Chapter 2. WIRELESS LANS AND PANS

2.1 INTRODUCTION

The field of computer networks has grown significantly in the last three decades. An interesting usage of computer networks is in offices and educational institutions, where tens (sometimes hundreds) of personal computers (PCs) are interconnected, to share resources (e.g., printers) and exchange information, using a high-bandwidth communication medium (such as the Ethernet). These privately-owned networks are known as local area networks (LANs) which come under the category of small-scale networks (networks within a single building or campus with a size of a few kilometers). To do away with the wiring associated with the interconnection of PCs in LANs, researchers have explored the possible usage of radio waves and infrared light for interconnection [1]. This has resulted in the emergence of wireless LANs (WLANs), where wireless transmission is used at the physical layer of the network. Wireless personal area networks (WPANs) are the next step down from WLANs, covering smaller areas with low power transmission, for networking of portable and mobile computing devices such as PCs, personal digital assistants (PDAs), which are essentially very small computers designed to consume as little power as possible so as to increase the lifetime of their batteries, cell phones, printers, speakers, microphones, and other consumer electronics. This chapter highlights the issues involved in the design of WLANs and PANs. It consists of the following sections:

1. Fundamentals of WLANs: The technical issues in WLANs must be understood in order to appreciate the difference between wired networks and wireless networks. The use of WLANs and their design goals are then studied. The types of WLANs, their components, and their basic functionalities are also brought out in this section.
2. IEEE 802.11 Standard: This section introduces a prominent standard in WLANs, the IEEE 802.11 standard. The medium access control (MAC) layer and the physical layer mechanisms are explained here. This section also covers some of the optional functionalities, such as security and quality of service (QoS).
3. HIPERLAN Standard: This section describes another WLAN standard, HIPER-LAN standard, which is a European standard based on radio access.
4. Bluetooth: This section deals with the Bluetooth standard, which enables personal devices to communicate with each other in the absence of infrastructure.
5. HomeRF: This section discusses the issues in home networking (HomeRF standard) and finally illustrates the technical differences between Bluetooth, HomeRF, and other technologies such as infrared [portable devices that use the infrared interface of the Infrared Data Association (IrDA) for transmission], which are the current technological alternatives in the PAN area.

2.2 FUNDAMENTALS OF WLANS

This section deals with the fundamental principles, concepts, and requirements of WLANs. This section also brings out WLAN types, their components, and some of their functionalities. In what follows, the terms "node," "station," and "terminal" are used interchangeably. While both portable terminals and mobile terminals can move from one place to another, portable terminals are accessed only when they are stationary. Mobile terminals (MTs), on the other hand, are more powerful, and can be accessed when they are in motion. WLANs aim to support truly mobile work stations.

2.2.1 Technical Issues

Here the technical issues that are encountered in the design and engineering of WLANs are discussed. In particular, the differences between wireless and wired networks, the use of WLANs, and the design goals for WLANs are studied.

Differences Between Wireless and Wired Transmission

• Address is not equivalent to physical location: In a wireless network, address refers to a particular station and this station need not be stationary. Therefore, address may not always refer to a particular geographical location.

• Dynamic topology and restricted connectivity: The mobile nodes may often go out of reach of each other. This means that network connectivity is partial at times.

• Medium boundaries are not well-defined: The exact reach of wireless signals cannot be determined accurately. It depends on various factors such as signal strength and noise levels. This means that the precise boundaries of the medium cannot be determined easily.

• Error-prone medium: Transmissions by a node in the wireless channel are affected by simultaneous transmissions by neighboring nodes that are located within the direct transmission range of the transmitting node. This means that the error rates are significantly higher in the wireless medium. Typical bit error rates (fractions of bits that are received in error) are of the order of 10^-4 in a wireless channel as against 10^-9 in fiber optic cables.

The above four factors imply that we need to build a reliable network on top of an inherently unreliable channel. This is realized in practice by having reliable protocols at the MAC layer, which hide the unreliability that is present in the physical layer.

Use of WLANs

Wireless computer networks are capable of offering versatile functionalities. WLANs are very flexible and can be configured in a variety of topologies based on the application. Some possible uses of WLANs are mentioned below.

• Users would be able to surf the Internet, check e-mail, and receive Instant Messages on the move.

• In areas affected by earthquakes or other such disasters, no suitable infrastructure may be available on the site. WLANs are handy in such locations to set up networks on the fly.

• There are many historic buildings where there has been a need to set up computer networks. In such places, wiring may not be permitted or the building design may not be conducive to efficient wiring. WLANs are very good solutions in such places.

Design Goals

The following are some of the goals which have to be achieved while designing WLANs:

• Operational simplicity: Design of wireless LANs must incorporate features to enable a mobile user to quickly set up and access network services in a simple and efficient manner.

• Power-efficient operation: The power-constrained nature of mobile computing devices such as laptops and PDAs necessitates the important requirement of WLANs operating with minimal power consumption. Therefore, the design of WLAN must incorporate power-saving features and use appropriate technologies and protocols to achieve this.

• License-free operation: One of the major factors that affects the cost of wireless access is the license fee for the spectrum in which a particular wireless access technology operates. Low cost of access is an important aspect for popularizing a WLAN technology. Hence the design of WLAN should consider the parts of the frequency spectrum (e.g., ISM band) for its operation which do not require an explicit licensing.

• Tolerance to interference: The proliferation of different wireless networking technologies both for civilian and military applications and the use of the microwave frequency spectrum for non-communication purposes (e.g., microwave ovens) have led to a significant increase in the interference level across the radio spectrum. The WLAN design should account for this and take appropriate measures by way of selecting technologies and protocols to operate in the presence of interference.

• Global usability: The design of the WLAN, the choice of technology, and the selection of the operating frequency spectrum should take into account the prevailing spectrum restrictions in countries across the world. This ensures the acceptability of the technology across the world.

• Security: The inherent broadcast nature of wireless medium adds to the requirement of security features to be included in the design of WLAN technology.

• Safety requirements: The design of WLAN technology should follow the safety requirements that can be classified into the following: (i) interference to medical and other instrumentation devices and (ii) increased power level of transmitters that can lead to health hazards. A well-designed WLAN should follow the power emission restrictions that are applicable in the given frequency spectrum.

• Quality of service requirements: Quality of service (QoS) refers to the provisioning of designated levels of performance for multimedia traffic. The design of WLAN should take into consideration the possibility of supporting a wide variety of traffic, including multimedia traffic.

• Compatibility with other technologies and applications: The interoperability among the different LANs (wired or wireless) is important for efficient communication between hosts operating with different LAN technologies. In addition to this, interoperability with existing WAN protocols such as TCP/IP of the Internet is essential to provide a seamless communication across the WANs.

2.2.2 Network Architecture

This section lists the types of WLANs, the components of a typical WLAN, and the services offered by a WLAN.

Infrastructure Based Versus Ad Hoc LANs

WLANs can be broadly classified into two types, infrastructure networks and ad hoc LANs, based on the underlying architecture.

Infrastructure networks contain special nodes called access points (APs), which are connected via existing networks. APs are special in the sense that they can interact with wireless nodes as well as with the existing wired network. The other wireless nodes, also known as mobile stations (STAs), communicate via APs. The APs also act as bridges with other networks.

Ad hoc LANs do not need any fixed infrastructure. These networks can be set up on the fly at any place. Nodes communicate directly with each other or forward messages through other nodes that are directly accessible.

Components in a Typical IEEE 802.11 Network

IEEE 802.11 is the most popular WLAN standard that defines the specification for the physical and MAC layers. The success of this standard can be understood from the fact that the revenue from the products based on this standard touched $730 million in the second quarter of the year 2003. The principles and mechanisms followed in this standard are explained later. In what follows, the basic components in a typical IEEE 802.11 WLAN [2] are listed.

The set of stations that can remain in contact (i.e., are associated) with a given AP is called a basic service set (BSS). The coverage area of an AP within which member stations (STAs or MTs) may remain in communication is called the basic service area (BSA). The stations that are a part of a BSS need to be located within the BSA of the corresponding AP. A BSS is the basic building block of the network. BSSs are connected by means of a distribution system (DS) to form an extended network.

DS refers to an existing network infrastructure. The implementation of the DS is not specified by the IEEE 802.11 standard. The services of the DS, however, are specified rigidly. This gives a lot of flexibility in the design of the DS. The APs are connected by means of the DS.

Portals are logical points through which non-IEEE 802.11 packets (wired LAN packets) enter the system. They are necessary for integrating wireless networks with the existing wired networks. Just as an AP interacts with the DS as well as the wireless nodes, the portal interacts with the wired network as well as with the DS. The BSSs, DS, and the portals together with the stations they connect constitute the extended service set (ESS). An ad hoc LAN has only one BSS. Therefore, ad hoc LANs are also known as independent basic service sets (IBSSs). It may be noted that the ESS and IBSS appear identical to the logical link control (LLC). Figure 2.1 gives a schematic picture of what a typical ESS looks like.

Figure 2.1. Extended Service Set.

Services Offered by a Typical IEEE 802.11 Network

The services offered by a typical IEEE 802.11 network can be broadly divided into two categories: AP services and STA services. The following are the AP services, which are provided by the DS:

• Association: The identity of an STA and its address should be known to the AP before the STA can transmit or receive frames on the WLAN. This is done during association, and the information is used by the AP to facilitate routing of frames.

• Reassociation: The established association is transferred from one AP to another using reassociation. This allows STAs to move from one BSS to another.

• Disassociation: When an existing association is terminated, a notification is issued by the STA or the AP. This is called disassociation, and is done when nodes leave the BSS or when nodes shut down.

• Distribution: Distribution takes care of routing frames. If the destination is in the same BSS, the frame is transmitted directly to the destination, otherwise the frame is sent via the DS.

• Integration: To send frames through non-IEEE 802.11 networks, which may have different addressing schemes or frame formats, the integration service is invoked.

The following are the STA services, which are provided by every station, including APs:

• Authentication: Authentication is done in order to establish the identity of stations to each other. The authentication schemes range from relatively insecure handshaking to public-key encryption schemes.

• Deauthentication: Deauthentication is invoked to terminate existing authentication.

• Privacy: The contents of messages may be encrypted (say, by using the WEP algorithm, which is explained later) to prevent eavesdroppers from reading the messages.

• Data delivery: IEEE 802.11 naturally provides a way to transmit and receive data. However, like Ethernet, the transmission is not guaranteed to be completely reliable.

2.3 IEEE 802.11 STANDARD

After the fundamental issues in WLANs are clearly understood, the reader is in a position to appreciate the de facto standards for WLANs. IEEE 802.11 is a prominent standard for WLANs, which is adopted by many vendors of WLAN products. A later version of this standard is the IEEE 802.11b [3], commercially known as Wi-Fi (wireless fidelity). The IEEE 802.11 standard, which deals with the physical and MAC layers in WLANs, was brought out in 1997. This standard is explained in this section.

It may be observed that IEEE 802.11 was the first WLAN standard that faced the challenge of organizing a systematic approach for defining a standard for wireless wideband local access (small-scale networks capable of transmitting data at high rates). As mentioned earlier, in contrast to other LAN standards, wireless standards need to have provisions to support mobility of nodes. The IEEE 802.11 working group had to examine connection management, link reliability management, and power management — none of which was a concern for other standards in IEEE 802. In addition, provision for security had to be introduced. For all these reasons and because of several competing proposals, it took nearly ten years for the development of IEEE 802.11, which was much longer compared to the time taken for the development of other 802 standards for the wired media. Once the overall picture and the ideas became clear, it took only a reasonable duration of time to develop the IEEE 802.11a and IEEE 802.11b enhancements. Under the IEEE 802.11 standard, MTs can operate in two modes: (i) infrastructure mode, in which MTs can communicate with one or more APs which are connected to a WLAN, and (ii) ad hoc mode, in which MTs can communicate directly with each other without using an AP.

2.3.1 Physical Layer

IEEE 802.11 supports three options for the medium to be used at the physical level — one is based on infrared [4] and the other two are based on radio transmission. The physical layer is subdivided conceptually into two parts — physical medium dependent sublayer (PMD) and physical layer convergence protocol (PLCP). PMD handles encoding, decoding, and modulation of signals and thus deals with the idiosyncrasies of the particular medium. The PLCP abstracts the functionality that the physical layer has to offer to the MAC layer. PLCP offers a service access point (SAP) that is independent of the transmission technology, and a clear channel assessment (CCA) carrier sense signal to the MAC layer. The SAP abstracts the channel which can offer up to 1 or 2 Mbps data transmission bandwidth. The CCA is used by the MAC layer to implement the CSMA/CA mechanism. The three choices for the physical layer in the original 802.11 standard are as follows: (i) frequency hopping spread spectrum (FHSS) operating in the license-free 2.4 GHz industrial, scientific, and medical (ISM) band, at data rates of 1 Mbps [using 2-level Gaussian frequency shift keying (GFSK) modulation scheme] and 2 Mbps (using 4-level GFSK); (ii) direct sequence spread spectrum (DSSS) operating in the 2.4 GHz ISM band, at data rates of 1 Mbps [using differential binary phase shift keying (DBPSK) modulation scheme] and 2 Mbps [using differential quadrature phase shift keying (DQPSK)]; (iii) infrared operating at wavelengths in 850-950 nm range, at data rates of 1 Mbps and 2 Mbps using pulse position modulation (PPM) scheme.

Carrier Sensing Mechanisms

In IEEE 802.3, sensing the channel is very simple. The receiver reads the peak voltage on the cable and compares it against a threshold. In contrast, the mechanism employed in IEEE 802.11 is relatively more complex. It is performed either physically or virtually. As mentioned earlier, the physical layer sensing is through the clear channel assessment (CCA) signal provided by the PLCP in the physical layer of the IEEE 802.11. The CCA is generated based on sensing of the air interface either by sensing the detected bits in the air or by checking the received signal strength (RSS) of the carrier against a threshold. Decisions based on the detected bits are made somewhat more slowly, but they are more reliable. Decisions based on the RSS can potentially create a false alarm caused by measuring the level of interference.

2.3.2 Basic MAC Layer Mechanisms

This section describes the MAC layer as specified by the IEEE 802.11 standard. The primary function of this layer is to arbitrate and statistically multiplex the transmission requests of various wireless stations that are operating in an area. This assumes importance because wireless transmissions are inherently broadcast in nature and contentions to access the shared channel need to be resolved prudently in order to avoid collisions, or at least to reduce the number of collisions. The MAC layer also supports many auxiliary functionalities such as offering support for roaming, authentication, and taking care of power conservation.

The basic services supported are the mandatory asynchronous data service and an optional real-time service. The asynchronous data service is supported for unicast packets as well as for multicast packets. The real-time service is supported only in infrastructure-based networks where APs control access to the shared medium.

Distributed Foundation Wireless Medium Access Control (DFWMAC)

The primary access method of IEEE 802.11 is by means of a distributed coordination function (DCF). This mandatory basic function is based on a version of carrier sense with multiple access and collision avoidance (CSMA/CA). To avoid the hidden terminal problem (which is explained later), an optional RTS-CTS mechanism is implemented. There is a second method called the point coordination function (PCF) that is implemented to provide real-time services. When the PCF is in operation, the AP controls medium access and avoids simultaneous transmissions by the nodes.

Inter-Frame Spacing (IFS)

Inter-frame spacing refers to the time interval between the transmission of two successive frames by any station. There are four types of IFS: SIFS, PIFS, DIFS, and EIFS, in order from shortest to longest. They denote priority levels of access to the medium. Shorter IFS denotes a higher priority to access the medium, because the wait time to access the medium is lower. The exact values of the IFS are obtained from the attributes specified in the physical layer management information base (PHYMIB) and are independent of the station bit rate.

• Short inter-frame spacing (SIFS) is the shortest of all the IFSs and denotes highest priority to access the medium. It is defined for short control messages such as acknowledgments for data packets and polling responses. The transmission of any packet should begin only after the channel is sensed to be idle for a minimum time period of at least SIFS.

• PCF inter-frame spacing (PIFS) is the waiting time whose value lies between SIFS and DIFS. This is used for real-time services.

• DCF inter-frame spacing (DIFS) is used by stations that are operating under the DCF mode to transmit packets. This is for asynchronous data transfer within the contention period.

• Extended inter-frame spacing (EIFS) is the longest of all the IFSs and denotes the least priority to access the medium. EIFS is used for resynchronization whenever physical layer detects incorrect MAC frame reception.

2.3.3 CSMA/CA Mechanism

Carrier sense with multiple access and collision avoidance (CSMA/CA) is the MAC layer mechanism used by IEEE 802.11 WLANs. Carrier sense with multiple access and collision detection (CSMA/CD) is a well-studied technique in IEEE 802.x wired LANs. This technique cannot be used in the context of WLANs effectively because the error rate in WLANs is much higher and allowing collisions will lead to a drastic reduction in throughput. Moreover, detecting collisions in the wireless medium is not always possible. The technique adopted here is therefore one of collision avoidance.

The Medium Access Mechanism

The basic channel access mechanism of IEEE 802.11 is shown in Figure 2.2 (a). If the medium is sensed to be idle for a duration of DIFS, the node accesses the medium for transmission. Thus the channel access delay at very light loads is equal to the DIFS. If the medium is busy, the node backs off, in which the station defers channel access by a random amount of time chosen within a contention window (CW). The value of CW can vary between CW_min and CW_max. The time intervals are all integral multiples of slot times, which are chosen judiciously using propagation delay, delay in the transmitter, and other physical layer dependent parameters. As soon as the back-off counter reaches zero and expires, the station can access the medium. During the back-off process, if a node detects a busy channel, it freezes the back-off counter and the process is resumed once the channel becomes idle for a period of DIFS. Each station executes the back-off procedure at least once between every successive transmission.

Figure 2.2. IEEE 802.11 DCF and RTS-CTS mechanism.

In the scheme discussed so far, each station has the same chances for transmitting data next time, independent of the overall waiting time for transmission. Such a system is clearly unfair. Ideally, one would like to give stations that wait longer a higher priority service in order to ensure that they are not starved. The back-off timer incorporated into the above mechanism tries to make it fair. Longer waiting stations, instead of choosing another random interval from the contention window, wait only for a residual amount of time that is specified by the back-off timer.

Contention Window Size

The size of the Contention Window (CW) is another important parameter. If the CW is small in size, then the random values will be close together and there is a high probability of packet collision. On the other hand, if the size of CW is very large, there will be some unnecessary delay because of large back-off values. Ideally, one would like the system to adapt to the current number of stations that are contending for channel access. To effect this, the truncated binary exponential back-off technique is used here, which is similar to the technique used in IEEE 802.3. The initial contention window is set to a random value between (0, CWmin) and each time a collision occurs, the CW doubles its size up to a maximum of CWmax. So at high load, the CW size is high and therefore the resolution power of the system is high. At low loads, small CW ensures low access delay. The specified values of CWmin and CWmax for different physical layer specifications are given in Table 2.1.

Table 2.1. IEEE 802.11 parameters

Acknowledgments

Acknowledgments (ACKs) must be sent for data packets in order to ensure their correct delivery. For unicast packets, the receiver accesses the medium after waiting for a SIFS and sends an ACK. Other stations have to wait for DIFS plus their backoff time. This reduces the probability of a collision. Thus higher priority is given for sending an ACK for the previously received data packet than for starting a new data packet transmission. ACK ensures the correct reception of the MAC layer frame by using cyclic redundancy checksum (CRC) technique. If no ACK is received by the sender, then a retransmission takes place. The number of retransmissions is limited, and failure is reported to the higher layer after the retransmission count exceeds this limit.

RTS-CTS Mechanism

The hidden terminal problem is a major problem that is observed in wireless networks. This is a classic example of problems arising due to incomplete topology information in wireless networks that was mentioned initially. It also highlights the non-transitive nature of wireless transmission. In some situations, one node can receive from two other nodes, which cannot hear each other. In such cases, the receiver may be bombarded by both the senders, resulting in collisions and reduced throughput. But the senders, unaware of this, may get the impression that the receiver can clearly listen to them without interference from anyone else. This is called the hidden terminal problem. To alleviate this problem, the RTS-CTS mechanism has been devised as shown in Figure 2.2 (b).

How RTS-CTS Works

The sender sends a request to send (RTS) packet to the receiver. The packet includes the receiver of the next data packet to be transmitted and the expected duration of the whole data transmission. This packet is received by all stations that can hear the sender. Every station that receives this packet will set its network allocation vector (NAV) accordingly. The NAV of a station specifies the earliest time when the station is permitted to attempt transmission. After waiting for SIFS, the intended receiver of the data packet answers with a clear to send (CTS) packet if it is ready to accept the data packet. The CTS packet contains the duration field, and all stations receiving the CTS packet also set their NAVs. These stations are within the transmission range of the receiver. The set of stations receiving the CTS packet may be different from the set of stations that received the RTS packet, which indicates the presence of some hidden terminals.

Once the RTS packet has been sent and CTS packet has been received successfully, all nodes within receiving distance from the sender and from the receiver are informed that the medium is reserved for one sender exclusively. The sender then starts data packet transmission after waiting for SIFS. The receiver, after receiving the packet, waits for another SIFS and sends the ACK. As soon as the transmission is over, the NAV in each node marks the medium as free (unless the node has meanwhile heard some other RTS/CTS) and the process can repeat again. The RTS packet is like any other packet and collisions can occur only at the beginning when RTS or CTS is being sent. Once the RTS and CTS packets are transmitted successfully, nodes that listen to the RTS or the CTS refrain from causing collision to the ensuing data transmission, because of their NAVs which will be set. The usage of RTS-CTS dialog before data packet transmission is a form of virtual carrier sensing.

Overhead Involved in RTS-CTS

It can be observed that the above mechanism is akin to reserving the medium prior to a particular data transfer sequence in order to avoid collisions during this transfer. But transmission of RTS-CTS can result in non-negligible overhead. Therefore, the RTS-CTS mechanism is used judiciously. An RTS threshold is used to determine whether to start the RTS-CTS mechanism or not. Typically, if the frame size is more than the RTS threshold, the RTS-CTS mechanism is activated and a four-way handshake (i.e., RTS-CTS-DATA-ACK) follows. If the frame size is below the RTS threshold, the nodes resort to a two-way handshake (DATA-ACK).

MAC as a State Machine

Figure 2.3 diagrammatically shows what has been discussed so far. It models the MAC layer as a finite state-machine, and shows the permissible transitions. It must be noted that the state-machine is simplistic and is given only to ease the understanding of the fundamental mechanisms at the MAC layer. The functioning of the finite state-machine is explained in what follows.

Figure 2.3. MAC state transition diagram.

If a node has a packet to send and is in the IDLE state, it goes into the WAIT_FOR_NAV state. After the on-going transmissions (if any) in the neighborhood are over, the node goes to the WAIT_FOR_DIFS state. After waiting for DIFS amount of time, if the medium continues to be idle, the station enters the BACKING_OFF state. Otherwise, the station sets its back-off counter (if the counter value is zero) and goes back to the IDLE state. During back-off, if the node senses a busy channel, the node saves the back-off counter and goes back to the IDLE state. Otherwise, it goes into one of three states. If the packet type is broadcast, the node enters the TRANSMITTING_BCAST state where it transmits the broadcast packet. If the packet type is unicast and the packet size is less than the RTS threshold, the node enters the TRANSMITTING_UNICAST state and starts transmitting data. If the packet size is greater than the RTS threshold, the node enters the TRANSMITTING_RTS state and starts transmitting the RTS packet. After the RTS transmission is over, the node enters the WAITING_FOR_CTS state. If the CTS packet is not received within a specified time, the node times out and goes back to the IDLE state, and increases the CW value exponentially up to a maximum of CW_max. If the CTS packet is received, the node enters the TRANSMITTING_UNICAST state and starts transmitting data. After the unicast packet is transmitted, the node enters the WAITING_FOR_ACK state. When the node receives the ACK, it goes back to the IDLE state and reduces the CW value to CW_min.

If a node receives an RTS packet when in IDLE state and if the NAV of the node indicates that no other on-going transmissions exist, the node enters the TRANSMITTING_CTS state and starts transmitting the CTS packet. After the CTS packet is transmitted, the node enters the WAITING_FOR_DATA state and waits for the data packet from the sender. On receiving the data packet, the node enters the TRANSMITTING_ACK state and starts transmitting the ACK for the data packet. When the ACK has been transmitted, the node goes back to the IDLE state. If the data packet is not received, the receiver returns to the IDLE state.

Fragmentation

Bit error rates in the wireless medium are much higher than in other media. The bit error rate in fiber optics is only about 10^-9, whereas in wireless, it is as large as 10^-4. One way of decreasing the frame error rate is by using shorter frames. IEEE 802.11 specifies a fragmentation mode where user data packets are split into several smaller parts transparent to the user. This will lead to shorter frames, and frame error will result in retransmission of a shorter frame. The RTS and CTS messages carry duration values for the current fragment and estimated time for the next fragment. The medium gets reserved for the successive frames until the last fragment is sent.

The length of each fragment is the same for all the fragments except the last fragment. The fragments contain information to allow the complete MAC protocol data unit (MPDU, informally referred to as packet) to be reassembled from the fragments that constitute it. The frame type, sender address, destination address, sequence control field, and indicator for more fragments to come are all present in the fragment header. The destination constructs the complete packet by reassembling the fragments in the order of the sequence number field. The receiving station ensures that all duplicate fragments are discarded and only one copy of each fragment is integrated. Acknowledgments for the duplicates may, however, be sent.

2.3.4 Other MAC Layer Functionalities

There are several other functionalities that the MAC layer provides in IEEE 802.11 WLANs. The functionalities described in this section are the point coordination function (PCF) which is used for QoS guarantees, timing synchronization, power management, and support for roaming.

Point Coordination Function

The objective of the point coordination function (PCF) is to provide guarantees on the maximum access delay, minimum transmission bandwidth, and other QoS parameters. Unlike the DCF, where the medium contention is resolved in a distributed manner, the PCF works by effecting a centralized contention resolution scheme, and is applicable only in networks where an AP polls the nodes in its BSS. A point coordinator (PC) at the AP splits the access time into super frame periods. The super frame period consists of alternating contention free periods (CFPs) and contention periods (CPs). The PC will determine which station has the right to transmit at any point of time. The PCF is essentially a polled service with the PC playing the role of the polling master. The operation of the PCF may require additional coordination to perform efficient operation in cases where multiple PCs are operating simultaneously such that their transmission ranges overlap. The IFS used by the PCF is smaller than the IFS of the frames transmitted by the DCF. This means that point-coordinated traffic will have higher priority access to the medium if DCF and PCF are concurrently in action. The PC controls frame transmissions so that contentions are eliminated over a limited period of time, that is, the CFP.

Synchronization

Synchronization of clocks of all the wireless stations is an important function to be performed by the MAC layer. Each node has an internal clock, and clocks are all synchronized by a timing synchronization function (TSF). Synchronized clocks are required for power management, PCF coordination, and frequency hopping spread spectrum (FHSS) hopping sequence synchronization. Without synchronization, clocks of the various wireless nodes in the network may not have a consistent view of the global time.

Within a BSS, quasi periodic beacon frames are transmitted by the AP, that is, one beacon frame is sent every target beacon transmission time (TBTT) and the transmission of a beacon is deferred if the medium is busy. A beacon contains a time-stamp that is used by the node to adjust its clock. The beacon also contains some management information for power optimization and roaming. Not all beacons need to be heard for achieving synchronization.

Power Management

Usage of power cords restricts the mobility that wireless nodes can potentially offer. The usage of battery-operated devices calls for power management because battery power is expensive. Stations that are always ready to receive data consume more power (the receiver current may be as high as 100 mA). The transceiver must be switched off whenever carrier sensing is not needed. But this has to be done in a manner that is transparent to the existing protocols. It is for this reason that power management is an important functionality in the MAC layer. Therefore, two states of the station are defined: sleep and awake. The sleep state refers to the state where the transceiver can not receive or send wireless signals. Longer periods in the sleep state mean that the average throughput will be low. On the other hand, shorter periods in the sleep state consume a lot of battery power and are likely to reduce battery life.

If a sender wants to communicate with a sleeping station, it has to buffer the data it wishes to send. It will have to wait until the sleeping station wakes up, and then send the data. Sleeping stations wake up periodically, when senders can announce the destinations of their buffered data frames. If any node is a destination, then that node has to stay awake until the corresponding transmission takes place.

Roaming

Each AP may have a range of up to a few hundred meters where its transmission will be heard well. The user may, however, walk around so that he goes from the BSS of one AP to the BSS of another AP. Roaming refers to providing uninterrupted service when the user walks around with a wireless station. When the station realizes that the quality of the current link is poor, it starts scanning for another AP. This scanning can be done in two ways: active scanning and passive scanning. Active scanning refers to sending a probe on each channel and waiting for a response. Passive scanning refers to listening into the medium to find other networks. The information necessary for joining the new BSS can be obtained from the beacon and probe frames.

2.3.5 Other Issues

Improvements in the IEEE 802.11 standard have been proposed to support higher data rates for voice and video traffic. Also, QoS provisioning and security issues have been addressed in extended versions of the standard. These will be discussed in the remainder of this section.

Newer Standards

The original standards for IEEE 802.11 came out in 1997 and promised a data rate of 1-2 Mbps in the license-free 2.4 GHz ISM band [5]. Since then, several improvements in technology have called for newer and better standards that offer higher data rates. This has manifested in the form of IEEE 802.11a and IEEE 802.11b standards, both of which came out in 1999. IEEE 802.11b, an extension of IEEE 802.11 DSSS scheme, defines operation in the 2.4 GHz ISM band at data rates of 5.5 Mbps and 11 Mbps, and is trademarked commercially by the Wireless Ethernet Compatibility Alliance (WECA) as Wi-Fi. It achieves high data rates due to the use of complimentary code keying (CCK). IEEE 802.11a operates in the 5 GHz band (unlicensed national information infrastructure band), and uses orthogonal frequency division multiplexing (OFDM) at the physical layer. IEEE 802.11a supports data rates up to 54 Mbps and is the fast Ethernet analogue to IEEE 802.11b.

Other IEEE 802.11 (c, d, and h) task groups are working on special regulatory and networking issues. IEEE 802.11e deals with the requirements of time-sensitive applications such as voice and video. IEEE 802.11f deals with inter-AP communication to handle roaming. IEEE 802.11g aims at providing the high speed of IEEE 802.11a in the ISM band. IEEE 802.11i deals with advanced encryption standards to support better privacy.

QoS for Voice and Video Packets

In order to offer QoS, delay-sensitive packets (such as voice and video packets) are to be given a higher priority to get ahead of less time-critical (e.g., file transfer) traffic. Several mechanisms have been proposed to offer weighted priority. Hybrid coordination function (HCF) can be used where the AP polls the stations in a weighted way in order to offer QoS. Extended DCF is another mechanism which has been proposed where the higher priority stations will choose the random back-off interval from a smaller CW. Performance of WLANs where voice and data services are integrated is studied in [6] and [7].

Wired Equivalent Privacy

Security is a very important issue in the design of WLANs. In order to provide a modest level of physical security, the wired equivalent privacy (WEP) mechanism was devised. The name WEP implies that this mechanism is aimed at providing the level of privacy that is equivalent to that of a wired LAN. Data integrity, access control, and confidentiality are the three aims of WEP. It assumes the existence of an external key management service that distributes the key sequence used by the sender. This mechanism relies on the fact that the secret key cannot be determined by brute force. However, WEP has been proven to be vulnerable if more sophisticated mechanisms are used to crack the key. It uses the pseudo-random number key generated by RSA RC4 algorithm which has been efficiently implemented in hardware as well as in software. This mechanism makes use of the fact that if we take the plain text, XOR (bit-by-bit exclusive OR) it with a pseudo-random key sequence, and then XOR the result with the same key sequence, we get back the plain text.

2.4 HIPERLAN STANDARD

The European counterparts to the IEEE 802.11 standards are the high-performance radio LAN (HIPERLAN) standards defined by the European Telecommunications Standards Institute (ETSI). It is to be noted that while the IEEE 802.11 standards can use either radio access or infrared access, the HIPERLAN standards are based on radio access only. The standards have been defined as part of the ETSI broadband radio access networks (BRAN) project. In general, broadband systems are those in which user data rates are greater than 2 Mbps (and can go up to 100s of Mbps). Four standards have been defined for wireless networks by the ETSI.

• HIPERLAN/1 is a wireless radio LAN (RLAN) without a wired infrastructure, based on one-to-one and one-to-many broadcasts. It can be used as an extension to a wired infrastructure, thus making it suited to both ad hoc and infrastructure-based networks. It employs the 5.15 GHz and the 17.1 GHz frequency bands and provides a maximum data rate of 23.5 Mbps.

• The HIPERLAN/2 standard intends to provide short-range (up to 200 m) wireless access to Internet protocol (IP), asynchronous transfer mode (ATM¹), and other infrastructure-based networks and, more importantly, to integrate WLANs into cellular systems. It employs the 5 GHz frequency band and offers a wide range of data rates from 6 Mbps to 54 Mbps. HIPERLAN/2 has been designed to meet the requirements of future wireless multimedia services.

¹ ATM networks are connection-oriented and require a connection to set up prior to transfer of information from a source to a destination. All information to be transmitted — voice, data, image, and video — is first fragmented into small, fixed-size packets known as cells. These cells are then switched and routed using packet switching principles.

• HIPERACCESS (originally called HIPERLAN/3) covers "the last mile" to the customer; it enables establishment of outdoor high-speed radio access networks, providing fixed radio connections to customer premises. HIPERACCESS provides a data rate of 25 Mbps. It can be used to connect HIPERLAN/2 deployments that are located far apart (up to 5 Km away). It offers point-to-multipoint communication.

• The HIPERLINK (originally called HIPERLAN/4) standard provides highspeed radio links for point-to-point static interconnections. This is used to connect different HIPERLAN access points or HIPERACCESS networks with high-speed links over short distances of up to 150 m. For example, the HIPERLINK can be employed to provide links between different rooms or floors within a large building. HIPERLINK operates on the 17 GHz frequency range.

Figure 2.4 shows a typical deployment of the ETSI standards. The standards excluding HIPERLAN/1 are grouped under the BRAN project. The scope of the BRAN has been to standardize the radio access network and the functions that serve as the interface to the infrastructural networks.

Figure 2.4. The ETSI-BRAN systems.

2.4.1 HIPERLAN/1

HIPERLAN/1 is a RLAN standard that was introduced by the ETSI in 1995. The standard allows nodes to be deployed either in a pre-arranged or in an ad hoc fashion. Apart from supporting node mobility, HIPERLAN/1 provides forwarding mechanisms (multi-hop routing). Thus, coverage is not limited to just the neighboring nodes. Using a clever framing scheme as explained later in this section, HIPERLAN/1 provides a data rate of around 23.5 Mbps without utilizing much power, thus having the capability to support multimedia data and asynchronous data effectively. This data rate is significantly higher than that provided by IEEE 802.11. The HIPERLAN/1 protocol stack is restricted to the two lower-most layers in the OSI reference model: the data link layer (DLL) and the physical layer. The DLL is further divided into the medium access control (MAC) sublayer and the channel access control (CAC) sublayer. The sections that follow describe the standard.

The Physical Layer

The tasks of the physical layer are modulation and demodulation of a radio carrier with a bit stream, forward error-correction mechanisms, signal strength measurement, and synchronization between the sender and the receiver. The standard uses the CCA scheme (similar to IEEE 802.11) to sense whether the channel is idle or busy.

The MAC Sublayer

The HIPERLAN/1 MAC (HM) sublayer is responsible for processing the packets from the higher layers and scheduling the packets according to the QoS requests from the higher layers specified by the HM QoS parameters. The MAC sublayer is also responsible for forwarding mechanisms, power conservation schemes, and communication confidentiality through encryption–decryption mechanisms.

Because of the absence of an infrastructure, the forwarding mechanism is needed to allow the physical extension of HIPERLAN/1 to go beyond the radio range of a single station. Topology-related data are exchanged between the nodes periodically with the help of special packets, for the purpose of forwarding.

In order to guarantee a time-bound service, the HM protocol data unit (HM-PDU) selected for channel access has to reflect the user priority and the residual lifetime of the packet (the time remaining for the packet to expire). The MAC layer computes the channel access priority for each of the PDUs following a mapping from the MAC priority to the channel access mechanism (CAM) priority. One among those PDUs which has the highest CAM priority and the least residual time will be selected for access to the channel.

The CAC Sublayer

The CAC sublayer offers a connectionless data service to the MAC sublayer. The MAC layer uses this service to specify a priority (called the CAM priority) which is the QoS parameter for the CAC layer. This is crucial in the resolution of contention in the CAM.

EY-NPMA

After a packet with an associated CAM priority has been chosen in the CAC sublayer for transmission, the next phase is to compete with packets of other nodes for channel access. The channel access mechanism is a dynamic, listen-and-then-talk protocol that is very similar to the CSMA/CA used in 802.11 and is called the elimination yield non-preemptive multiple access (EY-NPMA) mechanism. Figure 2.5 shows the operation of the EY-NPMA mechanism in which the nodes 1, 2, 3, and 4 have packets to be sent to the AP. The CAM priority for nodes 2 and 4 is higher with priority 2 followed by node 3 with priority 3, and node 1 with the least priority of 4. The prioritization phase will have k slots where k (can vary from 1 to 5 with k − 1 having higher priority than k) refers to the number of priority levels.

Figure 2.5. The operation of EY-NPMA.

The entire process of channel access occurs in the form of channel access cycles. A synchronization interval occurs after the end of every such cycle. This access cycle is comprised of three phases: prioritization, contention, and transmission.

1. Prioritization: This phase culls out nodes with packets of the highest CAM priority and lets them participate in the next phase. The prioritization phase consists of two events, namely, priority detection and priority assertion. During the priority detection period, a node listens to the channel for a number of time slots proportional to the CAM priority assigned to the packet that the node wants to send. In Figure 2.5, the nodes 2 and 4 wait for one slot and assert their priority in the second slot as they hold packets with higher priority, and nodes 3 and 1 wait for slots equal to their priority level. By listening to the channel, nodes 3 and 1 detect the existence of other nodes with higher priority and hence leave the prioritization phase. If a low-priority node has succeeded in waiting up to this slot, it enters the priority assertion period during which it sends a burst, signaling its selection to the next stage. In this process, the node(s) with the highest CAM priority will finish the prioritization phase first and hence will be selected for the next phase.
2. Contention: This phase is to eliminate as many nodes as possible, in order to minimize the collision rate during transmission. This phase extends to a maximum of 13 slots, each of the same width as that of the slots in the prioritization phase. In this phase, the nodes that transmitted a burst in the previous phase, resolve access to the channel by contention. This phase consists of two sub-phases, namely, the elimination phase and the yield phase. Nodes in this phase (nodes 2 and 4 in Figure 2.5) get to transmit a burst for a geometrically distributed number of time slots [the probability of a node's transmission extending to a slot length of k slots (where k < 12 slots) is 0.5^k+1] which is then followed by a sensing period of 1 slot. During this period, if a node detects another node's burst, it stops the contention process (node 2 in Figure 2.5). This period during which each contending node will have to listen to the channel for a slot duration is called the elimination survival identification interval. If the channel is sensed idle during this interval, the node reaches the yield phase. This period is also called elimination survival verification. This ensures that the node(s) which sent the elimination burst for the maximum number of slots will be chosen for the next phase. The next phase is the yield phase which complements the elimination phase; it involves each node listening to the channel for a number of time slots (up to a maximum of 15 slots, each with duration of the slot duration in the prioritization phase). This is in fact similar to the back-off state in which the probability of backing off for k slots is 0.1 × 0.9^k. If the channel is sensed to be idle during these slots, the node is said to be eligible for transmission. The node that waits for the shorter number of slots initiates transmission and other nodes defer their access to the next cycle to begin the process afresh.
3. Transmission: This is the final stage in the channel access where the transmission of the selected packet takes place. During this phase, the successful delivery of a data packet is acknowledged with an ACK packet.

The performance of EY-NPMA protocol suffers from major factors such as packet length, number of nodes, and the presence of hidden terminals. The efficiency of this access scheme varies from 8% to 83% with variation of packet sizes from 50 bytes to 2 Kbytes.

The above-described channel access takes place during what is known as the channel synchronization condition. The other two conditions during which channel access can take place are (a) the channel free condition, when the node senses the channel free for some amount of time and then gains access, and (b) the hidden terminal condition, when a node is eliminated from contention, but still does not sense any data transmission, indicating the presence of a hidden node.

Power Conservation Issues

The HIPERLAN/1 standard has suggested power conservation schemes at both the MAC and the physical layers.

At the MAC level, the standard suggests awake/sleep modes similar to the DFWMAC in IEEE 802.11. Two roles defined for the nodes are the p-savers (nodes that want to implement the function) and the p-supporters (neighbors to the p-saver that are deputized to aid the latter's power conservation). The p-saver can receive packets only at predetermined time intervals and is active only during those intervals, in the process saving power.

At the physical level, a framing scheme has been adopted to conserve power. The physical burst is divided into high bit rate (HBR) and low bit rate (LBR) bursts. The difference between the two bursts lies in the keying mechanisms employed for them – the HBR burst is based on Gaussian minimum shift keying (GMSK) that yields a higher bit rate, but consumes more power than frequency shift keying (FSK) used for the LBR bursts. The LBR burst contains the destination address of the frame and precedes the HBR burst. Any node receiving a packet, first reads the LBR burst. The node will read the HBR burst only if it is the destination for that frame. Otherwise, the burst is simply ignored, thereby saving the power needed to read the HBR burst.

Failure of HIPERLAN/1

In spite of the high data rate that it promised, HIPERLAN/1 standard has always been considered unsuccessful. This is because IEEE Ethernet had been prevalent and hence, for its wireless counterpart too, everybody turned toward IEEE, which came out with its IEEE 802.11 standard. As a result, hardly any manufacturer adopted the HIPERLAN/1 standard for product development. However, the standard is still studied for the stability it provides and for the fact that many of the principles followed have been adopted in the other standards. For further details on the standard, readers are referred to [8] and [9].

2.4.2 HIPERLAN/2

As seen earlier, the IEEE 802.11 standard offers data rates of 1 Mbps while the newer standard IEEE 802.11a offers rates up to 54 Mbps. However, there was a necessity to support QoS, handoff (the process of transferring an MT from one channel/AP to another), and data integrity in order to satisfy the requirements of wireless LANs. This demand was the motivation behind the emergence of HIPERLAN/2. The standard has become very popular owing to the significant support it has received from cellular manufacturers such as Nokia and Ericsson. The HIPERLAN/2 tries to integrate WLANs into the next-generation cellular systems. It aims at converging IP and ATM type services at a high data rate of 54 Mbps for indoor and outdoor applications. The HIPERLAN/2, an ATM compatible WLAN, is a connection-oriented system, which uses fixed size packets and enables QoS applications easy to implement.

The HIPERLAN/2 network has a typical topology as shown in Figure 2.6. The figure shows MTs being centrally controlled by the APs which are in turn connected to the core network (infrastructure-based network). It is to be noted that, unlike the IEEE standards, the core network for HIPERLAN/2 is not just restricted to Ethernet. Also, the AP used in HIPERLAN/2 consists of one or many transceivers called access point transceivers (APTs) which are controlled by a single access point controller (APC).

Figure 2.6. A typical deployment of HIPERLAN/2.

There are two modes of communication in a HIPERLAN/2 network, which are described by the following two environments:

• Business environment: The ad hoc architecture of HIPERLAN/1 has been extended to support a centralized mode of communication using APs. This topology corresponds to business environments. Accordingly, each AP serves a number of MTs.

• Home environment: The home environment enables a direct mode of communication between the MTs. This corresponds to an ad hoc architecture that can be operated in a plug-and-play manner. The direct mode of communication is, however, managed by a central control entity elected from among the nodes called the central controller (CC).

There are several features of HIPERLAN/2 that have attracted many a cellular manufacturer. These features are part of the discussion on the protocol stack of HIPERLAN/2 below. The HIPERLAN/2 protocol stack consists of the physical layer, convergence layer (CL), and the data link control (DLC) layer.

The Physical Layer

The physical layer is responsible for the conversion of the PDU train from the DLC layer to physical bursts that are suitable for radio transmission. HIPERLAN/2, like IEEE 802.11a, uses OFDM for transmission. The HIPERLAN/2 allows bit rates from 6 Mbps to 54 Mbps using a scheme called link adaptation. This scheme allows the selection of a suitable modulation method for the required bit rate. This scheme is unique to HIPERLAN/2 and is not available in the IEEE standards and HIPERLAN/1. More details on the physical layer can be found in [14].

The CL

The topmost layer in the HIPERLAN/2 protocol stack is the CL. The functions of the layer are to adapt the requirements of the different higher layers of the core network with the services provided by the lower layers of HIPERLAN/2, and to convert the higher layer packets into ones of fixed size that can be used by the lower layers. A CL is defined for every type of core network supported. In short, this layer is responsible for the network-independent feature of HIPERLAN/2.

The CL is classified into two types, namely, the packet-based CL and the cell-based CL. The packet-based CL processes variable-length packets (such as IEEE 802.3, IP, and IEEE 1394). The cell-based CL processes fixed-sized ATM cells. The CL has two sublayers, namely, the common part (CP) and the service-specific convergence sublayer (SSCS). The CP is independent of the core network. It allows parallel segmentation and reassembly of packets. The CP comprises of two sublayers, namely, the common part convergence sublayer (CPCS) and the segmentation and reassembly (SAR) sublayer. The CPCS processes the packets from the higher layer and adds padding and additional information, so as to be segmented in the SAR. For further information on the CP, readers are referred to [10].

The SSCS consists of functions that are specific to the core network. For example, the Ethernet SSCS has been standardized in [11] for Ethernet core networks. The SSCS adapts the different data formats to the HIPERLAN/2 DLC format. It is also responsible for mapping the QoS requests of the higher layers to the QoS parameters of HIPERLAN/2 such as data rate, delay, and jitter.

The DLC Layer

The DLC layer constitutes the logical link between the AP and the MTs. This ensures a connection-oriented communication in a HIPERLAN/2 network, in contrast to the connectionless service offered by the IEEE standards. The DLC layer is organized into three functional units, namely, the radio link control (RLC) sublayer on the control plane, the error control (EC) sublayer on the user plane, and the MAC sublayer. The following discussion describes the features of the DLC layer. For further details, readers are referred to [12] and [13].

The RLC Sublayer

The RLC sublayer takes care of most of the control procedures on the DLC layer. The tasks of the RLC can be summarized as follows.

• Association control function (ACF): The ACF handles the registration and the authentication functions of an MT with an AP within a radio cell. Only after the ACF procedure has been carried out can the MT ever communicate with the AP.

• DLC user connection control (DCC): The DCC function is used to control DLC user connections. It can set up new connections, modify existing connections, and terminate connections.

• Radio resource control (RRC): The RRC is responsible for the surveillance and efficient utilization of the available frequency resources. It performs the following tasks:

Dynamic frequency selection: This function is not available in IEEE 802.11, IEEE 802.11a, IEEE 802.11b, and HIPERLAN/1, and is thus unique to HIPERLAN/2. It allows the AP to select a channel (frequency) for communication with the MTs depending on the interferences in each channel, thereby aiding in the efficient utilization of the available frequencies.

Handoff: HIPERLAN/2 supports three types of handoff, namely, sector handoff (moving to another sector of the same antenna of an APT), radio handoff (handoff between two APTs under the same APC), and network handoff (handoff between two APs in the same network).

Power saving: Power-saving schemes much similar to those in HIPERLAN/1 and IEEE 802.11 have been implemented.

Error Control (EC)

Selective repeat (where only the specific damaged or lost frame is retransmitted) protocol is used for controlling the errors across the medium. To support QoS for stringent and delay-critical applications, a discard mechanism can be provided by specifying a maximum delay.

The MAC Sublayer

The MAC protocol is used for access to the medium, resulting in the transmission of data through that channel. However, unlike the IEEE standards and the HIPERLAN/1 in which channel access is made by sensing it, the MAC protocol follows a dynamic time division multiple access/time division duplexing (TDMA/TDD) scheme with centralized control. The protocol supports both AP-MT unicast and multicast transfer, and at the same time MT-MT peer-to-peer communication. The centralized AP scheduling provides QoS support and collision-free transmission. The MAC protocol provides a connection-oriented communication between the AP and the MT (or between MTs).

Security Issues

Elaborate security mechanisms exist in the HIPERLAN/2 system. The encryption procedure is optional and can be selected by the MT during association. Two strong encryption algorithms are offered, namely, the data encryption standard (DES) and the triple-DES algorithms.

2.5 BLUETOOTH

WLAN technology enables device connectivity to infrastructure-based services through a wireless carrier provider. However, the need for personal devices to communicate wirelessly with one another, without an established infrastructure, has led to the emergence of personal area networks (PANs). The first attempt to define a standard for PANs dates back to Ericsson's Bluetooth project² in 1994 to enable communication between mobile phones using low-power and low-cost radio interfaces. In May 1998, several companies such as Intel, IBM, Nokia, and Toshiba joined Ericsson to form the Bluetooth Special Interest Group (SIG) whose aim was to develop a de facto standard for PANs. Recently, IEEE has approved a Bluetooth-based standard (IEEE 802.15.1) for wireless personal area networks (WPANs). The standard covers only the MAC and the physical layers while the Bluetooth specification details the whole protocol stack. Bluetooth employs radio frequency (RF) technology for communication. It makes use of frequency modulation to generate radio waves in the ISM band.

² The project was named after Danish King Harald Blatand (A.D. 940-981) (who was known as Bluetooth due to his fondness for blueberries), who unified the Scandinavians by introducing Christianity.

Low power consumption of Bluetooth technology and an offered range of up to ten meters has paved the way for several usage models. One can have an interactive conference by establishing an ad hoc network of laptops. Cordless computer, instant postcard [sending digital photographs instantly (a camera is cordlessly connected to a mobile phone)], and three-in-one phone [the same phone functions as an intercom (at the office, no telephone charge), cordless phone (at home, a fixed-line charge), and mobile phone (on the move, a cellular charge)] are other indicative usage models.

2.5.1 Bluetooth Specifications

The Bluetooth specification consists of two parts: core and profiles. The core provides a common data link and physical layer to application protocols, and maximizes reusability of existing higher layer protocols. The profiles specifications classify Bluetooth applications into thirteen types. The protocol stack of Bluetooth performs the functions of locating devices, connecting other devices, and exchanging data. It is logically partitioned into three layers, namely, the transport protocol group, the middleware protocol group, and the application group.

The transport protocol group consists of the radio layer, baseband layer, link manager layer, logical link control and adaptation layer, and the host controller interface. The middleware protocol group comprises of RFCOMM, SDP, and IrDA (IrOBEX and IrMC). The application group consists of applications (profiles) using Bluetooth wireless links, such as the modem dialer and the Web-browsing client. The following sections discuss the concepts involved in the design of transport protocols in Bluetooth communications, and also provide an overview of the middleware and application layer protocols. Figure 2.7 shows the protocol stack of Bluetooth. The detailed specifications and explanation of the stack are available in [15], [16]. Readers may also refer to [17], [18] for more information.

Figure 2.7. Bluetooth protocol stack.

2.5.2 Transport Protocol Group

This group is composed of the protocols designed to allow Bluetooth devices to locate each other and to create, configure, and manage the wireless links. Design of various protocols and techniques used in Bluetooth communications has been done with the target of low power consumption and ease of operation. This shall become evident in the design choice of FHSS and the master–slave architecture. The following sections study the various protocols in this group, their purpose, their modes of operation, and other specifications.

Radio (Physical) Layer

The radio part of the specification deals with the characteristics of the transceivers and design specifications such as frequency accuracy, channel interference, and modulation characteristics. The Bluetooth system operates in the globally available ISM frequency band and the frequency modulation is GFSK. It supports 64 Kbps voice channels and asynchronous data channels with a peak rate of 1 Mbps. The data channels are either asymmetric (in one direction) or symmetric (in both directions). The Bluetooth transceiver is a FHSS system operating over a set of m channels each of width 1 MHz. In most of the countries, the value of m is 79. Frequency hopping is used and hops are made at a rapid rate across the possible 79 hops in the band, starting at 2.4 GHz and stopping at 2.480 GHz. The choice of frequency hopping has been made to provide protection against interference.

The Bluetooth air interface is based on a nominal antenna power of 0 dBm (1 mW) with extensions for operating at up to 20 dBm (100 mW) worldwide. The nominal link range is from 10 centimeters to 10 meters, but can be extended to more than 100 meters by increasing the transmit power (using the 20 dBm option). It should be noted here that a WLAN cannot use an antenna power of less than 0 dBm (1 mW) and hence an 802.11 solution might not be apt for power-constrained devices as mentioned in [19].

Baseband Layer

The key functions of this layer are frequency hop selection, connection creation, and medium access control. Bluetooth communication takes place by ad hoc creation of a network called a piconet. The address and the clock associated with each Bluetooth device are the two fundamental elements governing the formation of a piconet.

Every device is assigned a single 48-bit address which is similar to the addresses of IEEE 802.xx LAN devices. The address field is partitioned into three parts and the lower address part (LAP) is used in several baseband operations such as piconet identification, error checking, and security checks. The remaining two parts are proprietary addresses of the manufacturing organizations. LAP is assigned internally by each organization. Every device also has a 28-bit clock (called the native clock) that ticks 3,200 times per second or once every 312.5 μs. It should be noted that this is twice the normal hopping rate of 1,600 hops per second.

Piconet

The initiator for the formation of the network assumes the role of the master (of the piconet). All the other members are termed as slaves of the piconet. A piconet can have up to seven active slaves at any instant. For the purpose of identification, each active slave of the piconet is assigned a locally unique active member address AM_ADDR. Other devices could also be part of the piconet by being in the parked mode (explained later). A Bluetooth device not associated with any piconet is said to be in standby mode. Figure 2.8 shows a piconet with several devices.

Figure 2.8. A typical piconet.

Operational States

Figure 2.9 shows the state diagram of Bluetooth communications. Initially, all the devices would be in the standby mode. Then some device (called the master) could begin the inquiry and get to know the nearby devices and, if needed, join them into its piconet. After the inquiry, the device could formally be joined by paging, which is a packet-exchange process between the master and a prospective slave to inform the slave of the master's clock. If the device was already inquired, the master could get into the page state bypassing the inquiry state. Once the device finishes getting paged, it enters the connected state. This state has three power-conserving sub-states – hold, sniff, and park (described later in this section). A device in the connected state can participate in the data transmission.

Figure 2.9. Operational states.

Frequency Hopping Sequences

It is evident (in any wireless communication) that the sender and the receiver should use the same frequency for communication to take place. A frequency selection module (FSM) is present in each device to select the next frequency to be used under various circumstances. In the connected state, the clock and the address of the device (master) completely determine the hopping sequence. Different combination of inputs (clock, address) are used depending on the operational state. During the inquiry operation, the address input to FSM is a common inquiry address. This common address is needed because at the time of inquiry no device has information about the hopping sequence being followed. The address of the paged device is fed as input to the FSM for the paging state.

Communication Channel

The channel is divided into time slots, each 625 μs in length. The time slots are numbered according to the Bluetooth clock of the piconet master. A time division duplex (TDD) scheme is used where master and slave alternately transmit. The master starts its transmission in even-numbered time slots only, and the slave starts its transmission in odd-numbered time slots only. This is clearly illustrated in Figure 2.10 (a). The packet start shall be aligned with the slot start. A Bluetooth device would determine slot parity by looking at the least significant bit (LSB) in the bit representation of its clock. If LSB is set to 1, it is the possible transmission slot for the slave. A slave in normal circumstances is allowed to transmit only if in the preceding slot it has received a packet from the master. A slave should know the master's clock and address to determine the next frequency (from the FSM). This information is exchanged during paging.

Figure 2.10. Transmission of packets over a channel.

Packet-Based Communication

Bluetooth uses packet-based communication where the data to be transmitted is fragmented into packets. Only a single packet can be transmitted in each slot. A typical packet used in these communications has three components: access code, header, and payload. The main component of the access code is the address of the piconet master. All packets exchanged on the channel are identified by the master's identity. The packet will be accepted by the recipient only if the access code matches the access code corresponding to the piconet master. This also helps in resolving conflicts in the case where two piconets are operating currently on the same frequency. A slave receiving two packets in the same slot can identify its packet by examining the access code.

The packet header contains many fields such as a three-bit active slave address, a one-bit ACK/NACK for ARQ scheme [Automatic Repeat reQuest — anytime an error is detected, a negative acknowledgment (NACK) is returned and the specified frames are retransmitted], a four-bit packet type to distinguish payload types, and an eight-bit header error check code to detect errors in the header. Depending on the payload size, one, three, or five slots may be used for the packet transmission. The hop frequency which is used for the first slot is used for the remainder of the packet. While transmitting packets in multiple slots, it is important that the frequencies used in the following time slots are those that are assigned to those slots, and that they do not follow the frequency sequence that should have normally applied. This is illustrated in Figure 2.10 (b). When a device uses five slots for packet transmission, the next packet transmission is allowed in F(k+6) and not in F(k+2). Also note that the receiving time slot becomes F(k+5) as opposed to F(k+1). On this slotted channel, both synchronous and asynchronous links are supported.

Between a master and a slave there is a single asynchronous connectionless link (ACL) supported. This is the default link that would exist once a link is established between a master and a slave. Whenever a master would like to communicate, it would, and then the slave would respond. Optionally, a piconet may also support synchronous connection oriented (SCO) links. SCO link is symmetric between master and slave with reserved bandwidth and regular periodic exchange of data in the form of reserved slots. These links are essential and useful for high-priority and time-bound information such as audio and video.

Inquiry State

As shown in Figure 2.9, a device which is initially in the standby state enters the inquiry state. As its name suggests, the sole purpose of this state is to collect information about other Bluetooth devices in its vicinity. This information includes the Bluetooth address and the clock value, as these form the crux of the communication between the devices. This state is classified into three sub-states: inquiry, inquiry scan, and inquiry response.

A potential master sends an inquiry packet in the inquiry state on the inquiry hop sequence of frequencies. This sequence is determined by feeding a common address as one of the inputs to the FSM. A device (slave) that wants to be discovered will periodically enter the inquiry scan state and listen for these inquiry packets. When an inquiry message is received in the inquiry scan state, a response packet called the frequency hopping sequence (FHS) containing the responding device address must be sent. Devices respond after a random jitter to reduce the chances of collisions.

Page State

A device enters this state to invite other devices to join its piconet. A device could invite only the devices known to itself. So normally the inquiry operation would precede this state. This state also is classified into three sub-states: page, page scan, and page response.

In the page mode, the master estimates the slave's clock based on the information received during the inquiry state, to determine where in the hop sequence the slave might be listening in the page scan mode. In order to account for inaccuracies in estimation, the master also transmits the page message through frequencies immediately preceding and succeeding the estimated one. On receiving the page message, the slave enters the slave page response sub-state. It sends back a page response consisting of its ID packet which contains its device access code (DAC). Finally, the master (after receiving the response from a slave) enters the page response state and informs the slave about its clock and address so that the slave can go ahead and participate in the piconet. The slave now calculates an offset to synchronize with the master clock, and uses that to determine the hopping sequence for communication in the piconet.

Scatternets and Issues

Piconets may overlap both spatially and temporally, that is, many piconets could operate in the same area at the same time. Each piconet is characterized by a unique master and hence the piconets hop independently, each with its own channel hopping sequence as determined by the respective master. In addition, the packets carried on the channels are preceded by different channel access codes as determined by the addresses of the master devices. As more piconets are added, the probability of collisions increases, and a degradation in performance results, as is common in FHSS systems.

In this scenario, a device can participate in two or more overlaying piconets by the process of time sharing. To participate on the proper channel, it should use the associated master device address and proper clock offset. A Bluetooth unit can act as a slave in several piconets, but as a master in only a single piconet. A group of piconets in which connections exist between different piconets is called a scatternet (Figure 2.11).

Figure 2.11. A typical scatternet.

When a device changes its role and takes part in different piconets, it is bound to lead to a situation in which some slots remain unused (for synchronization). This implies that complete utilization of the available bandwidth is not achieved. An interesting proposition at this juncture would be to unite the timings of the whole of the scatternet as explained in [19]. But this may lead to an increase in the probability of packets colliding.

Another important issue is the timing that a device would be missing by participating in more than one piconet. A master that is missing from a piconet (by momentarily becoming a slave in another piconet) may miss polling slaves and must ensure that it does not miss beacons from its slaves. Similarly, a slave (by becoming a master or slave in another piconet) that is missing from a piconet could appear to its master to have gone out of range or to be connected through a poor-quality link.

Link Manager Protocol

Link manager protocol (LMP) is responsible for setting and maintaining the properties of the Bluetooth link. Currently, the major functionality of this layer is power management and security management. It also provides minimal QoS support by allowing control over parameters such as delay and delay jitter. Normally, a paging device is the default master of the piconet, but, depending on the usage scenario, the roles of the master and a slave could be switched and this is coordinated by exchange of LMP packets.

Power Management

The Bluetooth units can be in several modes of operation during the connection state, namely, active mode, sniff mode, hold mode, and park mode. These modes are now described.

• Active mode: In this mode, the Bluetooth unit actively participates in the piconet. Various optimizations are provided to save power. For instance, if the master informs the slave when it will be addressed, the slave may sleep until then. The active slaves are polled by the master for transmissions.

• Sniff mode: This is a low-power mode in which the listening activity of the slave is reduced. The LMP in the master issues a command to the slave to enter the sniff mode, giving it a sniff interval, and the slave listens for transmissions only at these fixed intervals.

• Hold mode: In this mode, the slave temporarily does not support ACL packets on the channel (possible SCO links will still be supported). In this mode, capacity is made available for performing other functions such as scanning, paging, inquiring, or attending another piconet.

• Park mode: This is a very low-power mode. The slave gives up its active member address and is given an eight-bit parked member address. The slave, however, stays synchronized to the channel. Any messages to be sent to a parked member are sent over the broadcast channel characterized by an active member address of all zeros. Apart from saving power, the park mode helps the master to have more than seven slaves (limited by the three-bit active member address space) in the piconet.

Bluetooth Security

In Bluetooth communications, devices may be authenticated and links may be encrypted. The authentication of devices is carried out by means of a challenge-response mechanism which is based on a commonly shared secret link key generated through a user-provided personal identification number (PIN). The authentication starts with the transmission of an LMP challenge packet and ends with the verification of result returned by the claimant. Optionally, the link between them could also be encrypted.

Logical Link Control and Adaptation Protocol (L2CAP)

This is the protocol with which most applications would interact unless a host controller is used. L2CAP supports protocol multiplexing to give the abstraction to each of the several applications running in the higher layers as if it alone is being run. Since the data packets defined by the baseband protocol are limited in size, L2CAP also segments large packets from higher layers such as RFCOMM or SDP into multiple smaller packets prior to their transmission over the channel. Similarly, multiple received baseband packets may be reassembled into a single larger L2CAP packet. This protocol provides QoS on certain parameters such as peak bandwidth, latency, and delay variation when the link is established between two Bluetooth units.

Host Controller Interface

This is the optional interface layer, provided between the higher (above LMP) and lower layers of the Bluetooth protocol stack, for accessing the Bluetooth hardware capabilities. Whenever the higher layers are implemented on the motherboard of a host device, this layer is needed. Such an approach could prove beneficial as the spare capacity of the host device (say, a personal computer) could be utilized. The specification defines details such as the different packet types as seen by this layer. Command packets that are used by the host to control the device, event packets that are used by the device to inform the host of the changes, and data packets come under this category.

2.5.3 Middleware Protocol Group

The basic functionality of the middleware protocol group is to present to the application layers a standard interface that may be used for communicating across the transport layer, that is, the applications need not know the transport layer's complexities, they can just use the application programming interfaces (APIs) or higher level functions provided by the middleware protocols. This group consists of the RFCOMM layer, service discovery protocol (SDP), IrDA interoperability protocols, telephony control specification (TCS), and audio.

The RFCOMM layer presents a virtual serial port to applications using the serial interface. Any application which is using the serial port can work seamlessly on Bluetooth devices. RFCOMM uses an L2CAP connection to establish a link between two devices. In the case of Bluetooth devices, there is no device which will be static and hence services offered by the other devices have to be discovered. This is achieved by using the service discovery protocol (SDP) of the Bluetooth protocol stack. Service discovery makes the device self-configured without manual intervention.

The IrDA interoperability protocol is not for communication between Bluetooth devices and Infrared devices. It is only for the existing IrDA applications to work on Bluetooth devices without any changes. The main protocols in the IrDA set are IrOBEX (IrDA object exchange) for exchanging objects between two devices and IrMC (infrared mobile communications) for synchronization.

Audio is the distinguishing part of Bluetooth. Audio is given the highest priority and is directly carried over the baseband at 64 Kbps so that a very good quality of voice is provided. Another important point to note here is that audio is actually not a layer of the protocol stack, but only a specific packet format that can be transmitted directly over the SCO links of the baseband layer.

Telephony control is implemented using the telephony control specification – binary (TCS-BIN) protocol. TCS defines three major functional areas: call control, group management, and connectionless TCS. Call control is used to set up calls which can be subsequently used to carry voice and data traffic. TCS operates in both point-to-point and point-to-multipoint configurations. One of the main concepts of TCS is that of the wireless user group (WUG). Group management enables multiple telephone extensions, call forwarding, and group calls. For example, consider multiple handsets and a single base set. When a call comes in to the base set, all the multiple handsets can receive this call. In a similar fashion, calls can also be forwarded.

The functionalities of TCS include configuration distribution and fast intermember access. Configuration distribution is the mechanism used to find the information about the other members in a group. Fast intermember access is a method for two slaves to create a new piconet. A WUG member uses the information from the configuration distribution and determines another member which it wants to contact. Then it sends the device's information to the master, which forwards it to this device. The contacted device then responds with its device address and clock information and places itself in a page scan state. Then the master contacts the device initiating the communication. This device now pages the contacted device and forms a new piconet. This explains how a new piconet is formed between two slaves with the help of the master.

In all the above cases, a connection-oriented channel is established. To exchange simple information such as adjusting volume or signaling information, establishing such a channel is overkill and hence connectionless TCS has been provided for having a connectionless channel.

2.5.4 Bluetooth Profiles

These profiles have been developed to promote interoperability among the many implementations of the Bluetooth protocol stack. Each Bluetooth profile specification has been defined to provide a clear and transparent standard that can be used to implement a specific user end function. Two Bluetooth devices can achieve a common functionality only if both devices support identical profiles. For example, a cellular phone and a headset both have to support the Bluetooth headset profile for the headset to work with the phone. The Bluetooth profiles spring up from the usage models. In all, 13 profiles have been listed and these can be broadly classified into the following four categories:

1. Generic profiles: The Generic access profile, which is not really an application, provides a way to establish and maintain secure links between the master and the slaves. The service discovery profile enables users to access SDP to find out which applications (Bluetooth services) are supported by a specific device.
2. Telephony profiles: The cordless telephony profile is designed for three-in-one phones. The Intercom profile supports two-way voice communication between two Bluetooth devices within range of each other. The Headset profile specifies how Bluetooth can provide a wireless connection to a headset (with earphones/microphones) for use with a computer or a mobile phone.
3. Networking profiles: The LAN Access profile enables Bluetooth devices to either connect to a LAN through APs or form a small wireless LAN among themselves. The dial-up networking profile is designed to provide dial-up connections via Bluetooth-enabled mobile phones. The FAX profile, very similar to the dial-up networking profile, enables computers to send and receive faxes via a Bluetooth-enabled mobile phone.
4. Serial and object exchange profiles: The serial port profile emulates a serial line (RS232 and USB serial ports) for (legacy) applications that require a serial line. The other profiles, generic object exchange, object push, file transfer, and synchronization, are for exchanging objects between two wireless devices.

Bluetooth is the first wireless technology which has actually tried to attempt to make all the household consumer electronics devices follow one particular communication paradigm. It has been partially successful, but it does have its limitations. Bluetooth communication currently does not provide support for routing. It should be noted that some research efforts are under way to accommodate this in the Bluetooth specification. Once the routing provision is given, inter-piconet communication could be enhanced. The issues of handoffs also have not yet been dealt with till now. Although master–slave architecture has aided low cost, the master becomes the bottleneck for the whole piconet in terms of performance, fault tolerance, and bandwidth utilization. Most importantly, Bluetooth communication takes place in the same frequency band as that of WLAN and hence robust coexistence solutions need to be developed to avoid interference. The technology is still under development. Currently, there are nearly 1,800 adopter companies which are contributing toward the development of the technology.

2.6 HOMERF

Wireless home networking represents the use of the radio frequency (RF) spectrum to transmit voice and data in confined areas such as homes and small offices. One of the visionary concepts that home networking intends to achieve is the establishment of communication between home appliances such as computers, TVs, telephones, refrigerators, and air conditioners. Wireless home networks have an edge over their wired counterparts because features such as flexibility (enabling of file and drive sharing) and interoperability that exist in the wired networks are coupled with those in the wireless domain, namely, simplicity of installation and mobility.

The HIPERLAN/2, as mentioned earlier, has provisions for direct communication between the mobile terminals (the home environment). The home environment enables election of a central controller (CC) which coordinates the communication process. This environment is helpful in setting up home networks. Apart from this, an industry consortium known as the Home RF Working Group has developed a technology that is termed HomeRF. This technology intends to integrate devices used in homes into a single network and utilize RF links for communication. HomeRF is a strong competitor to Bluetooth as it operates in the ISM band.

Technical Features

The HomeRF provides data rates of 1.6 Mbps, a little higher than the Bluetooth rate, supporting both infrastructure-based and ad hoc communications. It provides a guaranteed QoS delivery to voice-only devices and best-effort delivery for data-only devices. The devices need to be plug-and-play enabled; this needs automatic device discovery and identification in the network. A typical HomeRF network consists of resource providers (through which communication to various resources such as the cable modem and phone lines is effected), and the devices connected to them (such as the cordless phone, printers, and file servers). The HomeRF technology follows a protocol called the shared wireless access protocol (SWAP). The protocol is used to set up a network that provides access to a public network telephone, the Internet (data), entertainment networks (cable television, digital audio, and video), transfer and sharing of data resources (such as disks and printers), and home control and automation.

The SWAP has been derived from the IEEE 802.11 and the European digitally enhanced cordless telephony (DECT) standards. It employs a hybrid TDMA/CSMA scheme for channel access. While TDMA handles isochronous transmission (similar to synchronous transmission, isochronous transmission is also used for multimedia communication where both the schemes have stringent timing constraints, but isochronous transmission is not as rigid as synchronous transmission in which data streams are delivered only at specific intervals), CSMA supports asynchronous transmission (in a manner similar to that of the IEEE 802.11 standard), thereby making the actual framing structure more complex. The SWAP, however, differs from the IEEE 802.11 specification by not having the RTS-CTS handshake since it is more economical to do away with the expensive handshake; moreover, the hidden terminal problem does not pose a serious threat in the case of small-scale networks such as the home networks.

The SWAP can support up to 127 devices, each identified uniquely by a 48-bit network identifier. The supported devices can fall into one (or more) of the following four basic types:

• Connection point that provides a gateway to the public switched telephone network (PSTN), hence supporting voice and data services.

• Asynchronous data node that uses the CSMA/CA mechanism to communicate with other nodes.

• Voice node that uses TDMA for communication.

• Voice and data node that can use both CSMA/CA and TDMA for channel access.

Home networking also needs strong security measures to safeguard against potential eavesdroppers. That is the reason why SWAP uses strong algorithms such as Blowfish encryption. HomeRF also includes support for optional packet compression which provides a trade-off between bandwidth and power consumption.

Because of its complex (hybrid) MAC and higher capability physical layer, the cost of HomeRF devices is higher than that of Bluetooth devices. HomeRF Version 2.0, released recently, offers higher data rates (up to 10 Mbps by using wider channels in the ISM band through FHSS).

Infrared

The infrared technology (IrDA) uses the infrared region of the light for communication [20]. Some of the characteristics of these communications are as follows:

• The infrared rays can be blocked by obstacles, such as walls and buildings.

• The effective range of infrared communications is about one meter. But when high power is used, it is possible to achieve better ranges.

• The power consumed by infrared devices is extremely low.

• Data rates of 4 Mbps are easily achievable using infrared communications.

• The cost of infrared devices is very low compared to that of Bluetooth devices.

Although the restriction of line of sight (LoS) is there on the infrared devices, they are extremely popular because they are cheap and consume less power. The infrared technology has been prevalent for a longer time than Bluetooth wireless communications. So it has more widespread usage than Bluetooth. Table 2.2 compares the technical features of Bluetooth, HomeRF, and IrDA technologies.

Table 2.2. Illustrative comparison among Bluetooth, HomeRF, and IrDA technologies

2.7 SUMMARY

This chapter has discussed networks of a small scale which use tetherless communication (ability to move without restriction due to wires). In most of these networks, communication has been using radio waves of appropriate wavelength. In certain scenarios, infrared light has been used for transmitting data. Efficient protocols are used at the physical and MAC layers in order to make the transition from wired to wireless networks appear seamless to the higher layers on the protocol stack. This chapter has dealt with two prominent standards for WLANs, IEEE 802.11 and ETSI HIPERLAN, and two technological alternatives, Bluetooth and HomeRF, in the PAN area. Table 2.3 compares the technical features of different WLAN and PAN standards discussed in this chapter.

Table 2.3. A brief comparison among the different WLAN and PAN standards discussed in this chapter

The deployment considerations and choice of appropriate technology for a WLAN are network coverage, bandwidth requirement, expected traffic load, target users, security requirements, QoS requirements, scenario of deployment, and, finally, the cost of deployment.

2.8 PROBLEMS

1. Think of four scenarios where wireless networks can replace wired networks in order to improve the efficiency of people at their workplace. Briefly describe how in each case a wireless network will fit the role better than a wired network.
2. Compare and contrast infrastructure networks with ad hoc networks. Give example situations where one type of network is preferred to the other.
3. Does the IEEE 802.11 standard specify the implementation of the distribution system (DS)? If not, explain how DS is characterized.
4. Match the following pairs:

   

5. Why is a back-off timer used in the CSMA/CA mechanism?
6. How can DCF and PCF coexist in the BSS?
7. How are fragments of one packet sent via a WLAN that uses 802.11 without getting interleaved with fragments of other packets?
8. What are the functions of beacon frames?
9. What design trade-offs have to be considered when designing stations capable of conserving power?
10. Match the following pairs:

    

11. Give two points for and against the use of infrared and radio as a physical layer medium in a WLAN.
12. Choose the correct alternative from the choices enclosed in the parentheses.
    1. The power conservation problem in WLANs is that stations receive data (in bursts / constantly) but remain in an idle receive state (sporadically / constantly) which dominates the LAN adapter power consumption.
    2. There are two types of authentication schemes in IEEE 802.11 — the default is (shared key authentication / open system authentication) whereas (shared key / open system) provides a greater amount of security.
    3. The (DCF / PCF) mechanism is available only for infrastructure networks. In this mechanism, the AP organizes a periodical (CFP / CP) for the time-bounded information.
    4. The (FHSS / DSSS) is easier for implementation because the sampling rate is of the order of the symbol rate of 1 Mbps. The (FHSS / DSSS) implementation provides a better coverage and a more stable signal because of its wider bandwidth.
    5. The (IEEE 802.11 / HIPERLAN-2) camp is a connectionless WLAN camp that evolved from data-oriented computer communications. Its counterpart is the (HIPERLAN-2 / IEEE 802.11) camp that is more focused on connection-based WLANs addressing the needs of voice-oriented cellular telephony.
13. Why do we have four address fields in IEEE 802.11 MAC as against only two in IEEE 802.3 MAC frame?
14. Name the three MAC services provided by the IEEE 802.11 that are not provided in the traditional LANs, such as 802.3.
15. Determine the transfer time of a 22 KB file with a mobile data network (a) with a transmission rate of 10 Kbps and (b) repeat the same for 802.11 WLAN operating at 2 Mbps. (c) What is the length of the file that WLAN can carry in the time that mobile data service carried a 20 KB file? (d) What do you infer from the answers to the above questions?
16. What is normalized propagation delay? Determine the normalized propagation delay for the following: (a) IEEE 802.3 Ethernet (b) IEEE 802.11. Assume 802.11 LAN provides a coverage of 200 meters.
17. Discuss the deployment scenarios for various HIPERLAN standards in the ETSI BRAN system.
18. What is the probability that two HIPERLANs will have the same ID? Comment about the value that you have obtained.
19. What are the features of HIPERLAN/1 MAC sublayer that support QoS?
20. Compare the EY-NPMA and the CSMA/CA mechanisms.
21. Observe Table 2.4 and determine which node will get the chance of sending its packet.

    Table 2.4. EY-NPMA scheme

    

22. How is synchronization achieved in the LBR-HBR data burst in HIPER-LAN/1?
23. Compare the handoff procedures of the HIPERLAN/2 and the IEEE 802.11 standards.
24. What do you think are the advantages of using a FHSS for the operation of Bluetooth devices?
25. Assume that in one slot in Bluetooth 256 bits of payload could be transmitted. How many slots are needed if the payload size is (a) 512 bits, (b) 728 bits, and (c) 1,024 bits. Assume that the non-payload portions do not change.
26. If a master leaves the piconet, what could possibly happen?
27. Why is a device constrained to act as a master in at most one piconet?
28. With reference to the discussion on scatternets, illustrate the trade-off of synchronization and bandwidth using a timing diagram. (Hint: Consider the case when a Bluetooth device acts as a master in one piconet and as a slave in another.)
29. In the serial communications, the baud rate of transmission is specified. Legacy applications on Bluetooth devices specify the baud rate. In the case of the old serial communications, the data rate is the specified baud rate, but in the case of Bluetooth serial communications, it is not so. Why?
30. Based on your understanding of the Bluetooth protocol stack, suggest a possible implementation of FTP over Bluetooth without using TCP/IP.
31. Why is the RTS-CTS handshake avoided in the HomeRF technology?

BIBLIOGRAPHY

[1] F. Gfeller, "INFRANET: Infrared Microbroadcasting Network for In-House Data Communication," Proceedings of 7th European Conference on Optical Communication 1981, pp. P27-1–P27-4, September 1981.

[2] B. P Crow, I. Widjaja, L. G. Kim, and P. T. Sakai, "IEEE 802.11 Wireless Local Area Networks," IEEE Communications Magazine, vol. 35, no. 9, pp. 116-126, September 1997.

[3] R. V. Nee, G. Awater, M. Morikura, H. Takanashi, M. Webster, and K. W. Halford, "New High Rate Wireless LAN Standards," IEEE Communications Magazine, vol. 37, no. 12, pp. 82-88, December 1999.

[4] R. T. Valadas, A. R. Tavares, A. M. D. Duarte, A. C. Moreira, and C. T. Lomba, "The Infrared Physical Layer of the IEEE 802.11 Standard for Wireless Local Area Networks," IEEE Communications Magazine, vol. 36, no. 12, pp. 106-114, December 1998.

[5] B. Tuch, "An ISM Band Spread Spectrum Local Area Network WaveLAN," Proceedings of IEEE Workshop on WLANs 1991, pp. 103-111, May 1991.

[6] A. Zahedi and K. Pahlavan, "Capacity of a WLAN with Voice and Data Services," IEEE Transactions on Communications, vol. 48, no. 7, pp. 1160-1170, July 2000.

[7] J. Feigin, K. Pahlavan, and M. Ylianttila, "Hardware Fitted Modeling and Simulation of VoIP over a Wireless LAN," Proceedings of IEEE VTS Fall 2000, vol. 3, pp. 1431-1438, September 2000.

[8] Radio Equipment and Systems (RES); High PErformance Radio Local Area Networks (HIPERLAN) Type 1, Requirements and Architecture for Wireless ATM Access and Interconnection, 1998 (http://www.etsi.org).

[9] G. Anastasi, L. Lenzini, and E. Mingozzi, "Stability and Performance Analysis of HIPERLAN," Proceedings of IEEE INFOCOM 1998, vol. 1, pp. 134-141, March 1998.

[10] Broadband Radio Access Networks (BRAN); HIgh PErformance Radio Local Area Networks (HIPERLAN) Type 2; Packet Based Convergence Layer; Part 1: Common Part, 2000 (http://www.etsi.org).

[11] Broadband Radio Access Networks (BRAN); HIgh PErformance Radio Local Area Networks (HIPERLAN) Type 2; Packet Based Convergence Layer; Part 2: Ethernet Service Specific Sublayer (SSCS), 2001 (http://www.etsi.org).

[12] Broadband Radio Access Networks (BRAN); HIgh PErformance Radio Local Area Networks (HIPERLAN) Type 2; Data Link Control (DLC) Layer; Part 1: Basic Data Transport Functions, 2001 (http://www.etsi.org).

[13] Broadband Radio Access Networks (BRAN); HIgh PErformance Radio Local Area Networks (HIPERLAN) Type 2, Data Link Control (DLC) Layer, Part 2: RLC Sublayer, 2002 (http://www.etsi.org).

[14] Broadband Radio Access Networks (BRAN); HIgh PErformance Radio Local Area Networks (HIPERLAN) Type 2; Physical (PHY) Layer, 2001 (http://www.etsi.org).

[15] http://www.bluetooth.com/

[16] B. A. Miller and C. Bisdikian, Bluetooth Revealed: The Insider's Guide to an Open Specification for Global Wireless Communications, Prentice Hall PTR, New Jersey, November 2001.

[17] C. Bisdikian, "An Overview of Bluetooth Wireless Technology," IEEE Communications Magazine, vol. 39, no. 12, pp. 86-94, December 2001.

[18] http://www.palowireless.com

[19] J. Bray and C. F. Sturman, Bluetooth: Connect Without Cables, Prentice Hall PTR, New Jersey, December 2001.

[20] http://www.irda.org/