectural principles of the Internet, summarized from the Internet Architecture Board document [6]. The Internet technologies have been designed and developed by paying much attention to these principles. 1. The goal is connectivity. 2. The tool is the Internet Protocol. 3. The intelligence is end-to-end rather than hidden in the network. The first principle demands that all the computers should be able to communicate with each other. All the IP addresses are, in principle, known to all the computers connected to the Internet. Special arrangements are required if the communications should be restricted only between limited groups of computers. Different from the telephony network, other features of communications, such as quality, are given second priority. In the second principle, the Internet Protocol is a packet routing protocol, which is suitable for bulky data transfer between computers. The third principle comes from the fact that computers are powerful and flexible enough to implement services at the ends. This principle was very reasonable when the terminals shifted from dumb telephone terminals to computers. 3.2. Long term preparations for the disruptive innovation on the internet In addition to the impacts of the major terminal changes in the network architectures, as described in the previous

subsection, we can learn much about the terminal change impacts from the deployment history of the Internet. Disruptive innovations were elegantly studied by Prof. Clayton Christensen of Harvard University [7]. One of the disruptive innovations is the creation of the PC industry. The PC industry grew by using research and development results accomplished for mainframe and mini computers. For example, broad ranges of R&D were conducted for massive fast disc systems, efficient in terms of cost per byte from materials to systems. Exponential improvement of disc technologies brought out disruptive innovation in the computer industries. In the early 1980s, the emerging market of desktop personal computers was strengthened using the 5.25-inch disc drives which failed to address the needs of the mini computer market at that time. The 5.25-inch drive was small and lightweight. Priced at around $2000, it could be incorporated into desktop machines economically. The personal computers in those days were targeted at the niche market of computer nerds, but, as is well known, eventually eroded the mainframe and mini computer businesses. Disruptive innovation also occurred in the telecommunication industries. Very long term preparatory efforts were made before the disruptive innovation as described in Fig. 2. The first packets were successfully transmitted over ARPA-NET in 1969. From that point, it took more than 20 years for the Internet to become commercially sustainable. TCP/IP protocol standards were developed and applied to the net in 1983. Important companies and organizations such as Cisco Systems, SUN Microsystems and the Internet technology standardization organization, IETF, were established in these 20 years. The business model for the Internet Service Providers (ISP) was also invented as seen in the creation of MCI ISP service. Vast amounts of government funds, from both DARPA and the NSF, were used for government-industry-academic collaboration, to provide the researchers and engineers of the Internet with test beds where the Internet technologies could be developed, used for services, such as e-mail, and improved. Broadband network technologies, such as synchronous digital hierarchy (SDH) and wavelength division multiplexing (WDM), contributed to the rapid growth of the Internet, which was not intended by the investors in these technologies. Telecom industries made huge and long term

1867

H. Ichikawa et al. / Computer Networks 52 (2008) 1864–1872

Product Performance

- Government-industry-academic collaboration more than 20 years - R&D results for telephony service innovation drove disruptive innovation of Internet DARPA“ NSF“ VoIP/ Universal Super n atio Instant Messagin Networking” Computing” nov n I g Performance Demanded By inin (2000) a t s Su Napster(1999) High-end Market WDM

g

ion e vat ptiv u nno Google(1998) I r s g n Di vatio stainin SDH o Su In n PC

Performance Demanded By Low-end Market Web(1991), Archie(1990) UUNET Commercial Service(1987) TCP/IP, NSF Net/(Merit,MCI,IBM)(1983) AT&T Divestiture1984) vBNS(95) MCI(00) Time ARPA-net(1969) SUN Microsystems(1982) IIJ(1993) Cisco Systems(1984)

WS Mini-Computer, E-mail

Fig. 2. The Internet evolution through disruptive innovation.

investments in such technologies for so called broadband multimedia communications as advanced telecommunications services. The developed technologies brought out disruptive innovation, and the Internet is ultimately going to substitute for telephone networks. The bandwidth of the Internet has kept growing rapidly and has become powerful enough to even carry video streams. Such bandwidth is considered excessive, in some sense, for e-mail and simple web access. Thus, we might be at the beginning of another disruptive innovation. In the following, we will build a scenario of such disruptive innovation from the Internet towards ubiquitous wireless networks. 4. Appliances for ubiquitous wireless networks As discussed in the previous section, terminals or appliances are considered to determine the network architectures. In this section, we introduce the active RFID appliances for commercial solutions and the experimental wireless appliances for intelligent communications services [8,9] as the entry and ultimate appliances of the future, respectively. The applications of the former and latter wireless appliances are quite different, but both are equipped with batteries and similar sensors. We think active RFID will evolve into major appliances for future ubiquitous wireless networks. There have been many specific but economically useful active RFID systems developed. An active RFID system for international container tracking was developed, where an active RFID reader was included in each container. The reader reads the active RFID tags at each place the container passes through, to collect the information on the container traveled route. In addition to this, the reader monitors the temperature and vibrations of the container using its internal sensors. The collected data is kept in the storage of the reader and retrieved afterwards. This system was tested at the port of Kobe, Japan in January 2006. Passive RFID tags were also tested, using various

wireless bands of 2.4 GHz, 950 MHz, and 13.56 MHz, for whether they could be used for container tracking. The photo in Fig. 3 shows the scene of the test, where three PCs display the tracked data using 426 MHz and 303 MHz wireless bands for active RFID tags, and 953 MHz for passive RFID tags. The systems using the active RFID tags of 426 MHz and 303 MHz successfully tracked the container transfer, but passive RFID tags were proven to be inapplicable for container tracking. Based on this test result, the wireless band of 433 MHz was agreed to be used internationally for container tracking with active RFID. A wellengineered 433 MHz version of the above described system was successfully tested by tracking containers from the Hong Kong port to the Kobe port. The black object attached to a PC in the photo of Fig. 4 is an active RFID tag for another application system for inventory management. As a special type of inventory management, the system was built to monitor the position of PCs within a building for information security protection.

Fig. 3. Experiment of international container tracking system.

1868

H. Ichikawa et al. / Computer Networks 52 (2008) 1864–1872

Fig. 4. PC tracking system for information security protection.

Technologies for capturing and searching real world events are studied for intelligent communications in [8,9]. Fig. 5 illustrates the experimental system which comprises two modules: a query module and a search engine. Sensors attached to physical objects sense the physical information, such as motion of objects, and transmit the data to the sensor database. The query module accepts a query in the form of an English word set and translates it into a representation that the search engine can interpret. The search engine then searches the sensor database by calling the sensor grounder, which finds a data segment that satisfies the query condition. Also, the query module answers by, for example, displaying a video image recorded by video cameras or by sending a reply via e-mail. The sensor appliance is equipped with (a) micro-sensors, such as an accelerometer, a thermometer, a hydrometer, an illuminator, or a human detector, and (b) wireless networking and computing engine for communicating with the servers. The experimental system demonstrates that it is possible to answer questions such as ‘‘Who dropped this vase?” by finding the identifier of the questioned vase, receiving the data from the sensor attached to the vase, translating the sensed data into the movement of the vase, and retrieving the video image at the time when the sensor detected the vase dropped.

The active RFID solutions are useful, but presently form only niche markets. They do not seem to provide markets large enough to build ubiquitous wireless networks for their own purposes, while their aggregate will input vast amount of information to the networks in the future. The RFID appliance designs are not very independent from their applications. For instance, the data size and duty cycle of RFID radio protocols and the radio transmission power are carefully designed so that their functions should be suitable for their applications. The batteries of RFID appliances should operate just as long as needed for the applications. Power consumption is one of the most critical issues for RFID radio interface design, whether the RFIDs are active or passive. For example, the famous EPC five cents passive RFID tags can be read by readers 10 m away from the tags, while only a small amount of data from the tags, that is their identification numbers, can be transmitted. On the other hand, passive RFID tags carrying electronic money must send very secure radio protocols, so that the RFID radio signals cannot be transmitted over many meters. It is difficult to have a single standard for a wireless protocol, even for passive RFIDs. Active RFIDs have more options in their radio interface design. There can be many ways to use the limited battery power to meet the required functions, the reachable range of the radio signals, data size, duty cycle, and so on. Consequently, it is difficult to standardize the protocols of active RFIDs. 5. Grand challenges of ubiquitous wireless network deployment Many efforts have been made to clarify the grand challenges presented by ubiquitous or sensor networks. Most of these efforts are concerned with networks with huge numbers of wireless appliances. However, the history of the Internet indicates that it is important to focus on the deployment process from the smallest to the largest scale and to design the architectural principles that control the deployment process. If the principles have to be completely different from the ones for the Internet, the network will be a post-IP network. In this section, we discuss the grand challenges facing ubiquitous wireless network deployment. 5.1. Niche application

Fig. 5. Real world information capturing system.

As discussed in the previous section, it appears difficult to standardize the interface between wireless appliances and networks. Fig. 6 shows the frequency allocations available for RFID with the upper-limits of their radio transmission power and bandwidth. The communication range of wireless systems depends greatly on their frequency propagation conditions and radiation power. Therefore, the radio band will be selected according to the requirements of the target applications. As a result, there will be many kinds of readers receiving signals from wireless appliances. The cost of system or infrastructure construction and operation will be too great to provide niche markets with ubiquitous network access if many kinds of read-

1869

H. Ichikawa et al. / Computer Networks 52 (2008) 1864–1872

UHF Passive Short Range Devices

HF Passive IC Card Common

LF Passive Immobilizer

915MHz USA

< 135KHz

Use

869MHz

Microwave WiFi

EU 13.56MHz

433MHz EU US Japan 315MHz US JAPAN

0.1MHz

1MHz

10MHz

100MHz

2.45GHz 5.2GHz 5.8GHz

Worldwide Regional Almost worldwide

953MHz Japan

1GHz

10GHz

Active Smart Key Fig. 6. Frequency allocation available with RFID.

ers are installed, as is the case today. We need technologies that enable resources to be shared between applications. 5.2. Security/privacy issues Security and privacy issues should be carefully taken into consideration with new technology deployment. People are particularly concerned with the power that technology has for information distribution, in spite of the fruitful contributions expected. In the large scale deployment trials of passive RFID supported by MIT Auto-ID center, which was the RFID standardization organization, some part of the trials were prevented by a consumer group, Consumers Against Supermarket Privacy Invasion and Numbering (CASPIAN). They made an objection against the attachment of RFIDs outside shops to consumer products, such as safety razors, so as to track the product lifecycles related to consumer private lives. It is believed that supply chain management performance can be improved by substituting passive RFID for bar codes. RFID tags enable one to identify individual products by assigning different identification numbers to the individual products of the same category, while the same bar code number is assigned to all the products of a category. By tracking each individual product, the whole lifecycle of the individual product can be managed, from its production to destruction. This was opposed by CASPIAN because the tracking could reveal the precise records of consumers’ behavior, that is, it could compromise consumer privacy. This is very serious for passive RFID tag deployment because the additional performance improvement of the consumer product supply chain management becomes difficult by substituting RFIDs for bar codes. Active RFID appliances can collect much richer information than passive RFID, so consumer privacy protection should be considered mandatory in active RFID application systems. Also, some typically expected applications relate to the security of our lives. It would not be easy for individual niche ubiquitous wireless network application systems to satisfy the requirements of security and privacy with an acceptable cost for users.

5.3. Technological innovation Current technologies used for ubiquitous wireless networks are rather primitive. Innovative wireless technologies are expected to appear in the future for signal detection performance improvement, signal source location, and so on. Wireless appliances will also be much improved. Ubiquitous wireless networks should be able to accommodate themselves to such technological innovations. In addition, the ubiquitous wireless networks will be constructed over the Internet initially, so that efficient migration from the Internet should be taken into account. 6. Appliance defined ubiquitous network 6.1. Tools for making the infrastructure soft Digitized radio on fiber (DROF) [10,11] is a technology for transferring wireless space radio information to a remote receiver site, by digitizing the information and then transferring it over broadband networks. DROF enables us to locate receivers at the optimum sites with respect to system construction and operating cost. It will be possible for operators to take care of many different users’ receivers, as well as share receivers between users. Software defined radio (SDR) [12,13] is a technology with which to implement radio systems, using software instead of hardware circuits. Multi protocol readers employing SDR would be feasible if the required signal speed is not very fast. Fig. 7 illustrates a straightforward implementation that makes the fixed broadband network universal for wireless appliance systems by using DROF and SDR technologies. Wireless spaces are essentially extended to the receiver sites. 6.2. Architectural principles To cope with the above challenges, we propose an architecture for ubiquitous networks called appliance defined ubiquitous networks (ADUN) with the following

1870

H. Ichikawa et al. / Computer Networks 52 (2008) 1864–1872

Application

Extended Wireless Space

Protocol Reader

IF or Baseband

AD C

Broadband Network

Wireless Appliances

Decoder Wireless Space Info. Storage

Fig. 7. Wireless space extension with DROF and SDR.

architectural principles, which are quite different from those of the Internet. 1. Networks with ADUN architecture speak the protocols spoken by the appliances. Regarding current networks, it should be noted that terminals/appliances have been designed to speak the protocols expected by the networks. 2. Continuous radio information included in wireless spaces, rather than individual wireless appliance signals, is transported over the network. This means that radio information streams are transported over the networks, while packets are transported over the Internet. 3. Wireless appliances and radio signals received by antennas have owners. Signals from the spaces and the appliances are received and processed under the

control of the owners. In other words, the ADUN network comprises private networks. In contrast, any Internet host can send/receive packets to/from any other host. The hosts’ applications are responsible for confirming that the packets they send to the hosts are the intended ones. Internet hosts cannot prevent the arrival of packets from malicious hosts. 4. Signals from the wireless spaces and appliances of different owners are internetworked by administrators trusted by all the owners.

6.3. Architecture An owner’s view of an example ADUN network is shown in Fig. 8. There are two kinds of wireless appliances (WAP)

Fig. 8. Owner’s view of ADUN.

H. Ichikawa et al. / Computer Networks 52 (2008) 1864–1872

that are owned by individuals or used for public purposes. WAPs are located in private or public spaces. Radio signals in a certain frequency range in a private space are digitized and transported over a private wireless space network (WSN) to an information-sharing platform (ISP). The ISP is operated by an administrator trusted by all the ADUN users. The digitized radio streams from the owner’s private space are stored in a database, read by the owner’s decoder, and used by the owner’s applications. To share his/ her own and public WSN with others, the owner allows the ISP to use his/her WAP signal decoder. If a visitor is allowed by the owner to use the owner’s private WSN, the ISP extracts the visitor’s WAP data by decoding the owner’s radio stream and stores them in the visitor’s database. On the other hand, if the owner is allowed to use someone’s WSN, the ISP extracts the owner’s WAP data and stores them in the owner’s WAP database in the same manner. The ISP also extracts the public WAP data and stores them in the public WAP database; every user is allowed to use the public WSN. A wireless space analyzer finds new information by decoding and searching the stored data. For example, the context of a certain owner’s WAP will be inferred by decoding and searching other owners’ space data.

1871

6.5. Impacts on optical network architectures The wireless space can provide rich information through the use of both real time and non-real time processing, as seen in the radio telescope field. The virtually extended wireless spaces in ADUN should keep the rich information included in real wireless spaces as far as possible. This will need greater bandwidth than the above estimation for ADUN with respect to the transfer, and will also impose additional requirements. For example, antenna diversity is a technology for improving the performance of receivers using radio information over multiple radio paths. In order to use this technology, virtually extended radio paths should attempt to maintain the relationship between the corresponding real paths. For example, virtual radio paths with low latency will be required, but transmission delay stability and multi path synchronization will also be demanded. These requirements are different from those imposed by grid applications, but the two requirements do not seem to be contradictory for future optical networks. The ADUN will be well supported by the lambda access architecture introduced in Section 2 with the appropriate modifications. However, more detailed studies are needed.

6.4. Traffic generated by ADUN It is difficult to transfer all the radio information in a wireless space using DROF without any restrictions on, for example, the wireless frequencies, radio systems, or antenna sizes. As shown in Fig. 6, the wireless bands available for ubiquitous network appliances are restricted to those listed below. This restriction will not change greatly because new systems will be legally expected to use wireless bands without interfering with existing systems. 1. 312–440 MHz, 2. 869–960 MHz, 3. 2.4–2.4835 GHz. Simple radio systems have been applied to the 312– 440 MHz band, where a bandwidth of less than 1 MHz is assigned to each channel. On the other hand, a 26 MHz bandwidth is allowed for use with spread spectrum systems for the 915 MHz industry, science, medical (ISM) band in the USA. In the 2.4 GHz band, a bandwidth of 83.5 MHz is available for a typical radio system, IEEE 802.11g [14], which uses a 64 QAM modulation/demodulation system. The relationship between signal-to-noise ratio and bit error rate depends on the modulation-/-demodulation systems. When we take this into account, radio information over 83.5 MHz in the 2.4 GHz band is transported at about 8 Gbps over DROF. It should be noted that the latency condition assumed by IEEE 802. 11g, ‘‘short inter frame space” from the ‘‘data packet send” to ‘‘ACK receive,” remains a problem when extending the wireless space by DROF. In the 312–440 MHz band, it is desirable for ADUN to transport the 128 MHz band by down converting the 312– 440 MHz band; thus making the ADUN independent of radio systems. The 128 MHz band is estimated to generate a 5.1 Gbps stream over DROF.

7. Conclusion This paper introduced a ubiquitous wireless network architecture leading to the post-IP network and discussed its impacts on the optical network architectures supporting both the IP and post-IP networks. First, we introduced a lambda access network architecture with optical virtual concatenation for extremely broadband grid applications. With this architecture, the low latency requirements of the applications are supported by using a group of wavelengths to transfer bulk data. Computers are the end systems for such high end applications. In most research on advanced optical networks, end systems are assumed to be computers, but the optical networks are expected to work not only for computer applications, but also for the ultimate ubiquitous wireless applications in the future. The history of disruptive network innovations indicates that terminals or appliances determine the network architectures. Active RFID appliances are introduced to play major roles, even for the ultimate ubiquitous wireless applications of the future. The grand challenges are summarized from the viewpoint of ubiquitous network deployment. A new architecture called ADUN was introduced to cope with the grand challenges. The ADUN is quite different from the Internet in terms of architectural principles. The ADUN will be a post-IP network once it becomes more dominant than the Internet. With the ADUN, wireless spaces are virtually extended over optical networks by transporting 10–100 Gbps streams. Features, such as transmission delay stability and multi path synchronization, are needed in addition to low latency if we are to employ advanced digital radio signal processing technologies. This indicates an evolutionary direction for optical networks as the infrastructure for both Internet and post-IP networks.

1872

H. Ichikawa et al. / Computer Networks 52 (2008) 1864–1872

References [1] . [2] M. Tomizawa, J. Yamawaku, Y. Takigawa, M. Koga, Y. Miyamoto, T. Morioka, K. Hagimoto, Terabit LAN with optical virtual concatenation for grid applications with super-computers, in: OFC (Optical Fiber Communication Conference), 2005. [3] H. Ichikawa, M. Shimizu, K. Akabane, Appliance defined ubiquitous network, in: SAINT (International Symposium on Applications and the Internet), 2007. [4] EGEE (Enabling Grid for E-Science in Europe), . [5] J. Yamawaku et al., ECOC2004, Post-deadline paper Th4.5.3, 2004. [6] B. Carpenter (Ed.), Architectural Principles of the Internet, Internet Architecture Board, 1996. [7] C.M. Christensen, The Innovator’s Dilemma, Harvard Business School Pr, 1997. [8] T. Okadome, T. Hattori, K. Hiramatsu Y. Yanagisawa, A real-world event search system in sensor network environments, in: The Proceedings of the 7th International Conference on Mobile Data Management (MDM’06), 2006. [9] T. Okadome, T. Hattori, K. Hiramatsu, Y. Yanagisawa, Project pervasive association: toward acquiring situations in sensor networked environments, Journal of Computer Science and Network Security 6 (3B) (2006) 134–139. [10] P.M. Wala, A New Microcell Architecture Using Digital Optical Transport, VTC93, 1993, pp. 585–588. [11] S. Kuwano, Y. Suzuki, Y. Yamada, K. Watanabe, Digitized Radio-OverFiber System for Wide-Area Ubiquitous Wireless Network, MWP2006, P3, 2006. [12] W. Tuttlebee, Software Defined Radio, 2002. [13] H. Shiba, T. Shono, Y. Shirato, I. Toyoda, K. Uehara, M. Umehira, Software defined radio prototype for PHS and IEEE 802.11g wireless LAN, IEICE Transactions on Communications E85-B (12) (2002) 1703–1715. [14] ANSI/IEEE Standard 802.11g-2003: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications, Amendment 4: Further Higher Data Rate Extension in the 2.4 GHz Band, 2003.

Haruhisa Ichikawa received B.S., M.S. and Dr. Eng. degrees in electrical engineering from the University of Tokyo in 1974, 1976 and 1989, respectively. He joined NTT laboratories in 1976, where he was engaged in fundamental research on communications software and distributed computing. He proposed and conducted many R&D projects for software, Internet, information sharing platform, and ubiquitous networks, including business incubation. He was responsible for fundamental NTT R&D as the executive director of Science and Core Laboratory Group. He is a Professor at the University of Electro-Communications, Tokyo, since September, 2007. He is a member of IEEE, IEICE and IPSJ.

Masashi Shimizu received the B.E. and M.E. degrees in mechanical engineering from Keio University, Yokohama, in 1986, 1988, respectively. In 1988, he joined NTT Wireless System Laboratories, Yokosuka, Japan. He has been engaged in research on pointing control for deployable space antenna and surface error compensation by feed distribution control. His recent interest is active RFID and its applications. He is currently Senior Research Engineer, Supervisor, Wireless System Innovation Laboratory, NTT Network Innovation Laboratories. He is a member of IEICE.

Kazunori Akabane received his B.E. and M.E. degrees from Keio University, Yokohama, Japan, in 1994 and 1996, respectively. He joined NTT Wireless Systems Laboratories, Yokosuka, Japan, in 1996, where he has been involved in the research and development of wireless personal communication systems and software defined radio (SDR) systems. He is currently a Research Engineer at NTT Network Innovation Laboratories.

Osamu Ishida received the B.E. and M.E. degrees in electronic engineering from the University of Tokyo, Japan, in 1986 and 1988, respectively. In 1988 he joined NTT Transmission Systems Laboratories. His research interests include high-speed Ethernet transport, optical cross-connect switching, and component devices such as plainer lightwave circuits (PLC) and tunable diode lasers. He is currently a Senior Research Engineer, Supervisor of NTT Network Innovation Laboratories. He is a member of IEEE and IEICE.

Mistuo Teramoto received the B.E. and M.E degrees in engineering from Hokkaido University in 1982 and 1984, respectively. He joined NTT Electrical Communication Laboratories in 1984, where he was engaged in research on CAD system for communication devices. His present research includes data processing technologies in ubiquitous computing. He is currently Senior Research Engineer, Supervisor of NTT Network Innovation Laboratories.