The preceding sections detailed the various functions that the SNDCF performs. However, the order of operation is important and to ensure maximum efficiency, and compatibility, data must be operated on in the prescribed order. Within the SNDCF sublayer, the data transformation operations are depicted in Figure 5.18.
When data packets are passed to the SNDCF, the compression algorithms are applied. Following that, the data is segmented. This order of operation avoids segments that are then altered in size due to compression. Next, the segments are encrypted.
The encryption must be performed after data compression. In a way, data compression and encryption operate at odds to each other. Data compression looks for and takes advantage of patterns in the data stream. Encryption on the other hand, attempts to "randomize" the data. Therefore, encrypted data do not gain much from data compression techniques. In the CDPD system, data must be compressed prior to encryption, to allow achievement of the greatest efficiency.
Figure 5.19 further shows how all the data transformations are linked together within the CDPD system. The Network layer data packet is passed to the SNDCF. After compression, segmentation and encryption, the encrypted segments are passed to the data link layer. The Data Link Layer adds the proper framing headers and passes the sequence of frames to the MAC layer. The MAC layer concatenates the frames into a bit stream with frame flags between the data frames. Bit stuffing is performed to guarantee data transparency. This data bit stream is then broken into data blocks of 282 bits. The Reed-Solomon Forward Error Correction Code is appended to the 282 bit data block to form Reed Solomon (63,47) blocks of 378 bits. These Reed-Solomon blocks are then transmitted over the radio channel in accordance with the MAC protocol engine, taking into account the contention resolution and error recovery mechanisms.
It looks like a lot of processing and manipulation. However, each operation is necessary and addresses a specific characteristic of the radio data link.
The preceding sections have dealt with how the mobile device accesses the CDPD network. The emphasis has been on establishment and operation of a data link through a shared RF channel. However, the topic of how the mobile locates the "proper" RF channel to use has not been addressed.
To handle this topic, we must first define what is meant by "proper". In principle, a CDPD coverage area boundary is coterminous with the cell boundary perceived by cellular telephone users, in both physical space and frequency space. This is depicted in Figure 5.20. In addition, cellular telephone users should not have to be aware of the presence of CDPD. The most important result of these two requirements is that CDPD transmissions must not interfere with cellular telephone service.
In the development of the CDPD System Specification Release 1.1, the specification team examined a collection of cell coverage scenarios. These included differing terrain, vegetation and population density. From the diverse set of data collected, it became obvious that the theoretical view of cell boundary definition is unrealistic. An example of a real world cell is illustrated in Figure 5.21. From the illustration one can see that if the mobile operates with a selection threshold of -90 dB, it may transmit on the Cell 2 channel while well into the center of Cell 1. On the other hand, if the threshold is set at -80 dB, then the mobile may enter areas where both Cell 1 and Cell 2 are considered unacceptable. This thus creates a coverage hole.
From the various measurements, it became obvious that a single variable algorithm, such as received signal strength, would not satisfy all variations. So instead of using an algorithm that relied on a single parameter, the specifications adopts an approach that provides the M-ES with a large collection of parameters. These parameters help direct the M-ES to make the decision most appropriate for the coverage area characteristics.
The basis for the CDPD Radio Resource Management mechanism is that the M-ES selects the best channel. By this we mean that the M-ES must locate the strongest CDPD RF channel instead of settling on any channel with sufficient signal strength. Specific terrain effects may result in the M-ES successfully receiving the forward channel from an adjacent cell with good performance. This condition may confuse the M-ES into acquiring and operating on the channel from the distant cell. However, this condition results in the M-ES causing unacceptable interference with other CDPD mobiles and cellular telephones operating on the same RF channel frequency. By requiring the M-ES to select the best or strongest channel, the device will locate the stronger local cell's channel even though the distant channel provides an adequate signal level. This is illustrated in Figure 5.22.
In addition to selecting the best channel, the CDPD specification allows for some decision modifying parameters. For example, the "select the best channel" algorithm ideally would create a boundary between two cells at the locus of points where the signal strength from both cells are equal. As the mobile device crosses this imaginary equi-power line5.21 , it would switch to operate on the channel from the adjacent cell. However, this may not be desirable behaviour for the specific terrain. For example, if this imaginary line falls on a major thoroughfare, then mobile units travelling on this road will be continually moving from one cell to another. This generates undesirable traffic overhead. To address this condition, a hysteresis value is broadcast to the M-ESs. This RSSI Hysteresis value instructs the mobile devices to stay on its current channel until the difference between the current channel and the adjacent channel exceeds this broadcast value. An example of the "sticky" region using a 10 dB is illustrated in Figure 5.23.
Another way to address the above scenario is to relocate the boundary such that it does not fall on the roadway. This is accomplished in the CDPD specification by the use of a RSSI Bias value. This bias value is used by the M-ES when it compares the signal strength of the current cell and the adjacent cell. A negative RSSI Bias value instructs the M-ES to apply greater preference to the current cell's signal strength, thus effectively increasing the current cell's size5.22 . This also means that the current cell's coverage size can be reduced by a positive RSSI Bias value. An example of the use of the RSSI Bias value is shown in Figure 5.24.
The last parameter for radio resource management is concerned with interference management and not with cell selection. This is the Maximum Power Level parameter. This parameter is used by the network to ensure that any mobile operating within a particular cell must not exceed the broadcast Maximum Power Level. In CDPD, the mobiles dynamically adjust their transmission power level based on the signal strength it measures from the forward channel. The concept is that as the mobile moves away from the cell site antenna, the signal it receives weakens. This also means that for the base site to receive the mobile adequately, the mobile must increase its transmission power. The reverse is true also. As the mobile moves closer to the base site, it reduces its transmission power.
So why do we need a maximum power level limit? Once again, terrain effects come into play. In various parts of the country, there are cells situated in valleys between plateaus or mesas. The cell site may be situated at or near the bottom of the valley. As the mobile moves away from the base site and towards the crest of the mesa, it must increase its transmission power. However, if it is operating at peak power as it moves over onto the mesa, the flat terrain will allow the mobile's transmission to irradiate far and wide. This could result in unacceptable interference in distant cells using the same RF channel. Under these circumstances, the mobiles operating in the valley cell must be governed with a maximum power level chosen to minimize the unwanted interference effects on the mesa.
In the CDPD system, the mobile device has the responsibility selecting the "best" RF channel to use. However, there are two issues to address. First, the mobile device must possess the necessary data to make a valid decision. Unfortunately, much of this data is not directly accessible by the mobile device. The network infrastructure must pass the pertinent data to the mobiles.
The second issue is one of efficiency. Since the mobile is required to always operate on the "best" channel, it must frequently compare the current channel with possible alternate candidates. Unfortunately, most, if not all mobile devices will only have a single receiver, which means that when a mobile is assessing an alternate channel, it is unable to receive data broadcast on its normal data channel. Therefore, the algorithm must contain mechanisms to improve the efficiency of alternate channel scanning.
The CDPD system address these two concerns with a series of broadcast messages. The messages are transmitted periodically and consists of the following:
¥ Channel Stream Identification message
¥ Cell Configuration message
¥ Channel Quality Parameters message
¥ Channel Access Parameters message
The Channel Stream Identification message is illustrated in Figure 5.25. Its purpose is to identify the channel stream to all M-ESs able to receive its signal. The content of the message uniquely identifies the channel stream by the Cell Identifier and channel stream identfier tuple. The remainder of the PDU specifies the business relationship identifiers. Multiple CDPD service providers may select to operate under a single brand identity. The Wide Area Service Identifier (WASI) indicates this "branding" of the service. The Service Provider Identifier (SPI) specifies the business entity that is operating this network. These parameters have more to do with access control than with cell selection during movement of the mobile. The business relationship identifiers may be used by the mobile to determine if the current network is an appropriate one to access. The remaining fields contain the Power Product and Max Power Level parameters. These are used to direct the proper setting of dynamic power control mechanisms within the mobile devices.
The Cell Configuration message is illustrated in Figure 5.26. The MDBS of a cell transmits multiple cell configuration messages. It sends one Cell Configuration message containing data for itself and one Cell Configuration message for each and every one of its neighbor cells. Each Cell Configuration message contains a cell identifier. For the cell identified in the message, the following are indicated:
¥ The reference channel to be used for received signal strength comparisons. The need for a reference channel is discussed in greater detail in the following sections.
¥ The Effective Radiated Power Delta (ERP Delta). This is used to adjust for the difference between the power level of the reference channel and the actual power level of the CDPD data channel. The need for this parameter and its use is discussed in greater detail in the following sections.
¥ The Received Signal Strength Bias (RSSI Bias). This value is used to weigh the signal strength comparison between the CDPD channel in the current cell versus that of the indicated adjacent cell.
¥ The Power Product is the mobile dynamic power control parameter to be used in the cell identified in the message.
¥ The Maximum Power Level is the mobile dynamic power control parameter to be used in the cell identified in the message.
¥ The CDPD Channel List is a list of all RF channel numbers allocated for CDPD use in the cell identified.
¥ Other miscellaneous indicators to allow mobile optimization of the scanning function.
With the information from a full set of this data for all adjacent cells, a mobile device can quickly and effectively determine if the current channel is the most appropriate one. The basic scan algorithm involves a measurement of the signal strength of the reference channel of the adjacent cell. The received signal strength of the adjacent cell reference channel is then adjusted by the ERP Bias value associated with that cell5.23 . The resulting adjusted RSSI is compared against the received signal strength of the current cell. If the comparison indicates the adjacent cell is preferred, the mobile will need to perform a cell transfer procedure to move to that adjacent cell.
So, what is a "reference channel" and why is it necessary? In the CDPD system, each adjacent cell may use any one of approximately 300 RF channels. To scan all of them would be very time consuming. Given the Cell Configuration message about the adjacent cells, a mobile will need to scan the RF channels allocated for CDPD use. However, since the CDPD channel may change RF channels to avoid voice communications traffic, a scan action may miss the channel being used for CDPD data. The radio resource management mechanism identifies a continuously transmitting RF channel located at the same CDPD base site. This reliable RF signal allows the mobiles to quickly measure the signal strength from the specified adjacent cell. Unfortunately, these continuously transmitting RF channels may be operating at a different power level than the CDPD data channel it is used to represent. To allow the M-ES to account for this difference, an adjustment value called the Effective Radiated Power Delta (ERP Delta) is associated with the reference channel. After the mobile measures the signal strength of the reference channel, it subtracts the ERP Delta value. The result is a good approximation of the signal strength of the actual CDPD data channel from that base site.
The above answers the question as to why there is use a reference channel. It doesn't explain what a reference channel may be. The two characteristics a reference channel must have are that it must be a continuously transmitting signal and that it must be co-located at the CDPD base site. Both of these requirements may be met by either a dedicated CDPD data channel or a cellular telephone control channel.
Once the pertinent RF channel information is known about the possible adjacent channels, it is important to ensure that the mobile devices perform the scans at a rate appropriate for the current cell. The Channel Quality Parameters message (Figure 5.27)provides the necessary guidance to the mobile devices. The data broadcast includes the following:
¥ RSSI Scan Time
¥ RSSI Scan Delta
¥ RSSI Hysteresis
¥ RSSI Average Time
¥ BLER Threshold
¥ BLER Average Time
First and foremost, the mobile must decide on a delicate balance between frequent channel assessment and low overhead. However, without general knowledge about the cell topography and size, it is difficult for a M-ES to define effective scan triggers. Typically an M-ES would scan for alternate channels if the current channel's signal deteriorates significantly. However, it was found that there are many cell layouts where the channel's signal does not drop significantly even when a mobile has moved well into the coverage area of an adjacent cell. To handle this situation, a mobile must assess the adjacent channels periodically regardless of the current channel's signal history. What the network operator can provide with the Channel Quality Parameters RSSI Scan Time and RSSI Scan Delta to direct the mobiles to use values appropriate for each cell. For example, a cell that is small relative to the normal movement speed could have scan more frequently - a smaller RSSI Scan Time. A cell that suffers from excessive RF shadowing effects should ignore sudden fluctuations in signal strength - a larger RSSI Scan Delta.
The RSSI Hysteresis value is the amount of signal improvement that a M-ES must experience before it moves to the alternate channel. This parameter is useful for network operators to alleviate excessive cell transfers at some cell boundaries due to topography.
The RSSI Average Time parameter inform the mobiles on what length of time to average the signal strength readings to achieve reliable measurements.
The remaining parameters direct the mobiles to search for an alternate RF channel when the current signal, though strong, suffers from other channel degradation effects such as interference. These effects appear as channel impairments that increase the block error rates. The BLER Threshold sets the percentage block error rate that should be considered appropriate for operation within the current cell. The BLER Average Time instruct the mobile device to take the average of block error performance over the specified period of time.
The last message broadcast by the MDBS is the Channel Access Parameters message shown in Figure 5.28. This message contains the important parameters associated with the Medium Access Control function. Adjustment of these parameters for each cell may be necessary to account for cell size and traffic profile and loading.
Once the mobile device has accumulated all this data, it observes the received signal strength of the current signal. If the received signal strength changes by more than the RSSI Scan Delta value, it initiates scanning of all the neighbor cells. If the received signal strength has not changed by more than the RSSI Scan Delta for a period of time equal to RSSI Scan Time, it also initiates scanning of all the neighbor cells.
For each adjacent cell identified by the Cell Configuration messages received, the mobile assesses the signal strength of the reference channel. The comparison between current cell signal strength and adjacent cell signal strength is made. If the adjacent cell is considered to be a better cell, the mobile must initiate cell transfer procedures. If the current cell is considered to be better than all neighbor cells, the mobile must stay within the current cell.
This is a simple and logical algorithm to provide local mobility management.