Selected Papers: Wide Area Ubiquitous Network
Wide Area Ubiquitous Network Service System
The core network of NTT's wide area ubiquitous network (WAUN) should be designed considering scalability and support for low-performance terminals because the infrastructure of the ubiquitous society must economically handle an extremely large number of mobile low-power-consumption terminals distributed over a very wide area. In this paper, we describe the structure of the function distribution and a protocol suitable for WAUN.
NTT has proposed the wide area ubiquitous network (WAUN) as a new wireless network for building a ubiquitous society –. WAUN is an infrastructure for ubiquitously networked small computers, including sensors and actuators, that enables us to be networked anywhere and anytime with anybody and anything and enjoy a convenient life. There are some conventional approaches to building such a network for the ubiquitous society by, for example, extending existing networks such as the optical access network or the mobile telephone network by using ad-hoc network technology. However, ad-hoc networks were intended for building local area networks, so this technology is insufficient for building a ubiquitous network as an infrastructure that anyone can use. The infrastructure of the ubiquitous society must economically accommodate an extremely large number of terminals, for example microminiature sensors and actuators, which feature mobility, low performance, and low electricity consumption, distributed over a very wide area. WAUN requires a protocol for the wireless interface that is suitable for low-performance low-power-consumption terminals similar to RF-ID (radio frequency identification) tags that run on batteries. Its cell radius will be a few kilometers in order to accommodate many terminals, far more than the existing number of personal computers and cellular phones in each cell.
To make WAUN a reality, the WAUN core network must be designed considering scalability and support for low-performance terminals. For scalability, the WAUN core network will utilize function distribution, which is regarded as a characteristic of WAUN for achieving scalability. In addition, WAUN will use a simpler protocol than conventional cellular phone networks and IP (Internet protocol) networks and use an architecture in which the core network performs most of the functions currently performed by the terminals.
2. System configuration
The composition of WAUN is shown in Fig. 1. The WAUN core network consists of access points (APs), a management database (MDB), radio access network servers (RANSs), and IP gateways (IP-GWs).
An AP has a wide coverage (several kilometers) and can accommodate a lot of USs, which consist of wireless terminals and terminal equipment (TE). TEs are, for example microminiature sensors and actuators. Wired terminals are IP terminals that communicate with wireless terminals in WAUN and are accommodated through an IP-GW. A wired terminal receives data acquired by a TE and transmits commands to the TE. WAUN works as a middle box between a wireless terminal and a wired terminal. It provides a data upload service from the TE to the wired terminal and a data download service from the wired terminal to the TE.
A RANS works as a server to control the US (providing authentication, authorization, and location management). A US connects to different RANSs depending on its location. In contrast, an IP-GW always connects to the same wired terminal regardless of the location of the US. WAUN has an architecture for distributing the functions of the core network devices. When the radio accommodation design needs to be changed in response to a change in AP placement or an increase in the number of radio channels, etc., it is possible to change just the RANS settings and increase or reduce the number of RANSs. On the other hand, when there is a change in the use of a wired terminal, it is possible to change just the settings of the IP-GW and increase or reduce the number of IP-GWs. The WAUN architecture has sufficient flexibility to handle scalability requirements.
WAUN will accommodate a large number of USs and receive a small amount of data in a limited radio frequency band. These characteristics require the communication overhead to be small and communication opportunities to be obtained fairly by USs. The WAUN protocol should be designed considering these needs.
3.1 Protocol features
The structure of the protocol for WAUN is shown in Fig. 2. Two representative characteristics of the transmission protocol are described below.
1) WAUN core network controls communication between a US and wired terminal
In an IP network, a terminal itself decides when to transmit a message. A terminal makes a packet with a suitable header that includes the destination IP address, and the network carries the packet and delivers it to the destination terminal by routing based on the destination IP address. In contrast, WAUN uses a protocol in which the core network decides the destination terminal when a US sends data. There are three reasons for doing this.
2) WAUN uses several identifiers (IDs) for a US, and an ID sent over a radio link is assigned temporarily and often updated
WAUN uses different IDs on the radio links and in the wired network. The ID to use in the wired network is assigned in a fixed manner independent of the location of the US and it is long enough that a US can be distinguished uniquely while it is using a WAUN service. The ID to use on the radio link is temporary and often updated, and it is long enough to distinguish each US uniquely in the current WAUN. This is done because a US should be prevented from performing tracking for a stranger and because it enables transmission channels to be used effectively.
3.2 Basic protocols
This section describes some representative basic protocols (location registration, upload from a US to a wired terminal, and download from a wired terminal to a US). In WAUN, two kinds of data communication sequences are prescribed: the download sequence from a wired terminal to a US and the upload sequence from a US to a wired terminal. On the other hand, WAUN does not prescribe direct communication between two USs because a US has insufficient ability to judge a destination address by itself or to change it.
1) Location registration protocol
This prescribes two kinds of location registration protocols (Fig. 3): the protocols for when a US moves to a different AP accommodated by the same RANS and moves to a different RANS. In WAUN, the MDB stores master profile information about all USs. This information is used for security and for authentication, authorization and accounting (AAA). When a RANS receives a location register request from a US and the RANS has no profile information for the US concerned, it acquires the information from the MDB. After that, the local RANS controls the location management without reacquiring it from the MDB. Using two kinds of location registration protocols should reduce the load on the MDB and shorten the time needed for communication.
2) Upload/download protocols
The protocols for uploading data from a US to a wired terminal and downloading data from a wired terminal to a US are shown in Figs. 4 and 5, respectively.
This paper introduced the core network architecture of the WAUN that NTT has proposed as a new infrastructure for building a ubiquitous society. We described the structure of the function distribution and a protocol suitable for WAUN, which can economically handle an extremely large number of mobile low-power-consumption terminals distributed over a very wide area. We will prove the feasibility of WAUN through a field trial in future.