Hybrid wireless networks provide combined medium access methods selected from cellular, MANET, IP and 802. 1X. Effective resource management, route scheduling and relay strategies are key aspects in facilitating this heterogeneous environment. The hybrid relays have received significant research interest as a consequence of system capacity advantage and reasonable infrastructure cost. The novel HWN* infrastructure proposed here is also expected to provide stable, high-speed and user satisfied communication services in most situations including urban city with both cellular and MANET coverage, indoor environment and remote areas without cellular coverage thus is a challenging design goal. The section starts with a state of the art review for hybrid wireless networks with different design objectives. It then provides the rationale which motivates us to develop HWN* with system concept and architecture. The HWN* infrastructure is further supported by the novel algorithms and protocols proposed at Medium Access layer (MAC), NETWORK transport layer, and cooperation between cellular network, MANET and relay structure. Therefore, algorithms and protocols related to hybrid wireless networks are also briefly discussed.
Hybrid Wireless Networks
Hybrid wireless networks are defined as an integrated infrastructure that provides seamless services over several networks. However, most of research focuses on the infrastructure research itself and few algorithms have been proposed to explore hybrid wireless network practical usage at medium access layer and routing layer. Current algorithm suites are normally proposed for system capacity, relay station placement plan, cooperative resource sharing and path discovery to improve only the cellular network performance. These solutions leverage the presence of persistent resources to support relay networks, but rely on the fact that ad hoc multi-hop services being underestimated and multiple interface technology is not comprehensively considered. The service-oriented algorithm for hybrid wireless networks was still left for further investigation as most previous proposals only concentrate on the infrastructure design, which assume such services would be supported or extended based on existing cellular differentiated service methodologies.
For the design of effective resource and routing management algorithms in a complex system, centralised control approaches should be avoided as network scalability issues largely jeopardises the hybrid wireless network performance. The introduction of system flexibility by avoiding central control motivates this research to propose algorithms that consider cross layer communication issues and differentiated services in a decentralised manner that influence the global system performance through local conditional changes. The challenges for communication service provision have increased in many respects in recent years. Competition among providers demands a continuous reduction of production costs and improvement of service quality, which motives us to propose the cost-effective novel HWN* infrastructure and develop associated algorithms for the infrastructure realise with guaranteed QoS. The algorithm development does not require a brand new ubiquitous radio access technology or focusing on multi-hop based cellular networking technologies, but rather, the work integrally utilises existing cellular, ad hoc and relay technologies in an integrated fashion, and combines their advantages and overcomes their inefficiencies.
The research addresses several issues from handover management and routing perspectives, evaluates proposed algorithms in the HWN*, and finalises the HWN* management framework towards 4G mobile system. It is also concerned with the provision of fixed relay nodes to subscribers in both urban and remote areas and also implementing coordinated cellular and MANET radio access. It devises a framework for studying the trade-offs of interworking between the two active systems, identifies and answers the specific design space questions such as: How to share the resource between different service classes? How many relay nodes are sufficient in a reasonable cost? Where should we place RNs? What protocols are necessary to facilitate the load balancing between the two systems? How to handle network routing in a heterogeneous environment? Why the infrastructure uses relays and MANET to reduce the number of wired BSs? The study answers these questions in order to demonstrate that the proposed integrated infrastructure can be used for providing the enhanced data communication services.
Recent research attempts for MANET-cellular infrastructure integration include Multi-hop Cellular Network (MCN), Multi-Power Architecture for Cellular Networks (MuPAC), integrated Cellular and Ad hoc Relaying system (iCAR), Self-organising Packet Radio Networks with Overlay (SOPRANO) and the IST-WINNER project. The basic rational of the MCN is a cellular network evolution that concentrates on cellular radio access technology. Traditionally, a MT and a BS have a direct link in 2G or 3G cellular networks, but in the MCN a MT may reach the serving base station by using multi-hop relaying. The relay is called soft MT based relay which refers to other MTs act as relay clients using ad hoc frequencies. The proposal also states that the relay node could be an infrastructure node if the condition allows. However, it only generally concludes that the relay has capabilities like those of the BS such like an ad hoc node access point. It does not exactly solve problems such as how to choose packet delivery route or when relay mode should be used other than original cellular communications. The analysis and simulation results present the throughput comparison between conventional Single hop Cellular Network (SCN) and MCN. The throughput of MCN is better than that of SCN and increases as the transmission range decreases. The node uses a transmission range rthat is a fraction 1/k of the cell radius R where r = R/k. The parameter k is referred to as the reuse factor. The research explains these two observations by illustrating the different increasing orders, of mean number of channels such as simultaneous transmissions in a cell, and mean hop count, as the transmission range between source node and destination node decreases by k times. But unfortunately the actual gain will be lower. First of all, large control overheads are produced since every node may perform routing updates even when there is no topology change. If a MANET routing protocol e.g. ADOV is used for multi-hop routing, there is a high possibility of relay client MTs absence. The main disadvantage of the relay is the latency in route discovery and link failure. On the other hand, as the ad hoc frequency is assumed for packet relaying, a small topology change results in medium access failure on the fragile link.
MuPAC is an extended MCN with focus on node flexible power management. The system architecture is exactly the same as MCN and each MT uses a separate 802.11X interface. The multiple transmission powers helps MuPAC achieve maximised spatial reuse without substantially increasing the number of hops. The work certainly improves the system capacity performance since the possibility of a path break in MuPAC is lower than MCN once the algorithm is used. Throughput Enhanced Wireless in Local loop (TWiLL) is also an extended work based on MCN. It has been proposed by the same researchers and focuses on the power management of wireless multi-hop local loop to increase the hybrid system capacity. However, here it is highly possible that nodes increase the transmission power to reduce the path break probability and hop distance. It should also implement a distributed power management algorithm at each MT which is computationally expensive. Otherwise the power management has to be coordinated by BS central control because no dedicated relay node is used.
iCAR is not difficult to be evolved from the cellular network and it can be also categorised as one type of MCNs. Other than increase the system capacity through node power management or a longer hop distance. The infrastructure works on adaptive traffic load balancing in this multi-hop cellular infrastructure. It is the first hybrid wireless network that looks at the horizontal handovers between cellular accesses through MANET access. Extra cellular bandwidth available in surrounding cells can be borrowed to the congested cell through dedicated ad hoc relay modes named primary, secondary and cascaded relaying. The channel borrowing evaluation results indicate the improvement on packet congestion delay over conventional cellular networks and MCN, and it verifies that with a limited number of relay nodes, the call blocking and dropping probability in a congested cell as well as the overall system can be reduced. However, the simulation evaluation of iCAR system suffers unfair packet contention problem as all packets are treated exactly the same. An extra bandwidth can be allowed to any packets without considering packet priority, packet transmission requirement and QoS. For example, an urgent communication request or instant request can be blocked or terminated due to the contention and unreasonable channel borrowing. And this is also why differentiated traffic input should be introduced for hybrid wireless network evaluations. The iCAR has introduced a novel concept called managed mobility for relay nodes based on its signalling protocol and node mobility model. The relay node movement provides assistance on channel borrowing in congested communication areas but at the expense of complex and two layered route management (MT layer and Relay Layer). Furthermore, each relay station in iCAR must be equipped with location tracking system with extra cost. Another drawback of iCAR is that the system does not explore or underestimate the dedicated relay node assisted MANET communication mode capability. It only intends to reuse the cellular resource via multi-hop ad hoc relaying.
SOPRANO is a scalable architecture that assumes the use of multi-hop cellular and asynchronous CDMA with spreading codes to support high data rate Internet and multimedia traffic. The general idea of SOPRANO is not much different from iCAR other than IP network support and cross active network connections. It focuses on connection establishment and node power control based self-organisation, and investigates the formulation for an optimum transmission strategy. Again, a power management algorithm is proposed to adaptively selection transmission power to improve the system capacity, which is similar to MuPAC. The system presents high capacity bounds that illustrate how the technique helps in trading off conserved power for a multi-fold capacity advantage. However, both SOPRANO and iCAR rely heavily on soft MT based multi-hop traffic relay underestimate the usage of an alternative MANET with dedicated and location fixed RN infrastructure since fixed relay nodes are not considered.
European Commission Information Society Technologies (IST) WINNER project proposed the WINNER system. What WINNER focuses on is the Research & Development of a brand new beyond 3G radio interface technologies needed for a ubiquitous radio system. The RNs are deployed to incorporate with BSs realising an efficient and flexible spectrum usage and spectrum sharing environment, where the RNs only implement their new ubiquitous radio interface. The RNs are planned and share the same Radio Access Technology (RAT or refereed as interface) with BSs and MTs. Ad hoc communication is prohibited in this architecture and one can see the WINNER as an evolving node-oriented multi-hop cellular network with relay support. One disadvantage of WINNER can be the implementation cost. Significant hardware and software updates are required at base station radio network controller, relay node and end-user equipment to apply the novel radio interface. The telecom providers may not ready to replace anew radio access deployment without significant system performance improvement compared to current 3G cellular system.
The proposed HWN* has major differences from other hybrid wireless networks motioned previously. It has been summarised in a nutshell in the important issues comparison between HWN*, WINNER and SOPRANO architectures. Detailed comparison between iCar, MuPAC and MCN. Although advanced technologies such as location tracking make soft relay based infrastructure feasible, the route recalculating and reconfiguring in systems such as MCN, TWiLL and MuPAC, without dedicated relay nodes, are unstable. Meanwhile, there are other fundamental problems such as relay node absence and third party terminal relay security. Using dedicated and location fixed relay node provides straightforward method to enable reliable communication. The research will later compare system performance between HWN* with dedicated RNs, MCN with soft relay, SOPRANO and WINNER. SOPRANO is considered other than iCAR since CDMA based SOPRANO gives better results in terms of system capacity and network delay.
With dedicated and location fixed RNs support, the adaptive and scalable HWN* has four basic communication modes, which is cellular communication (also named BSON), RN supported cellular communication (BSON RN), MANET communication and RN supported MANET. With the assumption of coded and modulated digital communications, Node Xcan transmit information to Node Y via one or more RNs. The dedicated RN can be part of a cellular network and a MANET. The nodes X and Y can be a BS, a MT or a dedicated RN. And, the term communications include uplink communications (link from MT to BS), downlink communications (link from BS to MT), MT to MT communications or BS to RN communications. Two MTs may communicate directly or through an intermediate node (The node can be a RN or a group of RNs). The MT can be also accommodated into cellular network with dedicated RNs assistance. Therefore, the relay structure is viewed as a means of extending the communication coverage of either a cellular network or MANET. MANET and cellular network are mutually supported through the use of RN structure. The performance of individual MANET or cellular network, respectively, is also enhanced by RN support. Figure 1 presents the topology of the HWN* used for handover mobility management and cross-layer routing. The RNs create a mesh structure to support node communication through RN infrastructure. The procedure is similar to 802.1 IX node-to-infrastructure communication but a virtual backbone is constructed between RNs. The BSs may connect to an IP network via fixed lines or switching nodes. The HWN* assumes that there exists wireless connections between RN and BS, and between RN and RN.
Figure 1: Hybrid wireless network with dedicated and fixed relay nodes (HWN*)(© 2008, Chong Shen. Used with permission.)
The HWN* is a generic system and works with any off-the-shelf air interfaces, as an example, for the cellular part of HWN*, Time Division Multiple Access (TDMA) based system is used. This means that every individual transmitting channel required as part of the chain between any two terminals is created by allocation of time slots such as multiple time slots in Enhanced Data rate for GSM Evolution (EDGE). Variability of data rate is achieved by allocating differing number of time slots. The TDMA cellular interface allows the HWN* to take advantage of multi hop connections formed through RNs with more flexible implementation compared to Frequency Division Multiple Access (FDMA) (3GPP). The HWN* network deployment scenarios against physical layer link duplex model, medium access method, spectrum usage, node movement speed, transmission rate and overall scenario capacity. Traditional MANET scenario can not support high node mobility speed and data transmission rate without the presence of infrastructure node such as 802.11g access point or dedicated RN, thus high mobility and data rate can be realised in RN supported MANET. System capacity of cellular network or RN supported cellular are optimised as the introduction of RN optimises the resource sharing performance. We implement standard CSMA/CA for pure MANET and synchronised CSMA/CA for RN supported MANET mode considering IEEE 802.11e QoS standard for delay-sensitive traffic and the RN is given priority in terms of medium access. The RN priority access will be detailed in cross-layer routing algorithm proposal. Large scale deployment of dual-interfaced RNs is cost-effective as the equipment can be based on an integration of a modified 802.11 access point and cellular packet relaying function node. Each RN associates with one BS so that the radio resource usage in each RN can be coordinated. A decrease in BS density can be compensated by an increase in RN density, in order to maintain constant performance. The RN has properties and functionaries of:
- Two radio interfaces: The cellular interface and the MANET interface.
- The RN extends the cellular service range and optimises cell capacity.
- The RN minimises node transmitted power.
- The RN covers remote areas, supports inter network load balancing.
- The meshed RNs provide an alternative communication mode which is MANET with RN infrastructure for MANET based resource management and routing.
Theoretically, both the HWN* system capacity and the average packet delivery ratio per MT, compared to traditional 2G and 3G cellular networks, should be improved because the RNs provide relay capability as the substitution of a poor quality single-hop wireless link with a better-quality link encouraged whenever possible. The disadvantage of the RN integration is that whether in reality the infrastructure can be realised or not due to feasibility issues. Telecom providers should first agree, design and prototype such equipment. After identify air interfaces for MANET access and cellular access, both software and hardware are required to be upgraded on actual relay nodes.
No comments:
Post a Comment