Even though the current state-of-the-art standardization activities of the 3rd Generation Partnership Project (3GPP) long-term evolution (LTE) and worldwide interoperability for microwave access (WiMAX) have allowed for very high spectral efficiency transmissions, the laws of physics, combined with the Shannon’s capacity bound, show that high spectral efficiency would only be available when the distance between the transmitter and the receiver is small. Taking the approach of scaling the cellular radio access network (RAN) architecture to decrease the distance between the users and the base station (BS) is not practical from the cost perspective. Ubiquitous very high data rate coverage is also an extremely challenging problem with the conventional radio resource management (RRM) approaches, as the rates decrease substantially at the periphery of BS coverage regions (the well-known cell-edge coverage problem). The conventional cellular design also uses fixed (a priori) radio resource allocations and assignments, which are inefficient; this inefficiency becomes even worse in a dense network due to the increased interference. It is, therefore, necessary to examine new RAN architectures, which can cost effectively increase radio port density in the RAN coverage area, and related RRM optimization techniques, which effectively manage the interference.
In classical cellular RAN architecture, there is essentially one network element—the base station, so for this reason increasing the density of radio ports by increasing the number of base stations is not practical. A consensus is currently forming in the community about the next generation of advanced RAN architecture, which contains many other network elements, such as distributed antenna elements, femto BSs, and relays. Indeed relays are already part of the WiMAX 4G standard and are considered for addition into the LTE-advanced 4G standard. The new elements are to provide a high density of radio ports to (1) decrease the distance to the receivers and (2) enable new coordinated multipoint (CoMP) transmission and reception techniques, which promise high data rates.
In advanced RAN, elements other than the base station either do not implement all of the functionality of the base station, or are not directly connected to the Internet. The elements in the new RAN work together to provide dense radio port coverage as shown in the illustration below;
The radio ports are attached to the various elements sed throughout the RAN: full BSs, femto BSs, and relays. A full BS is a gateway to he Internet for multiple RAN elements, whereas relays connect to the full-BS station with wireless connections. Femto BSs connect to the RAN through the Internet and provide indoor coverage. he base station (RAN anchor) is an important element of the advanced RAN. t manages multiple radio ports and has a wired connection to the Internet. RAN anchors o not require radio resources to provide backhaul services. We distinguish two types of RAN anchors: full base station (full BS) and femto base station (femto BS). A full BS is a gateway to the Internet for multiple RAN elements, whereas a femto BS is a gateway to the Internet for indoor elements. In addition to the various types of base stations, RAN also uses many types of relays. Unlike base stations, which are directly connected to the Internet, relays connect to the Internet through direct wireless connections to RAN anchors or through multihop wireless connections over other relays, which connect directly to RAN anchors. A relay may have multiple radio ports attached to it, as a base station would and may have to participate in hand-off and other RRM procedures, as a base station would. However, a relay is not connected to the RAN with a wired connection—it must at least connect to a base station to get to the Internet and it may connect to the base station in the network layer by using multihop transmissions through other relays. The advanced RAN contains various types of relays, which vary in complexity. For example, a relay may be a fairly simple amplify-and-forward relay, which does not examine the data flow, or a much more complex decode-and-forward relay, which examines and forwards packets. Typically, a relay is also expected to have a shorter range than a base station, so it requires a lower power amplification and thus cheaper power amplifier than a base station. Cheaper power-amplifier circuitry also makes relays cheaper than a base station, from an engineering point of view. The essential part of the proposed RAN are radio ports, which perform the radio transmission. The radio ports are available densely throughout the RAN coverage area, so that the distance between the terminal and a radio port is always small. Because radio ports are deployed densely throughout the RAN coverage area, it may be possible for the user terminals to simultaneously send (and receive) radio signals to (and from) multiple radio ports. Similar technologies are already proposed for LTE-advanced is OFDMA macrodiversity, also known as coordinate multipoint transmission (CoMP). However, CoMP has to be back compatible to the existing LTE standard, so it may not be able to take full advantage of the high port density. In advanced RAN, the port density will have to be much higher to achieve rates in the range of tens of gigabytes per second, so new precoding technologies will be necessary. With multiple simultaneous transmission, the terminal can take advantage of spatial diversity if the transmissions and receptions through the multiple points is coordinated. Coordination of transmissions and receptions leads to potentially higher rates, as precoding may be done to take spatial properties channel of the distributed channel. The concept of CoMP transmission and reception has many names, such as distributed antenna ports”, and in the standardization process “multicell IMO,” “network MIMO,” and “network cooperative MIMO”, to mention a few. The proposed RAN is a mesh of RAN elements, where any one element can connect to any other element. Due to the flat hierarchy in the RAN, RRM does not belong to any given RAN element. In the proposed RAN, RRM is a network-wide set f protocols and algorithms that allow network elements with different capabilities o negotiate assignments of radio resources to users.
