Wimax handbook download

Video Audio icon An illustration of an audio speaker. Audio Software icon An illustration of a 3. Software Images icon An illustration of two photographs. Images Donate icon An illustration of a heart shape Donate Ellipses icon An illustration of text ellipses.

WiMAX handbook : building EMBED for wordpress. Want more? This will establish the secondary management connection of the SS and determine capabilities related to connection setup and MAC operation.

The version of IP used on the secondary management connection is also determined during registration. Connection Setup Now comes the connection setup, where data the content actually flows. Service flows are characterized by a set of QoS parameters, such as those for latency and jitter. To most efficiently utilize network resources, such as bandwidth and memory, WiMAX adopts a two-phase activation model in which resources assigned to a particular admitted service flow may not be actually committed until the service flow is activated.

In addition, the BS or the SS can dynamically establish service flows. Authorization Request and Authentication Information contains X. With successful authorization, SS regusters with the network 4.

The establishment of service flows is performed via a three-way handshaking protocol in which the request for service flow establishment is responded to and the response acknowledged. In addition to supporting dynamic service establishment, WiMAX supports dynamic service changes in which service flow parameters are renegotiated. These service flow changes follow a three-way handshaking protocol similar to the one dynamic service flow establishment uses.

The particular burst profiles used on a channel are chosen based on a number of factors, such as rain region and equipment capabilities. This is shown in Figure The SS performs the choice before and during initial ranging based on received DL signal quality measurements. Harsher environmental conditions, such as rain fades, can force the SS to request a more robust burst profile.

Alternatively, exceptionally good weather may allow an SS to temporarily operate with a more efficient burst profile. The BS still has ultimate control of the change. In this case, the BS may periodically allocate a station maintenance interval to the SS. The order of the burst profile change actions is different when transitioning to a more robust burst profile than when transitioning to a less robust one. The standard takes advantage of the fact that an SS is always required to listen to more robust portions of the DL as well as the profile that was negotiated.

Each scheduling service is associated with a set of rules imposed on the BS scheduler responsible for allocating the UL capacity and the request-grant protocol between the SS and the BS. The service provider preprovisions the services by entering the service flow information into the service flow database.

When the SS enters the BS by completing the network entry and authentication procedure, the BS downloads the service flow information from the service flow database. Figure provides an example of how the service flow information is populated.

Each SS has two service flows, identified by sfIndex, with the associated QoS parameters that are identified by qosIndex 1 and 2, respectively. Based on the MAC address, the BS will be able to find the service flow information that has been preprovisioned. Chapter 4 50 sfIndex and , with the preprovisioned service flow information.

This can be seen in Figure The service flows will then be available for the customer to send data traffic. The process begins with ranging and negotiation between the BS and SS followed by authentication and registration.

This design is noted for its robust nature. Networking is difficult enough in a predictable, managed wired environment. Talk about dropped packets! QoS refers, simply put, to reducing latency and jitter and avoiding dropping packets.

This chapter alleviates those fears by addressing both legacy- and WiMAX-specific fixes to ensure carrier-grade performance in an otherwise hostile environment. For instance, voice and video require low latency but tolerate some error rate. In contrast, generic data applications cannot tolerate error, but latency is not critical. The standard accommodates voice, video, and other data transmissions by using appropriate features in the MAC layer; this is more efficient than using these features in layers of control overlaid on the MAC.

In short, applying more bandwidth to the right channel at the right time reduces latency and improves QoS. The WiMAX standard supports adaptive modulation, effectively balancing different data rates and link quality.

The modulation method may be adjusted almost instantaneously for optimum data transfer. Adaptive modulation allows efficient use of bandwidth and a broader customer base. FDD, the legacy duplexing method, has been widely deployed in cellular telephony.

It requires two channel pairs, one for transmission and one for reception, with some frequency separation between them to mitigate self-interference.

In continuous FDD, the upstream and downstream channels are located on separate frequencies, and all CPE stations can transmit and receive simultaneously. The downstream channel is always on, and all stations are always listening to it. Traffic is sent on this channel in a broadcast manner using TDM.

In burst FDD, the upstream and downstream channels are located on separate frequencies. In contrast to continuous FDD, not all stations can transmit and receive simultaneously. Those that can transmit and receive simultaneously are referred to as full-duplex capable stations while those that cannot are referred to as half-duplex capable stations. A TDD frame has a fixed duration and contains one downstream subframe and one upstream subframe. The two subframes are separated by a guard time called transition gap TG , and the bandwidth that is allocated to each subframe is adaptive.

The TDD subframe is illustrated in Figure The PHY control field is used for physical information, such as the slot boundaries, destined for all stations. Payload data is encrypted, but message headers are unencrypted. This variation uses burst single-carrier modulation with adaptive burst profiling in which transmission parameters, including the modulation and coding schemes, may be adjusted individually to each SS on a frame-by-frame basis.

Randomization is performed for spectral shaping and to ensure bit transitions for clock recovery. Without FEC, error correction would require the retransmission of whole blocks or frames of data, resulting in added latency and a subsequent decline in QoS.

Throw more bandwidth at it! Throughput and latency are two essentials for network performance. Whereas throughput is the quantity of data that can pass from source to destination in a specific time, round-trip latency is the time it takes for a single data transaction to occur the time between requesting data and receiving it.

Latency can also be thought of as the time it takes from data send-off on one end to data retrieval on the other from one user to the other. Therefore, the better throughput bandwidth management, the better the QoS. How is that achieved? The two classes of SS allow a trade-off between simplicity and efficiency. RLC and other management protocols use bandwidth explicitly allocated to the management connections. It will typically, but need not, use the bandwidth for the connection that requested it.

For instance, if the QoS situation at the SS has changed since the last request, the SS has the option of sending the higher QoS data along with a request to replace this bandwidth stolen from a lower QoS connection.

This is detailed in Table This method uses less bandwidth. Furthermore, acknowledged protocols can take additional time, potentially adding delay. In the self-correcting protocol, all of these anomalies are treated similarly. After a time-out appropriate for the QoS of the connection or immediately, if the bandwidth was stolen by the SS for another purpose , the SS simply requests again.

For efficiency, most bandwidth requests are incremental; that is, the SS asks for more bandwidth for a connection. The SS has many ways to request bandwidth, combining the determinism of unicast polling with the responsiveness of contention-based requests and the efficiency of unsolicited bandwidth. For continuous bandwidth demand, the SS need not request bandwidth; the BS grants it unsolicited.

Bandwidth allocation and polling methods are detailed in Table To short-circuit the normal polling cycle, any SS with a connection running UGS can use the poll-me bit in the grant management Table Bandwidth Allocation Polling Methods Term Description Unicast polls Used for inactive stations and active stations that have explicitly requested to be polled. If an inactive station does not require bandwidth allocation, it responds to the poll by returning a request for 0 bytes.

Multicast and broadcast polls Used to poll a group of inactive stations when there is insufficient bandwidth to poll the stations individually. When a multicast group is polled, the members of the group that require bandwidth allocation respond to the poll. They use the contention resolution algorithm to resolve any conflicts that arise from two or more stations transmitting at the same time. If a station does not need bandwidth allocation, it does nothing; it is not allowed to respond with a bandwidth allocation of zero, as with the case of the individual poll.

Station initiated polls Used by stations to request that the BS poll them to request bandwidth allocation. Stations with active unsolicited grant service connections typically use the poll. A station initiating this type of poll sets a bit in the MAC header called the poll-me bit, typically to request to be polled more frequently in order to satisfy the QoS of the connection. When the base station receives the frame with the poll-me bit set, it polls the station individually.

Source: Ibe 60 Chapter 5 subheader to inform the BS that it needs to be polled for bandwidth needs on another connection. The BS may choose to save bandwidth by polling SSs that have unsolicited grant services only when they have set the poll-me bit. A more conventional way to request bandwidth is to send a bandwidth request MAC PDU that consists of simply the bandwidth request header and no payload.

GPC terminals can send it in either a request interval or a data grant interval allocated to their basic connection. A closely related method of requesting data is to use a grant management subheader to piggyback a request for additional bandwidth for the same connection within a MAC PDU. UGS is tailored for carrying services that generate fixed units of data periodically.

Here the BS schedules regularly, in a preemptive manner, grants of the size negotiated at connection setup without an explicit request from the SS. This eliminates the overhead and latency of bandwidth requests in order to meet the delay and delay jitter requirements of the underlying service. A practical limit on the delay jitter is set by the frame duration. If more stringent jitter requirements are to be met, output buffering is needed.

When used with UGS, the grant management subheader includes the poll-me bit as well as the slip indicator flag, which allows the SS to report that the transmission queue is backlogged due to factors such as lost grants or clock skew between the WiMAX system and the outside network. The BS, upon detecting the slip indicator flag, can allocate some additional bandwidth to the SS, allowing it to recover the normal queue state. Connections configured with UGS are not allowed to utilize random access opportunities for requests.

The real-time polling service is designed to meet the needs of services that are dynamic in nature but offers periodic dedicated request opportunities to meet real-time requirements. Providing fixed-size data grants at periodic intervals eliminates the overhead and latency associated with requesting transmission channels. Real-time polling service Designed to support real-time service flows that generate variable-size data packets, such as MPEG video, on a periodic basis.

This service is for stations that support realtime service when the flow is active and periodic unicast polls when the flow is inactive.

Non-real-time polling service Designed to support non-real-time flows that require variable-size data grants, such as FTP, on a regular basis. The service offers unicast polls on a regular basis to ensure that flows receive request opportunity even during network congestion. Best-effort service Designed to provide efficient service to besteffort traffic. Source: Ibe, Fixed Broadband Wireless Access Networks and Services increased, but this capacity is granted only according to the real need of the connection.

The real-time polling service is well suited for connections carrying services such as VoIP or streaming video or audio. The non-real-time polling service is almost identical to the realtime polling service except that connections may utilize random access transmit opportunities for sending bandwidth requests. Typically, services carried on these connections tolerate longer delays Chapter 5 62 and are rather insensitive to delay jitter.

The non-real-time polling service is suitable for Internet access with a minimum guaranteed rate. A best-effort service has also been defined. Neither throughput nor delay guarantees are provided. The SS sends requests for bandwidth in either random access slots or dedicated transmission opportunities.

The occurrence of dedicated opportunities is subject to network load, and the SS cannot rely on their presence. What Is FFT? Electromagnetic waves have sines and cosines and are analog in nature while digital data is a stream of 1s and 0s resulting in square waves.

How then can digital data be sent via an analog transmission? If 1, square waves are sent every second, the frequency components of sine waves are summed 1 KHz, 3 KHz, 5 KHz, and so on. Fast Fourier Transform is illustrated in Figure As the bit rate increases, the square wave frequency increases and the width of the square waves decreases.

FFT makes these computations more efficient by reducing the computation to NlogN. Very simply put, FFT makes the transmission of digital data square waves over the airwaves more efficient. Figure illustrates modulation schemes. By using a robust modulation scheme, WiMAX delivers high throughput at long ranges with a high level of spectral efficiency that is also tolerant of signal reflections. Dynamic adaptive modulation allows the BS to trade throughput for range.

For example, if the BS cannot establish a robust link to a distant subscriber using the highest order modulation scheme, QAM, the modulation order is Figure Modulation schemes focus the signal over distance. Throughput declines with distance Ex. Figure demonstrates how modulation schemes ensure throughput over distance. In general the greater the number of bits transmitted per symbol, the higher the data rate is for a given bandwidth.

However, the higher the number of bits per symbol, the more susceptible the scheme is to intersymbol interference ISI and noise.

Generally the signal-to-noise ratio SNR requirements of an environment determine the modulation method to be used in the environment. For this reason, where signals are expected to be resistant to noise and other impairments over long transmission distances, QPSK is the normal choice.

Essentially, these have been modulated with amplitude of zero. Although it would seem that combining the inverse FFT outputs at the transmitter would create interference between subcarriers, the orthogonal spacing allows the receiver to perfectly separate out each subcarrier. Figure illustrates the process at the receiver. The FFT output is then serialized into a single stream of data for decoding.

Note that when M is less than N, in other words fewer than N subcarriers are used at the transmitter, the receiver only serializes the M subcarriers with data. This translates into a spectrum efficiency of 3. If five of these 20 MHz channels are contained within the 5. With channel reuse and through sectorization, the total capacity from one BS site could potentially exceed 1 Gbps.

This is especially important in licensed spectrum use, where bandwidth and spectrum can be expensive. Here, OFDM delivers more data per spectrum dollar. QoS: Error Correction and Interleaving Error correcting coding builds redundancy into the transmitted data stream.

This redundancy allows bits that are in error or even missing to be corrected. The simplest example would be to simply repeat Kevin F. This is known as a repetition code. Although the repetition code is simple in structure, more sophisticated forms of redundancy are typically used because they can achieve a higher level of error correction. For OFDM, error correction coding means that a portion of each information bit is carried on a number of subcarriers; thus, if any of these subcarriers has been weakened, the information bit can still arrive intact.

Interleaving is the other mechanism used in OFDM systems to combat the increased error rate on the weakened subcarriers. Interleaving is a deterministic process that changes the order of transmitted bits. For OFDM systems, this means that bits that were adjacent in time are transmitted on subcarriers that are spaced out in frequency. Thus errors generated on weakened subcarriers are spread out in time; that is, a few long bursts of errors are converted into many short bursts.

Error correcting codes then correct the resulting short bursts of errors. A service flow is a unidirectional flow of packets that is provided a particular QoS see Chapter 4. The primary purpose of the QoS features defined here is to define transmission ordering and scheduling on the air interface. However, these features often need to work in conjunction with mechanisms beyond the air interface in order to provide end-to-end QoS or to police the behavior of SSs.

Service flows in both the UL and DL direction may exist without actually being activated to carry traffic. A service flow is characterized by a set of QoS Parameters, such as latency, jitter, and throughput assurances. The three types of service flows are listed in Table A service flow has at least an SFID and an associated direction. ProvisionedQoSParamSet A QoS parameter set provisioned via means outside of the scope of the standard, such as the network management system. The principal resource to be reserved is bandwidth.

This set also includes an additional memory or time-based resource required to subsequently activate the flow. Only an active service flow may forward packets. Authorization Module A logical function within the BS that approves or denies every change to QoS Parameters and Classifiers associated with a service flow. Service Flow Description Provisioned This service flow is known via provisioning by, for example, the network management system.

Some other mechanism has provisioned or may have signaled admitted service flows. Chapter 5 70 The Object Model The major objects of the architecture are represented by named rectangles, as illustrated in Figure Each object has a number of attributes; the attribute names that uniquely identify the object are underlined.

Optional attributes are denoted with brackets. The relationship between the number of objects is marked at each end of the associated line between the objects. The service flow is the central concept of the MAC protocol. It is uniquely identified by a big SFID. Service flows may be in either the UL or DL direction. Admitted and active service flows are mapped to a bit CID.

The service class is an optional object that may be implemented at the BS. A service class is defined in the BS to have a particular QoS parameter set.

The QoS parameter sets of a service flow may contain a reference to the service class name as a macro that selects all of the QoS parameters of the service class. The service flow QoS parameter sets may augment and even override the QoS parameter settings of the service class, subject to authorization by the BS. Service Classes The service class performs two functions. First, it allows operators to shift configuring service flows from the provisioning server to the BS.

Operators provision the SSs with the service class name; full implementation of the name is configured at the BS. This allows operators to modify the implementation of a given service to local circumstances without changing SS provisioning.

Second, it allows higher-layer protocols to create a service flow by its service class name. Such changes include requesting an admission control decision for example, setting the AdmittedQoSParamSet and requesting activation of a service flow Chapter 5 72 for example, setting the ActiveQoSParamSet.

The Authorization Module also checks reduction requests regarding the resources to be admitted or activated. This is further defined in Table The BS will be capable of caching the Provisional QoS parameter set and will be able to use this information to authorize dynamic flows that are a subset of the Provisional QoS parameter set.

Types of Service Flows The three types of service flows are described in Table Admission and activation requests for these provisioned service flows shall be permitted as long as the Admitted QoS parameter set is a subset of the Provisioned QoS parameter set, and the Active QoS parameter set is a subset of the Admitted QoS parameter set.

Requests to change the Provisioned QoS parameter set will be refused, as will requests to create new dynamic service flows. Static authorization defines a static system where all possible services are defined in the initial configuration of each SS. Dynamic authorization Communicates through a separate interface to an independent policy server that provides authorization module with advance notice of upcoming admission and activation requests and specifies proper authorization action to be taken on requests.

The Authorization Module then checks admission and activation requests from an SS to ensure the ActiveQoSParamSet being requested is a subset of the set provided by the policy server. Admission and activation requests from an SS that are signaled in advance by the external policy server are permitted. Enabling service flows follows the transfer of the operational parameters. The network assigns an SFID to provisional service flows.

Admitted service flows A two-phase activation model. First, the resources for a call are admitted; once the end-to-end negotiation is completed, the resources are activated. Service Flow Management Service flows may be created, changed, or deleted. This is accomplished through a series of MAC management messages listed in Table As Figure illustrates, the null state implies no service flow exists that matches the SFID in a message. In steady-state operation, a service flow resides in a nominal state.

As the transmission is over free space, it is important that the QoS measures account for what is perhaps the most difficult of datacom environments. Chapter 6 78 Interference—Some Assumptions The primary objection to wireless systems is the concern that there are or will soon be too many operators on the same frequency, which will cause so much interference that the technology will become unusable. This issue is not that simple.

Although this scenario may already be evident in the case of Wi-Fi variants largely limited to the 2. WiMAX currently has no problems, only solutions. This act primarily focused on three parameters: location, frequency, and power. The technology of the time did not permit consideration of a fourth element: time. In the modern sense, one might consider that a spectrum used by cell phones in a metropolitan area dense population with millions of users would command a very high price at a spectrum auction.

It is entirely possible that the wireless service provider may find a very low cost licensed spectrum and enjoy a protected spectrum, which will largely negate the concern over interference from other broadcasters the purpose of the Radio Act of in the first place. More specifically, it is the temperature equivalent of the RF power available at a receiving antenna per unit bandwidth, measured in units of degrees Kelvin.

Interference temperature density can be measured for particular frequencies using a reference antenna with known gain. Thereafter, it can be treated as a signal propagation variable independent of receiving antenna characteristics.

As illustrated in Figure , interference temperature measurements can be taken at receiver locations throughout the service areas of protected communications systems, thus estimating the real-time conditions of the RF environment.

Interference Temperature It matters what the signal level is here! These forms of interference manifest themselves as shown in Figure Figure illustrates a simplified example of the power spectrum of the desired signal and CoCh interference. In the case of a wider CoCh interferer as shown , only a portion of its power will fall within the receiver filter bandwidth. An out-of-channel interferer is also shown. Here, two sets of parameters determine the total level of interference.

This can be treated as CoCh interference. It cannot be removed at the receiver; its level is determined at the interfering transmitter. By characterizing the power spectral density psd of sidelobes and output noise floor with respect to the main lobe of a signal, this form of interference can be approximately computed similarly to the CoCh interference calculation, with an additional attenuation factor due to the suppression of this spectral energy with respect to the main lobe of the interfering signal.

Figure details the relationship of these lobes to the transmitter. Figure Main lobe, side lobes, and back lobe Side Lobes Back Lobe Main Lobe Base Station Chapter 6 82 Second, the receiver filter of the victim receiver does not completely suppress the main lobe of the interferer. No filter is ideal, and residual power passing through the stopband of the filter can be treated as additive to the CoCh interference present.

The performance of the victim receiver in rejecting out-of-channel signals, sometimes referred to as blocking performance, determines the level of this form of interference. Under the circumstances where a sharing agreement between operators does not exist or has not been concluded and where service areas are in close proximity, a coordination process should be employed.

Given that frequency spread, a for-profit service provider would be wise to consider a low-cost licensed frequency and avoid altogether the discussion of interference from other service providers.

The purpose of licensed frequency is to protect a broadcaster from other broadcasters interfering with his or her transmission. This is the original intent of the Radio Act of Recent changes in FCC policy now dictate that spectrum holders may resell their unused spectrum to other broadcasters, thus opening that spectrum to other operators.

The FCC even hints at forcing the resell of unused spectrum. The specifications for industrial, scientific, and medical ISM and unlicensed national information infrastructure U-NII stipulate multiple channels or frequencies. If interference is encountered on one frequency, the broadcaster can merely switch frequencies to a channel that is not being interfered with. This means that only three channels channels 1, 6, 11 do NOT overlap. Table indicates the channels of the unlicensed ISM band.

It is important to note that none of these channels overlap. A fundamental concept in any communications system is the link budget, a summation of all the gains and losses in a communications system. The link budget results in the transmit power required to present a signal with a given SNR at the receiver to achieve a target bit error rate BER.

That is, the interfering signal becomes too weak to present interference. In addition, if the distance between the BS and the subscriber device is greater than optimal, the signal weakens over the distance and becomes susceptible to interference, as the interfering signal is greater than the desired signal.

Figure illustrates coverage area using a series of cells. If the power level of the interfering signal gets close to the power level of the intended WiMAX signal, then interference will occur. The simplest solution is to increase the power level of the WiMAX signal in order to overcome the interfering signal. The limitation here is that the service provider must not interfere with licensed spectrum operators on similar unlikely spectrum.

A number of challenges arise from within a wireless network due to the nature of wireless transmissions. These sources of interference include multipath interference and channel noise. Both can be engineered out of the network. Chapter 6 86 Multipath Distortion and Fade Margin Multipath occurs when waves emitted by the transmitter travel along a different path and interfere destructively with waves traveling on a direct line-ofsight path.

This is sometimes referred to as signal fading. This phenomenon occurs because waves traveling along different paths may be completely out of phase when they reach the antenna, thereby canceling each other.

Because signal cancellation is almost never complete, one method of overcoming this problem is to transmit more power. Severe fading due to multipath can result in a signal reduction of more than 30dB. It is therefore essential to provide adequate link margin to overcome this loss when designing a wireless system.

Failure to do so will adversely affect reliability. The amount of extra RF power radiated to overcome this phenomenon is referred to as fade margin.

This distortion occurs at a receiver when objects in the environment reflect a part of the transmitted signal energy. Figure illustrates one such multipath scenario from a WMAN environment. Multipath-reflected signals arrive at the receiver with different amplitudes, different phases, and different time delays. Depending on the relative phase change between reflected paths, individual frequency components will add constructively and destructively. Consequently, a filter representing the multipath channel shapes the frequency domain of the received signal.

In other words, the receiver may see some frequencies in the transmitted signal that are attenuated and others that have a relative gain. Reflected Path Base Station Direct Path In the time domain, the receiver sees multiple copies of the signal with different time delays.

The time difference between two paths often means that different symbols will overlap or smear into each other and create ISI. Thus, designers building WLAN architectures must deal with distortion in the demodulator. OFDM relies on multiple narrowband subcarriers. In multipath environments, the subcarriers located at frequencies attenuated by multipath will be received with lower signal strength.

The lower signal strength leads to an increased error rate for the bits transmitted on these weakened subcarriers. Fortunately for most multipath environments, this affects only a small number of subcarriers and, therefore, only increases the error rate on a portion of the transmitted data stream. Furthermore, the robustness of OFDM in multipath can be dramatically improved with interleaving and error correction coding. Intersymbol interference is illustrated in Figure OFDM handles this type of multipath distortion by adding a guard interval to each symbol.

The guard interval is typically a cyclic or periodic extension of the basic OFDM symbol. In other words, it looks like the rest of the symbol but conveys no new information. Because no new information is conveyed, the receiver can ignore the guard interval and still be able to separate and decode the subcarriers.

When the guard interval is designed to be longer than any smearing due to the multipath channel, the receiver is able to eliminate ISI distortion by discarding the unneeded guard interval.

Hence, ISI is removed with virtually no added receiver complexity. It is important to note that discarding the guard interval does impact noise performance because the guard interval reduces the amount of energy available at the receiver for channel symbol decoding. In addition, it reduces the data rate, as no new information is contained in the added guard interval. Thus a good system design will make the guard interval as short as possible while maintaining sufficient multipath protection.

Single carrier systems could remove ISI by adding a guard interval between each symbol. However, this has a much more severe impact on the data rate for single carrier systems than it does for OFDM. Because OFDM uses a bundle of narrowband subcarriers, it obtains high data rates with a relatively long symbol period because the frequency width of the subcarrier is inversely proportional to the symbol duration.

Consequently, adding a short guard interval has little impact on the data rate. Single carrier systems with bandwidths equivalent to OFDM must use much shorter duration symbols. Hence, adding a guard interval equal to the channel smearing has a much greater impact on data rate. Because the cancellation of radio waves is geometry dependent, using two or more antennas separated by at least half of a wavelength can drastically mitigate this problem.

On acquisition of a signal, the receiver checks each antenna and simply selects the antenna with the best signal quality. This reduces but does not eliminate the required link margin that would otherwise be needed for a system that does not employ diversity. You've discovered a title that's missing from our library. Can you help donate a copy?

When you buy books using these links the Internet Archive may earn a small commission. Open Library is a project of the Internet Archive , a c 3 non-profit. See more about this book on Archive. This edition doesn't have a description yet.

Can you add one? Previews available in: English. Add another edition? Copy and paste this code into your Wikipedia page. Need help? Donate this book to the Internet Archive library.