High QOS with Traffic Control for Effective Bandwidth Management | Solutions and Recommendations

Overloading of a wireless network could occur because of the heavy data traffic. Traffic flow control for effective management of transmission resources, particularly the bandwidth, facilitates high quality of service (QOS). For example, by shaping (that is, spacing) the data traffic which usually comes in bursts unlike the voice traffic, the network bandwidth could be utilized more effectively. On the other hand, overloading of the wireless network even for short durations of time could result in degradation of QOS due to increased bit error rates.

Figure 1 depicts the different feedback controls used in our system for traffic flow adjustment for effective data transmission. Signal power of the mobile device or any other direct indicator of power for the RF link constitutes the inner power control just as in the traditional wireless systems. This feedback is provided every 1 -2 milliseconds. Similarly, outer power comprises a link error rate and/or interference indicator for the wireless link and may account for soft handoff power. This feedback may be provided every 50-100 milliseconds. The inner and outer power control loops conjointly provide feedback based on the signal strength of the RF link. The packet level control is provided every several hundred milliseconds by the queuing system in the WR based on the congestion status of the queues. Finally, our repertoire of traffic control mechanisms included the well known and well studied TCP flow control, which is provided through the acknowledgment messages between the two end points of the TCP flow. Based on the acknowledgment messages received, the source (WR) adjusts its transmission rate. Thus, with this mechanism, traffic flow is controlled by the congestion and/or interference state of the wireless links.

Figure 1: Feedback mechanism for traffic flow control

 

In our WR, there is also a provision for the traditional end-to-end rate control with a queue mechanism to shape up bursty traffic from a source into a smooth traffic flow of radio frames into the sink (mobile). The ACKs from the sink are also similarly queued up, and used as feedback for the source so that it can control its egress traffic flow.

An innovative flow control mechanism in the present work is Gang (or Group) flow control which seeks to shape the TCP flows from various sectors of the wireless network simultaneously, shown as Figure 2 with N acknowledge shapers corresponding to N sectors of the WR. Each shaper accepts the packets from wireless network and stores the packets in the acknowledge queues inside the WR. The acknowledge shaper can transmit acknowledge message over time to change the traffic flow for the sector based on the flow's power indicator of the RF link from wireless network.

Figure 2: Gang flow control for sectors of a wireless

 

Figure 2 illustrates two kinds of shaped acknowledge messages. In the first type shown for flows 1 and 2, the acknowledge messages are arranged in small groups and the groups are dispatched periodically. In the second type depicted for Flow N, on the other hand, the ACKS are evenly distributed overtime and transmitted. The difference of these two arrangements is that they have different transmit time. The first type of ACK shaping will affect the offsetting bursts for the traffic flow whereas the second type will affect the steady flow rate for the traffic flow.

The flows with unused bandwidth or lack of bandwidth will be identified for each TCP group for the interval related to retransmission time out for the TCP flow. The fair share of each TCP flow within the group will be calculated and used in adjusting the speed of acknowledge messages for the flows inside the gang. The fair share of each TCP flow may be used together with the RTT and arrival time for each traffic flow.

Call Processing and Soft Handoff Mechanism | Solutions and Recommendations

Figure 1 depicts interactions among various network elements for ingress data (i.e. data flowing from the mobile device to the edge router. Figure 2 depicts the interactions for the egress data flowing in the reverse direction. These figures are OO (object-oriented) style sequence diagrams for call flow in an UML (Universal Modeling Language) like notation.

Figure 1: Data transfer from the mobile device to an edge router

 

Figure 2: Soft hand off and data transfer from an edge router to the mobile

 

In both the scenarios described here, the first step is designation of one of the routers accessible to the mobile device as the primary router. In case of mobile originated calls, the mobile informs the WR from whom the strongest signal is received to take the responsibility as prime WR. On the other hand, in case of mobile terminated calls, the edge router determines the prime router based on the mobile location after locating and successfully paging the mobile. Once the primary router is determined, it initiates the process for setting MPLS tunnels between itself and the edge router as well as the secondary routers. The key innovation here is to employ these MPLS tunnels to emulate the BS-BSC and BSC-MSC (mobile switching center) A3/A7 interfaces in the legacy wireless systems. Distribution of call flow control this way among various routers this way results in enormous cost savings for the customers due to effective utilization network links by reduction of the call control and data paths. Figures 1 and 2 depict the call flows after creation of MPLS tunnels.

In case of ingress data, the radio frames transmitted by the device are received by all active routers including a primary router and a number of secondary routers as shown in Figure 1. The secondary routers simply forward the radio frames to the primary router. The SDU of the primary WR selects the best one among all such frames including the one directly received, inserts that into an IP packet, and transmits it to a back-haul network via an edge router for onward transmission to the other party.

The flow in case of egress data is naturally in the opposite direction as shown in Figure 2. The core network hands over the IP packets to an edge router for onward transmission to the primary router through a pre-established or dynamic MPLS path. The primary WR segments the packets into radio frames and multicasts them to all the secondary WRs in the active set via dynamically configured Label Switch Paths (LSPs). The primary and secondary WRs then transmit each one of these received radio frames after different amounts of delay offset to the mobile device so that the replicas of individual frames from different WRs arrive simultaneously at the destination. These synchronous radio frames received from different WRs are analyzed by the mobile device to not only obtain the best (correct) radio frame, but also assess the power levels of the frames. As the mobile device moves away from the primary WR, the power level of the radio frames from the mobile at the primary as well as that of the frames from the primary at the mobile drop. When the power level drops below a pre-configured threshold, the mobile device sends a control message to the primary WR indicating the power level of the prospective primary. The new primary WR could be one of the previous secondary wireless routers in an active set for the call. To achieve micro mobility, the current primary WR would signal the new primary an indication of the handover of its responsibility as primary WR. It would also supply to the latter the list of active WRs in the same control message. After receiving this message, the new primary WR receiving the strongest signals first confirms to the current primary that it is ready to take control of the traffic distribution, and then establishes multicast MPLS paths (LSPs) for the secondary routers in the active list. The LSPs provide synchronized framing for distribution and selection between neighbors of wireless traffic and fast rerouting for soft handoff using RSVP.

Solutions and Recommendations | WIRELESS ROUTER ARCHITECTURE

In the proposed architecture, the wireless routers (WRs) situated in different cells perform message routing in addition to the traditional Base Station (BS) functions. In fact, the traditional functions of the base stations, BSC (Base Station Controller), and MSC (Mobile Switching Center) in the legacy systems are lumped together and distributed across various WRs in the new wireless architecture. Figure 1 is a very high level depiction of various inter-router links as well as the control links between the proposed WRs, and the traffic and control interfaces. The inter-router links include: i) wireless specific virtual tunnels i. e. Multi Protocol Label Switch (MPLS) paths based on IP packets, ii) RSVP (Resources Reservation Protocol)/LDP (label Distribution Protocol) signaling channel, iii) routing message channel, iv) wireless specific virtual channel to carry MPLS based radio frames, and v) one or more wireless-specific control channels for call setup and maintenance including a signaling channel for usage by any signaling protocol (SP) such as extended RSVP and a routing channel for usage by any radio routing protocol (RRP) such as extended OSPF (Open Shortest Path First), RIP (Routing Information Protocol), or BGP (Boarder Gateway Protocol). The bold and dashed lines in the figure indicate the data and control channels, respectively.

Figure 1: Proposed wireless router links to peer routers and various control interfaces

 

For supporting wireless traffic services, the WRs access traffic and control interfaces that include media gateway controllers, WAP (Wireless Application Protocol) servers, policy management servers, call agent controllers, mobility managers, and AAA (Authentication, Authorization, and Accounting) servers. WRs communicate with these interfaces through MGCP (Media Gateway Controller Protocol), COPS (Common Open Policy Service), and other suitable protocols.

The WR architecture proposed herein facilitates wireless-access technology (e.g. CDMA or TDMA) independent network routing with the help of wireless interfaces to disparate wireless peripherals as shown in Figure 2, and hence is pivotal to an all-IP (Internet Protocol) radio access network (RAN) that seamlessly inter-works with the backhaul IP network with interfaces to various network peripherals. For communication with the backhaul networks at the other end, the WR includes wire-line interfaces to various network peripherals as well. At the heart of the WR lies the traffic control with various modules as follows: i) Quality of Service (QOS) engine for traffic conditioning and effective management of transmission resources, ii) Selection and Distribution Unit (SDU), iii) Central Processing Unit (CPU), iv) Call Processor, v) Timing Unit for synchronization purpose, vi) Communication Module with various traffic-controller interfaces to gateways, services, policy managers, IP routers, base-stations, call agents, and other remote nodes and resources, vii) Power and Interference Manager, viii) Radio resources Manager, ix) Mobility Manger, X) Packet Classification Unit, and xi) Security (IP SEC) Module.

Figure 2: Internal structure of the proposed wireless router

 

In the traditional mobile networks, an SDU is placed centrally at the Base Station Controller (BSC) to manage the call processing at a number of base stations. It selects the best frame from a number of incoming radio frame instances from the mobile via different base-stations (BSs) for onward transmission to the intended destination through backhaul network, and distributes similarly the messages received from the backhaul to the target mobiles. Based on the quality of the frames received from different base-stations, an SDU also manages soft-handovers (that is, call redirections from one BS to the other). In our architecture, BS is replaced by the more versatile WR module. Additionally, it incorporates an innovative concept of a distributed SDU with every WR housing an SDU. These distributed SDUs in the present architecture facilitate distribution of intelligence, and switching and control functionality to individual cell sites.

The main advantage here is that the radio frames need not be transmitted to the BSC over the backhaul links. A very efficacious use of the backhaul network bandwidth this way, in turn, results tremendous cost savings for the customer. Additionally, this approach helps in averting traffic congestion. Common switching points leading to delayed/dropped traffic are reduced if not altogether eliminated. However, these advantages could be realized only by effective inter-router communication methods for mobility management by call redirection (soft handoff) and MPLS path reconfiguration as the mobile device transitions between cells. The proposed innovation should also be implemented without compromising the quality of service (QOS). In the following subsections, we describe the mechanisms for soft handoff and QOS in the proposed architecture.

Issues, Controversies, Problems | WIRELESS ROUTER ARCHITECTURE

The legacy mobile wireless architecture shown on the right half of Figure 1 with a hierarchy of BS (Base Station)s, BSC (Base Station Controller)s, and MSC (Mobile Switching Center)s has a number of shortcomings with respect to handling of large volumes of data in B3G networks: i) wireless frame selection for handoff management is done at BSCs resulting in the duplicate traffic flow on the backhaul, ii) even during the ideal periods of a call, transmission resources are reserved resulting resources wastage in contrast to the IP-networks equipped with the statistical multiplexing scheme, iii) only 15% of the BS-BSC traffic is payload, and the rest is overhead, iv) BSs forward erroneous frames also BSC and this results in dead payload on the backhaul, v) uneven utilization of links makes the system inefficient, cost-ineffective, and unsuitable for deployment of new data intensive services, vi) transmission delays in long BS-BSC links could cause soft-handoff failures and call drops and thereby contribute to performance degradation, and vii) single-point failures of the legacy system entities or the links between them results in low system availability.

Figure 1: A novel distributed wireless router based RAN communicating with a legacy network (each pink dot represents a wireless router)

The demand for high quality services despite large volumes of call traffic necessitates a drastic reduction in the long backhaul (BS-MSC) control and data paths used in the legacy systems. An elegant approach to address the problem is to build the next generation RANs as distributed configurations of simple but functionally comprehensive wireless router (WR)s that could replace and at the same time inter-work, as shown in Figure 1, with the hierarchy of legacy wireless system entities i.e. the BSs, BSCs, and MSCs. Thus these new versatile WRs, need to have integrated into them several overlaying features of an all-IP 3G wireless network depicted in Figure 2. Even though the functional modules of the service layer are not detailed out in the figure in view of the ever growing number of wireless/wire-line services, typical entities that provide network services are call servers, bandwidth brokers, SLA (Service Level Agreement) managers, billing servers, HLR (Home Location Register)s, HSS (Home Subscriber Server)s, MGW (Media Gateway)s, SGW (Signaling Gateway)s, legacy servers, DNS (Domain Name Servers)s, and so on. The control layer supports the services with entities such as QOS (Quality of Service)/Mobility/Location/Power managers, Call Agents, and AAA (Authentication, Authorization, and Accounting) mangers. In the legacy systems, these two functional layers roughly correspond to the MSC and BSC functions, respectively. For execution of these functions, the legacy system entities (BSs, BSCs, and MSCs) need to communicate with one another. This is supported by backhaul IP networks with wire-line topologies. A part of the inter-BS communication is supported by wireless routing. Finally, the communication between the BSs at different cell sites and the mobile device constitutes the physical layer functionality of the wireless system.

Figure 2: Several overlaying features that need to be integrated into a WR

Since the WR replaces the mobile network entities in the proposed RAN architecture, it needs to incorporate the control and routing functionalities of the legacy systems. In particular, it should have the following features: i) it should support all the data and signaling protocols for inter-router communication as well as communication with various service and control entities, ii) it should facilitate effective hand-off management to achieve nearly zero call drop rate, iii) it should provide high QOS by effective bandwidth management through traffic shaping, and iv) it should be capable of dynamically configuring its operational parameters in collaboration with its neighbors, and adapt itself to RF topology and other changes. Overall, it should render the RAN both high performing and highly available at a low cost.

IP-TRIGGERED RESOURCE ALLOCATION STRATEGY (ITRAS)

To support IP QoS, the Internet Engineering Task Force (IETF) recommends integrated services (IntServ)  and differentiated services (DiffServ). They are expected to be effective also in all-IP-based 4G networks. Since 4G networks will support multimedia traffic, we need to visit the issue of providing IP QoS in IP based wireless access networks, and propose ITRAS for QoS support in 4G networks, where the decision of radio resource allocation follows IntServ or DiffServ policy.

IP QoS

IntServ  uses Resource Reservation Protocol (RSVP) to reserve bandwidth during the session setup. As a first step of RSVP, the source sends a QoS request message of PATH to the receiver through intermediate routers which run an admission and a policy control. If the sender receives RESV returned from the receiver through the reverse route as an indication of QoS guarantee, it initiates the session. If each router along the path receives packets, it classifies and schedules them. IntServ ensures strict QoS, but each router has to implement RSVP and maintain per-flow state, which brings difficulty in a large-scale network.

DiffServ, on the other hand, does not need any signaling protocol and cooperation among nodes. As the QoS level of a packet is indicated by the DS field of IP header (TOS field in IPv4, Traffic Class field in IPv6), each domain can deal with it independently. Once the packet is classified, each router can mark, shape or drop it according to network status. Since DiffServ is not so rigorous as IntServ, it is scalable in supporting QoS statistically.

QoS of Wireless Access Networks

In general, a wireless access network has the capability of managing QoS independently of the IP network because it becomes a bottleneck for providing end-to-end QoS. QoS control can be made possible by using some access and scheduling methods. Recently the QoS of IEEE 802.11 WLAN system is supplemented by IEEE 802.11e standard (IEEE 802.11e, 2005). It defines Extended Distributed Contention Access (EDCA) that assigns a small backoff number to high priority traffic, and Hybrid Coordination Function (HCF) that improves the conventional polling scheme of Point Coordination Function (PCF). Also, cdma2000 lx EV-DO and WCDMA-HSDPA (High Speed Downlink Packet Access) adopted opportunistic scheduling to exploit channel fluctuation. The opportunistic scheduling has brought an implementation issue in designing various scheduling algorithms for QoS.

The Third Generation Partnership Projects (3GPP and 3GPP2) define four traffic classes and their related parameters for QoS provisioning. There exist gateways between IP backbone and access networks that perform protocol conversion and QoS mapping between IP and access networks. However, direct translation is difficult since access networks have their own QoS attributes that require strict QoS provisioning within them.

Meanwhile, the importance of unified QoS management grows in 4G networks as QoS management for both access network and IP network becomes cumbersome in all-IP networks. If each network has an individual QoS model, it needs a rule that integrates their QoS models to ensure end-to-end QoS. For the unified QoS management, we propose ITRAS that considers L1, L2 and L3 together. In ITRAS, L1 and L2 allocate radio resources and logical channels, respectively, according to the QoS indication of L3.

C ITRAS

ITRAS concerns the information about IntServ and DiffServ for the resource management of L1 and L2. When IntServ sets up a real-time session, MAC reserves a dedicated channel. On the contrary, when DiffServ is used for low mobility users, MAC can exploit either a dedicated or shared channel. If the shared channel is allocated for DiffServ, the wireless scheduler runs a scheduling algorithm for QoS provisioning. In contrast, the dedicated channel allocation needs admission control that allows a limited number of users into the network for QoS support. Therefore IP QoS information helps MAC and PHY manage resources of the following in a flexible manner.

• Cell type - microcell or macrocell
• Multiple access - OFDMA or FHOFDMA
• MAC channel - dedicated or shared
• PHY scheduling - priority or fairness

IntServ is easy to be involved in radio resource management because wireless access is usually accompanied by signaling. When an MS requests a real-time service in a 4G network, the corresponding AR can initiate IntServ and allocate a dedicated channel. For a downlink call, the AR can adjust the bandwidth of a dedicated channel with the aid of RSVP As real-time traffic usually requires a constant data rate, the dedicated channel is recommended to use power control rather than AMC. In this aspect, FH-OFDMA and CDMA may have more suitability than OFDMA for real-time services.

Regarding DiffServ in 4G networks, it is enough for an MS to set the DS field properly for uplink packets because the BS controls radio resources before transferring them to the AR. For downlink traffic, the AR classifies packets according to the DS field and chooses a multiple access method, and accordingly the BS allocates a dedicated or shared channel. The dedicated channel has the advantage of simple management, while the shared channel goes well with the DiffServ because both require scheduling. Contrary to scheduling in routers which need to handle a lot of flows, wireless scheduling takes care of not many connections, which allows to use per-user buffer. So the wireless scheduler can exploit an algorithm with high granularity of radio resources. Figure 1 summarize tightly coupled resource management among three layers through a unified QoS strategy.

 

Figure 1: The coupled layering for resource management

Further Issues

Implementing ITRAS needs further study. Specifically, when the subnet-based all-IP network is deployed, an AR should cooperate with its subordinate BSs in performing ITRAS functions. While resource management functions are primarily handled by BSCs in the GSM networks, more functions will be imposed on BSs in 4G networks. Basically ARs will be responsible for IP QoS and BSs will play the primary role of resource management. Another challenge is the application of ITRAS to the macro/micro cellular network. In this case, a coordinator is needed in deciding whether an incoming session is served by a macrocell or a microcell. It will also have the capability of load balancing by triggering vertical handoff.

Along with the architectural evolution towards all-IP network, one of the most salient trends for future network design is emerging in the form of Fixed Mobile Convergence (FMC). The integration of wireline and wireless technologies and services realized by FMC is expected to offer benefits to both operators and consumers by delivering enhanced user experience over a unified framework. IP Multimedia Subsystem (IMS) (Poikselka, 2004) lies at the heart of this network convergence. It is a framework that provides a variety of IP based services. This framework enables wireline, wireless and cable operators to offer rich multimedia services across both legacy circuit switched and new packet switched network infrastructures. Also, together with QoS provisioning, security should be guaranteed in 4G mobile networks.

Dual-Linked BS Model | NETWORK ARCHITECTURES

WiMAX and 3G-LTE systems consider flat network architectures but mobility is still implemented in a hierarchical manner, as shown in Figure 1. An MS traveling within the subnet while changing BSs performs L2 handoff only without changing the MIP attachment. When it moves into another AR area, it triggers L3 handoff. On the other hand, the pure all-IP network suffers from a long handoff latency and high signaling overhead since it incurs L3 protocol at each handoff.

The subnet-based network reduces the frequency of L3 handoff that is accompanied by relatively long latency. Nevertheless, an MS still experiences a long latency when it performs L3 handoff. For seamless L3 handoff, we develop a dual-linked BS model where some BSs are connected to two neighboring ARs at the same time as shown in Figure 2. Obviously, this approach can be extended to support the case where a BS is linked with more than two ARs by adding more links as many as neighboring ARs to that BS. Here, we will use the terminology of "dual" as the general implementation term.

 

Figure 1: The subnet-based access network that has a dual-linked BS

In the conventional subnet-based model, an MS performs L2 and L3 handoffs at the same time when it crosses the boundary of a subnet. This may cause a serious problem of communication blackout because L2 handoff typically exploits a conservative method in preventing the ping-pong effect. It happens like this. An MS starts an L2 handoff when the signal power of the corresponding BS is weak. As an L3 handoff is accompanied by a long latency, the signal may turn too weak during the L3 handoff, resulting in a blackout.

In contrast, the presented network model with some dual-linked BSs decouples L2 and L3 handoffs, thereby providing a flexible handoff mechanism. Since each dual-linked BS can access both ARs of new and old, it helps L3 handoff to be performed independently of L2 handoff when an MS stays in its coverage. An MS entering the area of the dual-linked BS will prepare the L3 handoff. More explanations are given next.

Future Movement Prediction

Generally an MS is able to sense the presence of neighboring BSs since each BS broadcasts its pilot signal. When the MS enters the service area of a dual-linked BS, it triggers L2 handoff. Completing the handoff, it predicts the movement by detecting the pilot strengths of neighboring BSs. When it is likely to move into some other subnet, it prepares L3 handoff. The handoff can be initiated by either the MS or the BS. If L3 handoff is triggered too early, there exists a possibility of too many L3 handoffs, resulting in the pingponging effect. On the other hand, if too late, L2 and L3 handoffs are incurred at the same time. In this case, the handoff delay may not be reduced, because L3 handoff dominates the overall delay.

This motivates to design an algorithm that initiates early L3 handoff following the concept of the existing L2 handoff algorithm. The graph in Figure 2 shows an example of L2 and L3 handoff triggers. In this scenario, if the measured pilot signal strength at the MS from a new BS (BS3) exceeds that of the old BS (BS2) by Th₁ for the time interval I₁ the L2 handoff towards the new BS is triggered (Holma & Toskala, 2000). If the new BS belongs to a different subnet, the L3 hand off is initiated according to the thresholds Th₂ and I₂. In this case, the L3 handoff must start before the next L2 handoff. Deciding the threshold values is an implementation issue.

 

Figure 2: An example of handoff in the dual-linked BS model

Mobile IP Handoff

Following a proper movement detection, an MS performs MIP handoff. The MS begins MIP handoff by sending a request message. After the MS obtains a CoA (care-of-address) for the new subnet, the AR forwards the request message to the MS's Home Agent (HA) to update the MS's location information. During this process, packets arriving at the BS via the old AR will be transferred to the MS by using the new CoA. This is possible because the BS can access the two ARs, which is a unique feature in this scheme. In contrast, in a conventional network, some packets arriving at the old BS or AR will be dropped or forwarded via old and new ARs, so the forwarded packets will experience some latency.

In the dual-linked BS model, each dual-linked BS maintains a table for mapping between old and new CoA during the handoff procedure. Indeed, the handoff latency in MIP is mainly incurred by HA registration. In the new architecture, however, there exists only a little handoff latency since every packet arrives at the same BS via either old or new AR during the handoff. Figure 2 shows an example of the handoff scheme. An MS and its corresponding BS perform L2 handoff whenever it crosses a cell boundary, while performing L3 handoff separately from L2 handoff.

The advantage of the dual-linked BS model is exhibited in Figure 3. In the conventional model, the new BS or AR may receive packets for an MS to which it does not have a direct connection while the MS is performing the handoff between subnets. In this case, the new BS or AR may drop or buffer packets destined for the MS. This case also occurs when the whole handoff process has not completed yet even if the MS has established a new connection already. Therefore, as shown in Figure 3, some packets should be forwarded to the corresponding location during the handoff process. Unless the system supports packet forwarding, packets will be dropped. In contrast, the dual-linked BS model has almost no packet loss during the handoff, because the BS has a connection to each of old and new ARs.

 

Figure 3: Packet forwarding comparison between the conventional and new models

Macro/Micro Cellular Systems Combining Multiple Access | NETWORK ARCHITECTURES

The semi hierarchical cellular systems can be extended for solving the mobility problem. Generally cells are categorized into macrocells, microcells and picocells depending on its size. Macrocells and microcells are usually deployed in rural and urban regions, respectively, while the picocells are in a building. In some region such as a hot-spot zone, an MS can access both macrocells and microcells as in Figure 1. A service model is proposed  where macrocells and microcells cover high speed and low speed MSs, respectively. This structure is effective, because a high speed MS has to change cells frequently if covered by microcells.

 

Figure 1: The model of a hierarchical cellular network

We extend the hierarchical cell structure by integrating multiple access techniques. Some systems under development are based on OFD-MA (e.g., WiMAX) that combines OFDM and frequency division multiple access (FDMA). Since OFDMA systems have lots of channels in a frequency domain, it has higher allocation granularity than OFDM system. It also has the ability of taking advantage of adaptive modulation and coding (AMC), but its application is limited to low mobility. If an MS using OFDMA has high mobility, it cannot perform coherent detection properly due to the long symbol.

Meanwhile, a FH-OFDMA system, which combines frequency-hopping and OFDMA, has the advantage of exploiting diversity (Y. Kim et al, 2003). Though it experiences difficulty in supporting high data rates and AMC, it can overcome channel fading and multiuser interference through a FH pattern. Accordingly, it is a viable combination that microcells for low mobility use OFDMA that has fine granularity, while macrocells for high mobility use FH-OFDMA that is robust to channel fading and interference.

Each cell plane can handle traffic classes differently as well. High rate data services are suitable for OFDMA that has high spectral efficiency and supports various data rates by AMC. In contrast, low rate services like voice are adequate for FH-OFDMA that is easy to use diversity. If an MS has the capability of supporting dual modes, it can switch cells according to mobility and traffic type in a manner of using vertical handoff that offers an additional merit of load balancing.

In summary, the hierarchical network that consists of OFDMA microcells and FH-OFDMA macrocells has the ability of supporting various users with different mobility and traffic types. Since the considered network is based on a common OFDMA platform, it is more manageable than other heterogeneous networks. Also, it provides the flexibility in network planning. 4G networks will be most probably overlaid with 2G or 3G cellular networks. As existing cellular networks are basically designed for circuit-switched voice service, in some regions, they will keep undertaking voice users and high mobility users like the FH-OFDMA macrocell system, while 4G networks focus on data traffic users by using the OFDMA microcell system.