3 Performance Enhancement via Directional Antennas
3.3 MAC Layer: DMAC
3.3.3 Directional Network Allocation Vector (DNAV)
The value of network allocation vector indicates the duration of the ongoing transmission in the vicinity, and the node must reserve the channel for those acting nodes by deferring its own transmission. Directional NAV is an enhancement of NAV,
which reserves the channel only in a certain range of directions. The design of DNAV is to release the medium which is not necessarily reserved, and thus spatial reuse can be achieved. If a node receives an RTS or a CTS packet from its neighbors, a DNAV is set. Each DNAV is tagged with two important values; the direction where the control packet comes from and the duration of the corresponding data transmission.
A node cannot transmit any signals whose direction is in the range of unexpired DNAVs. Another important factor is the width of DNAV, which is based on the beamwidth formed by the directional antennas. The DNAV width of a node must be larger than the beamwidth.
Assume that the width of DNAV is 2w degrees and the beamwidth is 2b degrees.
For a node transmitting safely, it must refer to its DNAV table and the difference between the transmitting direction and all the DNAVs must exceed (w+b) degrees.
In other words, the antenna beam intended to a certain direction must not overlap with any unexpired DNAVs. In the following discussing, we set that the width of DNAV is 45°; the range of DNAV1 is from 315° to 45°; the range of DNAV2 is from 45° to 135°; the range of DNAV3 is from 135° to 225°, and the range of DNAV4 is from 225° to 315°. Figure 3.7 shows a simple example where spatial reuse is achieved.
Node B has data packets to send to node C, and data transmission is initiated after RTS/CTS exchanging. Node A received RTS and CTS packets from nodes B and C, respectively. Two DNAVs with 45° width, DNAV1 and DNAV3, are set upon the reception of these control packets. The numbers in the parentheses represent the relative angles between node A and these two active nodes. If node A has a packet to be sent to node D or node E, it must refer to its DNAV table to check if there is any transmission ongoing in the corresponding direction. Node A cannot transmit any signal to node E until the expiration of DNAV3 which was set upon the reception of the CTS packet from node C. However, transmission to node D is not deferred by
any DNAV, and can be ignited. As mentioned in Section 3.3.1, the DOA is always the most effective direction towards the transmitter with the minimum path loss. This also implies that transmissions towards the DOA of the transmitter could cause the most interference. Therefore, the most effective way to avoid collisions is to set DNAV according to the DOA even the signal DOA does not match where the transmitter is physically located. The width of the DNAV can also be adjusted for the control of aggressiveness of a transmitter. A DNAV with a narrower width makes more directions available for transmission, and thus makes the transmitter more aggressive.
To best exploit the advantages of directional antennas, the routing protocols should be modified as well. The relative direction information of neighbor nodes can be further utilized. In the following section, we introduce the basic concepts of routing protocols in ad hoc networks, and achieve our goal by modifying a well-known routing protocol, the DSDV routing protocol.
3.4 Routing Layer
A wireless ad hoc network is a collection of mobile nodes that are free to move arbitrarily in a certain area where interconnections between nodes could change continuously. Moreover, in a wireless ad hoc network, a node can communicate with every other node in a certain range directly or use other nodes as relays. Due to lack of infrastructure, nodes themselves function as routers which discover and maintain routes to other nodes in the network. Discovering a route is to form a path from a source node to a destination node by selecting nodes in the network as relays.
Maintaining a route is to take actions to reconstruct a broken route when one of the
relays on the route is no longer available. Without an effective construction of routes, packets cannot be delivered reliably from one node to another, especially under a fast changing topology. The design of efficient routing protocols is the important issue in such dynamic wireless networks.
In order to fully exploit the advantages of directional antennas, we want to find a suitable routing protocol. Therefore, our goal is to establish a routing mechanism which can provide the information for physical and MAC layers. Before proceeding with the discussion of the proposed routing protocol, the classification of current routing protocols will be introduced in the following section.
3.4.1 Classification of Current Routing Protocols
There have been several routing protocols proposed for wireless ad hoc networks [21]-[32]. Depending on when the route is computed, these routing protocols may be categorized as: table-driven routing and on-demand routing protocols [33]. Table-driven routing protocols are also called proactive routing or pre-computed routing protocols. As implied by the name, table-driven routing protocols attempt to maintain a table which consists of up-to-date routing information from each node to every other node in the network. In order to maintain a consistent network view, updates of routing information are propagated throughout the network periodically or whenever the link state of the network changes. The advantage of table-driven routing protocols is that a route to the destination is already available when a source wants to send packets to a certain node in the network.
However, the propagation of routing information could result in flooding of update packets, which consumes a lot of the wireless network bandwidth. Destination sequence distance vector (DSDV) [21], clusterhead gateway switching routing
(CGSR) [22], wireless routing protocol (WRP) [23], global state routing (GSR) [24], and fisheye state routing (FSR) [25] are examples of table-driven routing.
On-demand routing protocols is also called reactive routing protocols. In this method, routes are created only when they are needed. In other words, the route may not exist in advance and it is computed just before the packet is sent. When a source needs a route to send packets to a destination, it initiates a route discovery process within the network. This process is completed once a route is found, and the route will be maintained by a route maintenance procedure which includes the detecting and rebuilding of a broken route. The major advantage of on-demand routing protocols is that the precious bandwidth of wireless ad hoc networks is greatly saved.
The bandwidth consumption due to the exchange of routing information is limited because only those routes that are needed will be maintained. However, the source node must wait until such a route can be discovered. Dynamic source routing (DSR) [26], ad hoc on-demand distance vector (AODV) [27], temporally ordered routing algorithm (TORA) [28], dynamic source tracing (DST) [29], associativity based routing (ABR) [30], and signal stability-based adaptive (SSA) [31] are examples of on-demand routing. Furthermore, hybrid methods make use of both to come up with a more efficient one which minimizes the overhead incurred during route discovery and maintenance. Zone routing protocol (ZRP) [32] is an example of hybrid methods.
3.4.2 Modification of DSDV Routing Protocol
In order to establish a routing mechanism which provides information for physical and MAC layers, routing protocols must be redesigned for this purpose. The DSDV routing protocol is a well-designed routing protocol with completed routing functionality. To provide provides information for physical and MAC layers, we propose some modifications to this routing protocol. The DSDV routing protocol described in [21] is a table-driven routing algorithm. In distance vector routing algorithms, every node maintains for each destination a set of distance vectors, each of which includes destination ID, next hop, distance, and so on. In order to keep the distance estimates up-to-date, each node exchanges distance vectors with its neighbors periodically. When a node receives distance vectors from its neighbors, it updates its distance vectors and the shortest distance to every other node is computed.
By combining the next hop of nodes on the path from the source to the destination, a route with the shortest distance is completed in a distributed manner. The distance vector algorithm described above is the classical distributed Bellman-Ford (DBF) algorithm [34]. A significant problem with the DBF algorithm is slow convergence.
A node could take a very long time to build a path to a certain destination, especially after significant changes under the network topology. Another major performance problem with DBF algorithm is that the algorithm could cause routing loops. The primary cause of the routing loops is that nodes choose their next hops in a completely distributed fashion based on information which can possibly be stale, and therefore incorrect. One approach to the routing loop problem is to tag each routing table entry with a sequence number so that nodes can quickly distinguish stale routes from the new ones and thus avoid formation of routing loops.
In the DSDV routing protocol, each node must maintain a routing table
containing the next hop information for all of the possible destinations within the network. Each entry of the routing table is tagged with a sequence number assigned by the destination. As mentioned before, the sequence numbers enable the mobile nodes to distinguish stale routes from new ones, which can avoid the formation of routing loops. Update packets of routing table are periodically transmitted throughout the network to maintain table consistency. An entry with a newer sequence number is always preferred. As for those entries with the same sequence number, the one with a smaller hop count is chosen as the next hop. In other words, DSDV selects the shortest path based on the number of hops to the destination. Routing functionality is completed by exchanging routing table information throughout the network. To alleviate the potentially large amount of bandwidth required by update packets, two types of route update are defined. The first is known as a full dump which carries all the available routing information. The other one is called an incremental. An incremental routing update carries information which has changed since the last full dump. The full dump routing update can be transmitted relatively infrequently when mobile nodes move slowly. Routing information exchanging can be completed by merely incremental routing updates. When movement becomes frequent, and the size of an incremental routing update increases, a full dump can be scheduled. After a full dump broadcast, the size of the following incremental routing update will be smaller.
By employing these two types of update packets, the network traffic can be largely reduced.
Detailed description of the proposed modification to the DSDV routing protocol is introduced in the following sections. The DOA information from the physical layer is utilized here to determine the neighbor distribution in each direction. A neighborhood table records the relative directions to each reachable neighbor of a node, and the omnidirectional received power Pro and thus the M-null directional
antennas with DMAC protocol as well as power control strategy can be implemented.
3.4.2.1 Route Discovery
As in the DSDV routing protocol, the proposed modification requires each mobile node to maintain a routing table which lists all the possible destinations within the network. As shown in Table. 3.1, each entry is tagged with some important routing information such as a sequence number, the metric and the next hop to each destination, and the install time of each entry. The sequence number, as mentioned before, provides a judgment on the freshness of a route. Every time a destination node advertises its routing table, the corresponding sequence number is increased.
Upon receiving the routing information, a route with a more recent sequence number is always preferred. For those candidates with the same sequence number, a route with the smallest metric is selected, and the corresponding node is chosen as the next hop. In a multi-hop environment, the next hop succeeds the data packets and relays forward the destination. The metric represents the hop count of each route all the way to the destination. The install time field indicates when the entry is installed in the routing table. Examining the install time of each route helps to determine when to delete stale routes. In fact, not all of the information in the routing table is exchanged through the network. As shown in Table. 3.2, update packets contains only information about reachable destinations and the corresponding metrics and sequence numbers.
In addition to the routing table, we add a neighborhood table is maintained in each mobile node in DSDV routing protocol. As shown in Table. 3.3, a neighborhood table lists all the reachable neighbors, the corresponding directional angles, and the
omnidirectional received power Pro. Each entry in the neighborhood table is also tagged with the install time as in routing tables. Every time a node hears a signal from one of its adjacent nodes, the DOA and the power receiving of this signal are estimated at the physical layer. Since the signal can be received by this node, the sender of the signal represents a reachable neighbor of this node. If this neighbor is never heard before, it is added into the neighborhood table along with the DOA. If this neighbor is already recorded in the neighborhood table, the DOA is updated.
Therefore, a node can be aware of all its reachable neighbors and the corresponding direction to each. The use of the omnidirectional received power parameter Pro is to calculate the power control factor
β
. We will discuss the cross-layer power control exhaustively in Chapter 4.The neighborhood table not only helps to guides the antenna beam to the right direction, but also provides neighbor angles to form M beamforming nulls in physical layer. Recall that at the MAC layer, a DRTS packet requires a DOA caching mechanism to be sent directionally toward an intended node. That is, when relaying a data packet, a node must check its routing table for the next hop. After making sure that it has a route to the destination, a DRTS packet is sent directionally toward the next hop according to the DOA information recorded in the neighborhood table.
While forming the directional mode, the antennas can also form M beamforming nulls to suppress the neighbor interference as described in Section 3.2.3.
3.4.2.2 Route Maintenance
The mobility of nodes and some other reasons may cause broken links. The broken link may be detected by the MAC layer protocol, or it may also be inferred if no broadcasts have been received for a while from a former neighbor. Since a broken link could result in serious transmission error, a broadcast routing update containing this information should be arranged immediately. In the routing table, a broken link is described by a metric of . Once a node detects a broken link, i.e. an unreachable next hop, it immediately assigns an metric to any route through that next hop and generates an updated sequence number. This is the only situation when the sequence number is generated by any mobile node other than the destination node.
Sequence numbers defined by the originating mobile nodes are generated as even numbers, and sequence numbers generated to indicate metrics are odd numbers.
Upon receiving the notification of a broken link, a node updates its routing table and continues to advertise the broken link. When a node receives an metric, and it has a later sequence number with a finite metric, it triggers a route update broadcast to disseminate the important news about that destination.
∞
∞
∞
∞
In summary, the amount of overhead required for the proposed routing strategy is the extra memory for neighborhood table which includes neighbor DOAs and omnidirectional received power Pro. Consequently, all the physical, MAC, and routing functionality can benefit from the establishment of neighborhood table.
Furthermore, the directional antenna adopted at the physical layer has no effect on the amount of overhead. In other words, a smaller sector size or beamwidth of the directional antennas does not result in more overhead.
3.5 Operation of the Cross-Layer M-Null Directional antennas System
In Sections 3.2, 3.3, and 3.4, we have introduced the M-null directional antennas model, DMAC, and DSDV in physical, MAC, and routing layers, respectively. Now, we present our directional antennas system operating across these layers. Figure 3.8 shows the flowchart of the procedure of forming M-null directional antennas. Before transmission, the first step is that every node broadcasts its routing table. When a node receives the other node’s broadcast signals, the DOAs of neighbor nodes are estimated at the physical layer, and the neighborhood table is updated as mentioned before. When a node has a packet to transmit, the second step is to form M-null directional antennas upon MCMV beamforming to transmit DRTS directionally.
From neighborhood table, the direction of destination and the directions of the other neighbor node which are set to form beamforming nulls are obtained. A node randomly sets M beamforming nulls at first. If interference comes from other direction, directional antennas memorize its direction, and a beamforming null is directed to it by replacing a stale beamforming null. While the transmission is in progress, the node can resist new interference by replacing the stale beamforming nulls set in the initial stage.
Figure 3.9 is an example of a transmission which uses the 4-null directional antennas to resolve the hidden terminal and blocking problem mentioned in Section 2.2.2. Assume that in this figure, a node equipped with five antennas can form one main beam, three side lobes, and four beamforming nulls. If node A intends to send data to node B, it first sends a DRTS packet to node B. According to the neighborhood table, node A forms a main beam toward node B, and four
beamforming nulls randomly toward nodes C, D, E, and F. After receiving DCTS, node B forms a main beam toward node A, and four beamforming nulls randomly toward node C, D, E, and F. During the transmission, node A will not be interfered by nodes C, D, E, and F, so the hidden terminal problems, discussed in Section 2.2.2, will not occur. In addition, nodes C, D, E, and F will be not blocked by node A, because they do not sense the gain of the side lobes of node A. In this scenario, we use the 4-null directional antennas model to address the hidden terminal and blocking problems and fully utilize the potential of spatial reuse.
Assume that node A does not form a beamforming null to node G, as Figure 3.9 shown. When node A sends DRTS directionally to node B with gain of Gd, it also sends DRTS using side lobes to node G with side lobe gain of 1/10Gd. In this case, node G cannot sense the DRTS, since node G listens omnidirectionally with gain of Go in idle state. If node G has packets to send to node C, node G sends DRTS to node C with gain of Gd, then the DRTS interferes with node A, because node A does not form a beamforming null in the direction of node G in the initial stage. Then, node A memorizes the direction of node G, and change the direction of a beamforming null to direct it to node G. Through using the 4-null directional antennas, the two transmission pairs (A←→B, G←→C) can survive at the same time. Although node A is interfered by node G at first, it can also avoid the interference by rearranging beamforming nulls.
3.6 Computer Simulation
To evaluate the performance of the proposed system architecture, we use the
To evaluate the performance of the proposed system architecture, we use the