Innovations in Radio Access

As mentioned in the previous section, a single mobile transmitter and large service areas impeded the progress of the mobile telephone service. To address this challenge, AT&T Bell and other telecommunication providers introduced the concept of using multiple antenna towers within a geographical service area.⁴ Each antenna tower provides transmit and receive functions for a smaller coverage area, dubbed a “cell.” The antenna tower, known in 1G as a base station (BS), is at the heart of each of these cells, and multiple BSs could be placed strategically to form a cellular network throughout the desired service area. Figure 1-3 illustrates this concept.

Figure 1-3 shows multiple transmission stations, each covering a smaller area (or “cell”), which provided a simple, scalable, and extensible solution to the limitations of MTS and IMTS. Now mobile (or rather cellular) service subscribers were more likely to be in close proximity of a base station (BS) throughout the coverage zone. Using this approach, more “cells” could be added to the network, thus easily expanding the service coverage area if required.

While this cellular approach addressed some of the problems in the original MTS-based mobile design, it introduced the problem of radio wave interference between neighboring cells. The pictorial representations of a cellular network give an illusion of clean radio coverage boundaries between cells, but the reality is not so. Radio waves transmitted from base stations travel through the cell coverage area, but they do not magically stop at theoretical cell boundaries. The cell boundaries are rather amorphous, where cellular signal from adjacent base stations overlap each other (as shown by circular shapes in Figure 1-3). This results in signal interference, thus distorting the signal and sometimes making it harder for the user handset to extract meaningful information from it.

Adjusting the transmission power could help minimize this interference but does not eliminate it. The problem was solved by using non-overlapping frequency ranges in neighboring cells. The solution encompassed dividing the available frequency spectrum into smaller ranges, and then using one of these subdivided frequencies in each cell. The same subdivided frequency spectrums can be reused in multiple cells, provided the cells have sufficient geographical separation among them to avoid service-impacting interference from neighboring cells. Figure 1-4 shows the various frequency reuse patterns, where each unique frequency range is represented by a number. Clusters of 4, 7, or 12 frequencies are commonly used frequency reuse patterns. These patterns can be repeated over a larger geography, thus allowing for expanded coverage area using the same frequencies.

It must be noted that Code Division Multiple Access (CDMA), an air-interface access method, uses a different principle. CDMA uses special codes that allow reuse of the same frequency across all cells. CDMA is further explained later in the section, “Third Generation (3G).”

An Introduction to Mobile Transport

While the cellular network concept made it easier for the user equipment to communicate with geographically disperse base stations, it also highlighted the need for a robust mobile transport. The system now required a transport mechanism to connect these distributed cellular base stations to the central exchange. In the case of early 1G networks, all base stations within the coverage area were directly connected to the central exchange, typically using analog leased lines. Virtually all the 1G network systems were developed independently and used proprietary protocols over these leased lines for communication between the base station and MSC. This could be considered the very first mobile transport network, and although the underlying protocols and communication mechanisms have evolved significantly over the years, the fundamental concept of a mobile transport originated from these very first 1G networks.

Emergence of a Mobile Core

The base stations used a point-to-point connection to the exchange or central office within the coverage area. This central office, referred to as the Mobile Switching Center (MSC), provided connectivity services to cellular subscribers within a single market. The MSC performed its functions in conjunction with other subsystems located in the central office, including the following:

• Home Location Register (HLR): A database that contained the information of all mobile users for the mobile operator.

• Visitor Location Register (VLR): A database that temporarily stored the subscriber profile of mobile users currently within the coverage area serviced by the MSC.

• Authentication Center (AuC): This subsystem provided security by authenticating mobile users and authorizing the use of mobile services.

The MSC was responsible for all functions in the 1G cellular network, including the following:

• User authentication and authorization: Done through the AuC subsystem using the HLR.

• Security and fraud prevention: Done by comparing locally stored phone data in the AuC’s Equipment Identity Register (EIR) with equipment information received. This helped the MSC deny services to cloned or stolen phone units.

• Cellular subscriber tracking and mapping to base station within its coverage area: Each base station would provide the MSC with this information for subscribers associated with that base station.

• Voice call connectivity, including local and long-distance calls: Any calls between cellular subscribers within the coverage area were connected directly through the MSC, while call requests to cellular subscribers in other coverage areas or to traditional PSTN subscribers were routed through the PSTN network. Using PSTN as the backbone ensured universal connectivity between all cellular and traditional land-line subscribers.

• Subscriber handoff between different base stations: This was assisted by the MSC, which constantly kept track of subscriber signal strength received through base station(s). When the MSC determined that the subscriber signal was stronger from a base station different from the one that subscriber was currently registered to, it switched the subscriber to the new base station. In cellular terminology, this process is known as a handoff. In 1G networks, cellular handoff was initiated by the MSC, as shown in Figure 1-5.

• Roaming between different MSCs: This refers to both roaming within the mobile provider coverage zone (intra-operator roaming) and roaming between different mobile providers (inter-operator roaming). Initially, inter-operator roaming and registration was a manual process, but it was subsequently replaced by automatic registration.

• Billing: Billing was also managed by the MSC as it kept track of all subscribers, their airtime usage, and call type (such as local or long distance).

Figure 1-5 shows a local cellular handoff within an MSC service region as well as a subscriber roaming between different MSC regions, including both inter-operator and intra-operator roaming.

1G cellular networks provided a robust framework for scalable mobile telephony services. Cellular radio access network, mobile transport from base station to MSC, and modular functional blocks within the MSC provided an architecture that laid the foundation for subsequent generations to build upon. Because the MSC was handling all the major functions of managing and monitoring of user devices as well as call control functions, it limited the overall scalability of the mobile system. Second generation (2G) mobile architecture aimed to address these limitations.

2G Innovations in Radio Access

2G introduced a number of technical enhancements over its 1G counterpart, primarily focused on providing ease of use and scaled services through digital modulation, multi-access, enhanced security, and handset flexibility.

2G Mobile Transport

In order to provide a more efficient and scalable network, 2G introduced the concept of a base station controller (BSC). Now, instead of a direct point-to-point connection from each base transceiver station (BTS) to MSC, multiple BTSs would connect to a BSC that provides connectivity to MSC. The BTS in 2G was the equivalent of a base station (BS) in 1G. Figure 1-6 provides an overview of the end-to-end 2G mobile network.

As shown in Figure 1-6, multiple BSCs acted as an aggregation point for a group of BTSs. The physical link between the BTS and BSC was a full or fractional T1 or E1 link, with specialized protocols for communication between the two. The control functions of the BTS, such as frequency channel allocation, user signal level measurement, and cellular handoff between BTSs (previously the responsibility of MSC), were now handled by the BSC.

Modular transport provided extensibility to add a new BSC and/or BTS when desired. As the BSC acted as an aggregator for multiple BTSs, the architecture allowed the MSC to scale better by controlling more BTSs using the same number of links. Additionally, depending on the vendor, the BSC also provided switching center capabilities, thus further reducing the load at the MSC. BSCs were connected to the MSC using Frame Relay over full or fractional T1/E1 links.

2G Mobile Core

With the introduction of the BSC, some 2G networks, such as GSM, started to distinguish between the radio network and the switching network. The terms base station subsystem (BSS) and network switching subsystem (NSS) were introduced to highlight this architecture. The BSC and BTS functions belonged to the BSS, while MSC and the databases it used collectively comprised the NSS, performing all validation and switching functions. This distinction led to how the mobile networks are architected today; NSS evolved into the mobile core, and BSS evolved into radio access network (RAN). The transport network providing the connectivity between the RAN and mobile core evolved into what is commonly known as the mobile backhaul.

The 2G mobile core offered a number of key enhancements over the previous generation in terms of efficiency, interoperability, and services, some of which are covered here.

2G Technology Summary

2G development was an interesting time in mobile standardization and adoption, as multiple technologies were competing in standard bodies as well as in the marketplace. GSM quickly emerged as the dominant market force owing to its widespread adoption in the European market. GSM emerged as the major 2G mobile standard; its users could get “roaming service” in multiple countries and could easily switch handsets using SIM cards. All these “features” propelled GSM to frontrunner status in the race for mobile dominance.

At the turn of the century, GSM had more than two-thirds of the market share compared to rest of the mobile technologies, with over 788 million subscribers in 169 different countries.⁶

3G Innovations in Radio Access

The frequency resource was still a significant constraint in meeting the growing bandwidth demand. In addition to adding more frequency bands, 3G also offered more efficient air interface sharing among a growing number of subscribers and services.

Improvements in digital signal processing made it possible to consider more complex media access technologies than Frequency Division Multiple Access (FDMA) and Time Division Multiple Access (TDMA). This resulted in the acceptance of Wideband Code Division Multiple Access (WCDMA) as a standard for 3G. WCDMA is a flavor of CDMA that was used in some mobile networks of second generation, such a IS-95. WCDMA and CDMA are both multi-access technologies based on the same principle of separating user transmissions using codes. Compared to CDMA, WCDMA uses higher code rates, wider channel widths of 5MHz instead of 1.25MHz, and some other differences.

Unlike previous generation radio access, WCDMA does not assign a separate frequency or a timeslot to different users. Instead, their signals are transmitted at the same time and using the same frequency. This media access technology is based on a counterintuitive approach of transmitting a signal using substantially more bandwidth than would be necessary with other modulation techniques. Each bit or group of bits (a symbol) of an original digital signal is encoded by a few rectangular pulses (called “chips”), based on a special sequence of bits, called a “spreading code.” When an original digital signal is multiplied by such a code, it could be said that the signal is being spread over the spectrum, as a higher rate signal of rectangular pulses occupies a wider frequency spectrum compared to a similar but lower rate signal. The spectrum of an original signal becomes wider, and the energy of the original symbols is also distributed over the wider frequency band. Figure 1-9 illustrates the use of spreading codes over a digital signal in WCDMA.

The reverse operation, or de-spreading, requires exactly the same code to recover the original signal. The energy of each original symbol is combined during de-spreading operations and results in the recovery of the symbol. This process is also referred to as “correlation.”

Spreading codes are generated in such a way that they maintain a mathematical property of orthogonality. When a de-spreading operation is done with a code different from the one used for spreading, it results in recovery of nearly zero energy for each symbol that appears as negligible noise. Different spreading codes are used to encode different data channels, making it possible to distinguish them at the receiver. Therefore, these are also referred to as “channelization codes.”

When both transmitter and receiver use the same code, the de-spreading operation recovers the original signal while effectively filtering out any other signals. This allows simultaneous transmission of multiple signals encoded with different orthogonal spreading codes using the same frequency band. Figure 1-10 illustrates the signal transmission and de-spreading operation in WCDMA.

The spreading codes used in WCDMA have a fixed rate of 3.84M chips per second, which is related to the width of a radio channel (5MHz). The number of chips used to encode a single bit (or more strictly, a symbol) of the original signal is known as “spreading factor” and may vary in WCDMA transmissions between 4 and 512.

When spreading codes don’t start at the same time, they might have decreased orthogonality. This becomes a challenge in WCDMA due to asynchronous mode of base stations operation. To cope with that, WCDMA uses an additional code, called a “scrambling code.” An already spread signal is further encoded by a scrambling code without increasing the rate of the signal. This operation ensures better orthogonality and helps to better distinguish transmissions from different mobile devices and base stations. Different scrambling codes are used by each individual transmitter in a cell.

WCDMA transmissions use Quadrature Phase-Shift Keying (QPSK) modulation in the downlinks and dual-channel QPSK in the uplinks. Without going into the complexity of modulation theory, suffice to say that QPSK modulation converts 2-bit symbols into a radio signal with four distinguishable phase shifts. Thus, each 2 bits of a downlink channel in UMTS are modulated by a single-phase shift of QPSK. In the uplink direction, however, 2 bits comprising a single symbol for QPSK are provided by two separate channels: control and data, hence the name dual-channel QPSK. Multiplexing of two uplink channels into a single transmission improves air interface efficiency when significant asymmetry of the traffic patterns exists between uplink and downlink.

The very foundational principle of WCDMA, that all mobile devices within the cell transmit on the same radio frequency, can lead to a situation where a single powerful transmission jams all other (weaker) signals. This is also known as “near-far problem” in CDMA systems. To address this problem, it is critical to control power of each mobile device transmitting in the cell. Each radio node constantly evaluates radio signal parameters, such as signal-to-interference ratio (SIR), and instructs each device to reduce or increase its power level.

In addition to these, there were many other innovations such as Adaptive Multi-Rate (AMR) codecs for voice, switching off transmissions during voice silence or gaps in data streams with discontinuous transmissions (DTX), and so on, but their details fall outside the scope of this book. Collectively, these innovations helped to create a robust, efficient, and fast UMTS air interface, reaching the peak speeds of 2Mbps in the downlink and 768Kbps in the uplink; however, later 3GPP releases boosted achievable data rates significantly.

3G Mobile Transport

3G UMTS introduced the concept of NodeB, which terminates the air interface from mobile devices and is considered the demarcation point between radio access networks and the mobile backhaul network. Similar to its predecessor (BTS) in 2G, NodeB required connectivity to its controller in the mobile core. In 3G, the controller is called the Radio Network Controller (RNC), and NodeB relies on the mobile transport to provide robust connectivity between the two.

Initial 3G implementations used a similar approach for mobile transport as their predecessor (that is, point-to-point T1/E1 links between the NodeB and RNC). As bandwidth consumption continued to grow, however, the 1.5/2Mbps capacity offered by T1/E1 links quickly became saturated. The use of 5MHz channels, along with the efficient coding techniques, significantly increased the bandwidth requirements in the mobile backhaul (that is, from the cell site to the mobile core). One of the solutions was to use multiple T1/E1 links; another option was an upgrade to higher capacity T3/E3 links. This was a costly proposition, however, given the dedicated use of such point-to-point links, and the industry moved toward higher capacity yet cost-effective alternates such as Asynchronous Transfer Mode (ATM) and IP in the mobile backhaul.

Around the same time when the first 3G networks were being deployed, Internet providers and network operators were also deploying high-speed data networks to meet the growing Internet bandwidth requirements. One of the leading technologies for such networks was ATM, which provided higher speeds (155Mbps or more), traffic prioritization through quality of service (QoS), and predictable traffic delay. These properties made ATM a suitable transport mechanism for mobile traffic, where voice required careful handling due to its time-sensitive nature and data required higher bandwidth.

With 3GPP standardizing the use of ATM, and subsequently Internet Protocol (IP), for mobile transport, this presented service providers with an opportunity to consolidate their Internet and mobile transport networks. The potential to use a single transport network tempted the service providers and network operators to embrace a common technology for mobile and Internet transport in the hopes of optimizing operation, reducing operational expenses, and extracting maximum return on investment on their deployments. For mobile communication networks (MCNs), this meant that instead of using purpose-built, mostly point-to-point links, base stations and NodeBs could utilize general-purpose high-speed data networks to connect to the mobile core. While this concept of a “single converged network” was introduced in 3G, it did not see significant adoption in most service providers until well into the 2010s. Some service providers preferred to maintain separate physical networks for mobile and traditional data networks, while many others made substantial strides towards consolidating the two. 3G heralded the arrival of the mobile backhaul (MBH) era for MCNs. Whereas, previously, transport was simply a collection of dedicated point-to-point links from a base station to the mobile core, MBH networks provided a blueprint for robust, reliable, multiservice connectivity within and between the radio access network and the mobile core. MBH networks are discussed in more detail in the next chapter.

3G Mobile Core

As previously mentioned, the 3GPP definition of the core network in Release 99 did not feature many changes to the GSM/GPRS standards. Although new interfaces were defined to interact with UTRAN, the main constituents of the core network remained largely unchanged: MSC, GMSC, HLR, SGSN, GGSN, and so on. However, scalability challenges in the circuit switched domain became a reality due to mobile networks’ expansion and consolidation over large geographic areas. Major core network changes were therefore introduced in 3GPP Release 4 to address these challenges by splitting the functions of MSC into two entities: MSC server (MSC-S) and media gateway (MGW). Figure 1-11 shows an overview of the 3G/UMTS architecture and components.

The MSC-S is sometimes also referred to as a “call server” and implements call control and signaling. Put simply, MSC-S takes care of call routing decisions and negotiating voice bearer requirements when a mobile subscriber makes or receives a circuit switched voice call. MSC-S is also responsible for HLR interrogation, maintaining VLR, and the generation of call detail records for the billing system. An MSC-S connecting its mobile networks with other networks is called Gateway MSC-S (GMSC-S).

While MSC-S and GMSC-S implement call control and signaling, the actual circuit switching is performed by media gateways (MGWs). In other words, MGWs provide bearers for circuit switched voice, perform media conversion between TDM and ATM or IP-based voice, transcoding, and other call-related functions such as echo cancellation.

Both ATM and IP were defined by 3GPP as acceptable transport for MGW and MSC connectivity. The signaling between MSC-S, which was originally done using a protocol called Signaling System 7 (SS7), can now be carried over ATM or IP. When carried over IP, it is referred to as SIGTRAN, short for signaling transport.

Although some significant changes were introduced in the circuit switched part of 3G CN, the packet switched domain remained largely unchanged. The functional split between SGSN and GGSN allowed for scaling the packet switched domain efficiently by adding more SGSNs and/or GGSNs where and when needed. SGSNs and GGSNs were retrofitted with new interfaces to communicate with RNCs in the UTRAN over ATM or IP.

The terms circuit switched core and packet switched core also emerged to distinguish between two major functional domains of 3G CN, with the latter further simplified to just packet core. Even though the circuit switched core could now use an ATM or IP backbone for connectivity, interestingly enough it remained separate from the packet-switched core’s IP backbone in most implementations.

3G Enhancements

Though 3GPP had originally defined its Release 99 for the third generation mobile wireless specifications, the consortium continued to define new specifications in the subsequent years. Unlike Release 99, where the number “99” was influenced by the year (1999) when most of it was developed, the subsequent releases were simply given a sequential name, starting with Release 4.

These released specifications had varying degrees of influence on the RAN, transport, and mobile core of 3G mobile systems. These influences will be discussed in this section.

3GPP Release 8 and Beyond

After 3GPP Release 4, which had brought significant changes to the mobile core, all the releases were more focused on improving the air interface data rates. However, Release 8 once again specified major changes to the mobile core using the new specifications called Evolved Packet Core (EPC). On the radio side, it specified Enhanced Universal Terrestrial Radio Access Network (E-UTRAN).

These enhancements, known as Long-Term Evolution (LTE), paved the path for 4G networks, which are covered in Chapter 3.

Figure 1-12 shows a progression of uplink and downlink speeds across various 3GPP releases.

3G Technology Summary

2G passed the baton to the third generation of mobile networks at a critical turning point in mobile networking history.

3G enhanced mobile services on two major fronts. Early 3GPP releases focused on major changes in the radio access technologies with the use of WCDMA, establishing a solid foundation for air interface evolution. Through the separation of MSC functions into MSC-S and MGW, 3G kickstarted the adoption of ATM/IP transport in the mobile core and eventually in the mobile transport. The transition from TDM to ATM/IP promised operational expense reduction and maximization of the return on investment on future deployments.

Later 3GPP releases introducing HSPA provided tangible benefits to end users through higher upload and download speeds. The result was a massive growth in the subscriber base. By the end of 2008, the number of mobile subscriptions surpassed 4 billion globally.⁹