0615a12906a97c4a3612f9fbd6d408da.ppt
- Количество слайдов: 66
Wireless Networks and Mobile Ad Hoc Networks D. Manivannan 1
Overview • Wireless network evolution • Mobile ad hoc networks • Issues and challenges in Mobile Ad Hoc Networks – Routing, multicasting, resource discovery, Qo. S, security, etc. • Sensor networks 2
Wireless networks evolution • Characteristics of wireless communication – Higher interference results in lower reliability • Infrared signals suffer from interference from sunlight and heat resources and can be absorbed/blocked by various objects and materials. Radio signals usually are less prone to be blocked but can be interfered with other electric signals • Broadcast nature of transmission means all devices are interfering with each other – Low bandwidth availability causes degraded Qo. S, delay, longer connection setup; bandwidth typically few Kbps to a few Mbps – Highly variable network conditions • Higher data loss rate due to interference, mobility, decrease in received signal strength due to distance 3
Wireless networks evolution… • Characteristics of wireless communication… – Limited computing and energy resources for mobile nodes • Limited computing power, memory, disk size, battery capacity, limitation of device size, weight and cost – Limited transmission resources • Medium sharing, limited availability of frequencies with restricted regulations – Device size limitation: due to portability requirements limits in user interfaces and displays – Weaker security: radio interface is accessible to everyone, network security is more difficult to implement. 4
Wireless networks evolution… • Types of wireless networks – a classification – By network formation and architecture • Infrastructure-based networks: e. g. , WLANs, cellular networks • Infrastructureless (ad hoc) networks: there is no prearrangement regarding the specific role each node should assume. For example, two laptops can set up an independent network whenever one is within the range of the other. In ad hoc networks, each node is expected to behave as a router and take part in the discovery and maintenance of routes to other nodes as well as forwarding packets for other nodes. 5
Wireless networks evolution… • Types of wireless networks – a classification – By communication coverage area • WWANs (Wireless Wide Area Networks): these are inrastructure-based networks that rely on networking infrastructures like cellular networks and satellite networks. • WMANs (Wireless Metropolitan Area Networks): These are sometimes referred to as fixed wireless. These are infrastructure-based networks that enable users establish broadband wireless connection among multiple locations within a metropolitan area. For example, multiple office buildings in a city could be connected without incurring the high cost of laying fiber or copper cables. IEEE has set up a specific 802. 16 working group on Broadband wireless access standards for developing standards for this (Wi. MAX forum is active in the implementation of part ofstandard). For more info visit www. ieee 802. org/16 6
Wireless networks evolution… • Types of wireless networks – a classification… – By communication coverage area… • WLANs (Wireless Local area networks): These enable users to establish wireless connections within a corporate or campus building or public place such as airport, coffee shop, etc. WLANs can be used for flexible data communication in temporary offices or other spaces to supplement an existing LAN. Offices, homes, coffee shops, airports, conferences, are typical hotspots for WLAN installations. – Typical WLAN implementations include 802. 11 (Wi-Fi – Wireless Fidelity), IEEE standard and Hiper. Lan/2, developed by European Telecommunications Standards Institute Broadband Radio Access Networks project. Typical transmission range varies for 2 Mbs to 600 Mbps. (802. 11 a: 2 Mbps; 802. 11 b: 11 Mbps; 802. 11 g: 54 Mbps; 802. 11 n: 600 Mbps) 7
Wireless networks evolution… • Types of wireless networks – a classification… – By communication coverage area… • WPANs (Wireless Personal Area Networks): these enable users to establish ad hoc wireless communication among personal wireless devices such as PDAs, cell phones, laptops, home appliances etc. IEEE has set up a specific 802. 15 working group for developing standards for this. For more info visit www. ieee 802. org/15. Initial release of the standard is based on Bluetooth (www. bluetooth. org) which is a short range cable-replacement technology. Bluetooth is gaining popularity due to low complexity, low power consumption, interoperablity with 802. 11 networks. Now (2012), cumulative Bluetooth devices shipped surpassed 9 billion. They are used by BMW, IBM, HP, Toyota, etc. In 2012 more than 17000 companies belong to the Special Interest Group(in 1998 five companies founded this group). – Typical coverage area is up to 10 m 8
Wireless networks evolution… • Types of wireless networks – a classification… – By communication coverage area… • WBANs (Wireless Body Area Networks): this is related to wearable computers that are distributed over the body (e. g. , head-mounted displays, microphones, earphones, sensors, etc). Main requirements of a WBAN are: – Ability to connect heterogeneous devices – Autoconfiguration capability (adding removing devices should be transparent to user) – Ability to connect with other BANs (to exchange data with other people) or PANs (to access Internet, for example) 9
Wireless networks evolution… • Types of wireless networks – a classification… – By access technology: Depending on the specific standard, frequency and spectrum usage, wireless networks can be categorized based on the access technology used • • GSM networks TDMA networks CDMA networks Satellite Networks Wi-Fi (802. 11) networks Hiper. Lan/2 networks Bluetooth networks Infrared networks 10
Wireless networks evolution… • Types of wireless networks – a classification… – By network applications: based on the specific usage, wireless networks can be classified as • • • Enterprise networks Home networks Tactical networks Sensor networks Pervasive networks Wearable networks 11
Wireless networks evolution… • Forces Driving Wireless technology evolution – Existing technologies • 1 G introduced in the 80’s which supported analog cell phones using FDMA. E. g. , NMT and TACS in Europe and AMPS in North America • 2 G introduced digital mobile systems and added fax, data and messaging capability. • 3 G systems offer increased bandwidth – 128 Kbs when a mobile moves at high speeds, 384 Kbps at pedestrian speeds, and upto 2 Mbps in stationary applications. Europe and Asia are promoting W-CDMA and EDGE whereas North America is promoting cdma 2000 and each is developed by different standard bodies. 12
Wireless networks evolution… • Forces Driving Wireless technology evolution… – The need to integrate various types of networks – 3 G cellular networks, WLANs and WPANs – The need to support high-speed multimedia services – The need for convergence in network infrastructure • Solution: an all IP network. – The need to support high mobility and device portability – The need to add location intelligence: i. e. , support for location-based information services – The need for greater standard interoperability 13
Wireless networks evolution… • 4 G networks architecture and capabilities – Are touted as hybrid broadband networks that integrate different network topologies and platforms • First level of integration addresses the integration of wireless networks with varying transmission characteristics such as WLAN, WPAN as well as mobile ad hoc networks • The second level of integration includes the fixed networkbackbone infrastructure – All-IP networks : It is based on the assumption that future networks will be entirely packet switched using protocols evolved from those in today’s Internet (networks are likely to provide upto 100 Mbps bandwidth). • This means the core 4 G network can be designed and can evolve independently from access networks. This approach can tap the rich protocol suits already available such as Vo. IP protocols such as SIP, H. 323, etc. 14
Wireless networks evolution… • 4 G networks … – Lower cost and higher efficiency: • Equipments are likely to be cheaper • An open converged IP network means further reduction in the network buildout and maintenance cost • No need to purchase extra spectrum, as 2 G/3 G spectrum can be reused for 4 G and much of the spectrum needed for WLANs and WPANs is free. 15
Wireless networks evolution… • 4 G networks … – Ultra high speed for multimedia applications: • At speeds upto 100 Mbs, will be able to provide high -bandwidth wireless services such as TV, music, browsing Internet, real-time video steaming, and other multi-media applications whether one is in office or home – Ubiquitous computing: • one of the goals of 4 G networks is to provide pervasive computing environments that can seamlessly and ubiquitously support users in accomplishing their tasks • These devices personalize themselves in our presence to find information or software needed 16
Wireless networks evolution… • 4 G networks … – Support for ad hoc networking: • Non-infrastructure-based mobile ad hoc networks are expected to be an important part of the 4 G architecture – Location Intelligence: • To support ubiquitous computing requirements, 4 G terminals need to be more intelligent in terms of user’s location and service needs • Outdoor wireless applications can use GPS for obtaining location information. Indoor devices can use techniques like infraread, ultrasound. • Possible location-based services include: finding nearest service providers such as gas station, restaurants; searching for special offers; searching for other co-located users. 17
Mobile ad hoc networks (MANETs) • Characteristics of MANETs – Wireless: nodes communicate wirelessly and share the same media, radio, infraread, etc – Ad-hoc based: a temporary network formed dynamically by a collection of nodes in an arbitrary manner as the need arises – Autonomous and infrastructure-less: does not depend on any established infrastructure or centralized administration – Multihop routing: no dedicated routers are necessary. Every node acts as a router and forwards each others packet to enable information sharing – Mobility: each node is free to move about while communicating with other nodes • On the web about MANETs: – www. ietf. org/html. charters/manet-charter. html – protean. itd. nrl. navy. mil/manet_home. html 18
Mobile ad hoc networks (MANETs)… • Applications of MANETs – Tactical networks: Military communications, operations. Automated battlefields – Sensor networks: collection of embedded sensor devices to collect data in hazardous environments, natural habitats, etc. – Emergency operations: search and rescue operations, disaster recovery – Educational applications: set up virtual classroom or conference 19
Mobile ad hoc networks (MANETs)… • Design issues and constraints: – Infrastructureless: lack of centralized entity means network management has to be distributed across nodes which makes fault-detection and management difficult – Dynamically changing network topology: • Results in route changes, frequent network partitions and possibly packet loss – Physical layer limitation: • Communication is inherently broadcast in nature • Limited transmission range results in specific problems such as hidden terminal problem and exposed terminal problem • Collisions are inherent to the medium and has higher probability for packet loss compared to wireline networks. 20
Mobile ad hoc networks (MANETs)… • Design issues and constraints: – Limited link bandwidth and quality: • Wireless links have significantly lower capacity than wired links – Variation in link and node capabilities: • Each node may be equipped with one or more radio interfaces that have varying transmission/receiving capabilities, which can result in asymmetric links • Designing network protocols for such heterogeneous networks can be complex – Energy constrained operation: • Nodes have limited battery power. Hence designing energyaware protocols is important 21
Mobile ad hoc networks (MANETs)… • Design issues and constraints: – Network robustness and reliability: • A node can act selfishly or a node may fail to forward packets to save energy • Misbehaving nodes can have severe impact on overall performance of the network • Lack of centralized monitoring and management points means these types of behaviors cannot be detected. – Network security: Nodes are more vulnerable to information and physical security threats. Some security requirements in ad hoc networks: • Confidentiality: preventing eavesdropping • Access control: protecting access to wireless network infrastructure • Data integrity: preventing tampering with traffic (i. e. , accessing, modifying, or injecting traffic) • Denial of service attacks by malicious nodes 22
Mobile ad hoc networks (MANETs)… • Design issues and constraints: – Network scalability: • Most of the existing network protocols designed for ad hoc networks have been designed for small networks. So, designing protocols that would scale for large networks (tens of thousands of nodes) such as tactical networks and sensor networks is important • Many issues, such as addressing, location management, routing, configuration management, interoperability, security, etc. need to be addressed – Quality of service: A Qo. S guarantee is essential for successful delivery of multimedia network traffic. Qos involves several metrics such as throughput, packet loss, delay, error-rate, etc. 23
Overview of research in MANETs • • • Media access control (MAC) optimization Routing in Ad Hoc networks Multicasting and broadcasting Resource discovery Quality of service Security 24
Overview of research in MANETs… Media access control (MAC) optimization • Two issues that arise in wireless networks if we use CSMA/CD based MAC protocols – Hidden terminal problem: • Two terminals A and C cannot detect each others transmissions due to being outside the transmission range of each other but their transmission ranges are not disjoint. – Exposed terminal problem: • This results from situations in which a permissible transmission from a mobile station (sender) to another station has to be delayed due to irrelevant transmission activity between two other mobile stations within the sender’s transmission range. (this can result in throughput reduction) 25
MANETs… Media access control (MAC) optimization • Techniques proposed to handle the hidden terminal problem – Many protocols such as MACA, FAMA (floor acquisition multiple access), MACA-BI (MACA by invitation) have been proposed to solve the hidden terminal problem – Basic idea behind these solutions is • When a node wants to send a packet to a neighbor, it sends a RTS (request to send) control packet. The receiver then consents to the communication by sending a CTS (consent to send packet). Upon receiving the CTS packet, 26 the sender can start transmission.
Routing protocols for MANETs • Important criteria and considerations used in designing and comparing new routing protocols include: – Simplicity and ease of implementation – Rapid route convergence – routes should be loop-free and optimal, and possibly multiple routes should be available – Distributed but light weight in nature – can quickly adapt to changes in topology and traffic pattern resulting from mobility and failure conditions; control overhead should be minimum – Bandwidth, power and computing efficient with minimum overhead – Scalable, secure and reliable 27 – Supporting quality of service requirements
Routing protocols for MANETs… • Two main classes of routing protocols – Proactive routing protocols • Attempt to maintain a consistent up-to-date routing information between every pair of nodes in the network by propagating, proactively, route update messages at fixed time intervals • Routing information is maintained in tables at each node and hence this is also called table-driven – Reactive on demand routing protocols • In this approach, a route to a destination is established only when there is a demand for it, usually initiated by a source node through a route discovery process. • A route discovered is maintained until (i) the destination becomes inaccessible along every path from the source, (ii) the route is no longer used, or (iii) it has expired 28
Routing protocols for MANETs… • Some proactive routing protocols – Destination sequenced distance vector (DSDV) protocol • It is a distance vector protocol with extensions to make it suitable for MANETs – Clustered gateway switch routing (CGSR) protocol • This extends DSDV with cluster framework which increases scalability – Wireless routing protocol (WRP) • Four tables are used to maintain distance, link cost, routes, and message transmission information. – Optimized link state routing protocol (OLSR) • It is an optimization of pure link state protocol 29
Routing protocols for MANETs… • Some reactive routing protocols – Dynamic source routing (DSR) – Ad hoc on-demand distance vector (AODV) routing – Routing Protocol with Selective Forwarding (RPSF) – Temporally ordered routing algorithm (TORA) – Associativity-based routing (ABR) – Signal stability routing (SSR) 30
Routing protocols for MANETs… Proactive routing protocol - DSDV • Destination-Sequenced Distance Vector routing (DSDV) (Perkins and Bhagwat SIGCOMM ’ 94) – Each node maintains a route to every other node – The routing table contains the following information for each entry. (destination address, dest. seq. number, next-hop, hop-count, and install time) The dest. seq number is originated by the destination. – It uses both event-triggered and periodic routing table updates – Every time interval, each node broadcasts to its neighbors its current sequence number, along with any routing table updates. The routing table update is of the form (destination address, dest. seq number, hopcount) 31
Routing protocols for MANETs… Proactive routing protocol - DSDV • Upon receiving an update message, the neighboring nodes use this information to update their routing table entries using an iterative distance vector approach • In addition to periodic updates, a node also sends eventtriggered updates to announce important link changes such as link removals • If a node learns two disjoint paths for a destination, it selects the path associated with the greatest associated sequence number. • Two optimizations are also implemented for improvement in performance – Full update: these are transmissions of the entire routing table. These are performed relatively infrequently – Incremental updates: these updates include only the entries that have changed since the last full update. These are transmitted frequently. Once the number of routing changes become too large to fit in a single NPDU (network protocol data unit), a full update is transmitted 32
Routing Protocols for MANETs… Proactive… DSDV… • DSDV’s mechanism to damp routing fluctuations – Routing updates for a given destination can propagate along different paths at different rates – To prevent a node announcing a route change for a destination while another better route is still enroute for the same destination, DSDV requires nodes to wait a settling time before announcing a new route. The settling time is the weighted average time that routes to a particular destination fluctuate. 33
Routing Protocols for MANETs… A Reactive Protocol - AODV • Ad Hoc On-Demand Distance Vector (AODV) routing protocol (C. E. Perkins and E. M. Royer) – Like many of the reactive protocols, route discovery cycle involves a broadcast network search and a unicast reply containing discovered paths – It relies on per node sequence numbers for ensuring routes are loop-free – Each node maintains a route table which contains the next hop information for destination nodes – Each entry in the route table has an associated lifetime period; if the route is not utilized within this time period, the route is expired. 34
Routing Protocols for MANETs… A Reactive Protocol – AODV… • Route discovery under AODV – If a node does not have a route to a destination to which it wants to send packets, it creates a route request (RREQ) packet – RREQ includes the id of the destination, the last sequence number for the destination and the source’s address and current sequence number. RREQ also contains a hopcount initialized to 0, and a RREQ ID, which is per node monotonically increasing counter which is incremented every time a new request is initiated – The source address together with the RREQ ID uniquely identifies the request and can be used to detect duplicate request. – After creating the RREQ, the source broadcasts this request 35 to all its neighbors
Routing Protocols for MANETs… A Reactive Protocol – AODV… • When a node receives the RREQ – It creates a reverse route to the source node with the node from which it received the RREQ as the nexthop to the source node – Increments the hopcount in RREQ by 1. If it does not have an unexpired route to the destination, it rebroadcasts the RREQ with the new hopcount value. RREQ floods the network in this manner until it reaches the destination or a node that has a route to the destination. – If it has an unexpired route to the destination, and the destination sequence number of the route is greater than or equal to the destination sequence number received in the RREQ ( destrt ≥ dest. RREQ), then it sends a RREP (this ensures that the most recent route is selected and also guarantees 36 loop freedom)
Routing Protocols for MANETs… A Reactive Protocol – AODV… • When a node receives the RREQ… – The RREP contains the source node address, destination node address and the destination sequence number as given by the route table entry. In addition, the hopcount field of the RREP is set to the node’s distance from the destination (if the destination itself is creating RREP, it is set to 0). – The reverse route that was created as the RREQ was forwarded is utilized to route the RREP back to the source node. – When an intermediate node receives the RREP, it creates a forward entry for the destination node; it uses the node from which it received the RREP as the next hop toward destination; the hopcount for that route is the hopcount in the RREP, incremetned by 1; this entry will be used to route packets to the destination if the source chooses this path. Then the RREP is forwarded – After the source receives the RREP, it can use it for data transmission. If the source receives more than one RREP, it uses the one with the largest sequence number and smallest hopcount. 37
Routing Protocols for MANETs… A Reactive Protocol – AODV • Route maintenance in AODV – When a link break along an active path occurs, the node upstream of the break (i. e. , closer to the source node) invalidates the routes to each of the destinations (which use this node as next hop) in its route table. Then, it creates a RERR message. – The RERR message contains the list of nodes that are now unreachable due to the loss of the link – The RERR message is sent to the upstream neighbors that were utilizing this link. Thus, the RERR message propagates to the source and the source can repair the route if it is still needed. 38
Routing Protocols for MANETs… A Reactive Protocol – AODV • An Optimization in AODV – To improve performance and reduce overhead, source nodes can use expanding ring search approach to search by modifying the TTL (time to live) field of the RREQ packet. – Incrementally larger areas of the network are searched until a route to the destination is found. – Network-wide flooding of RREQ messages is prevented if route can be found in a local area. 39
Routing Protocols for MANETs… A Reactive Protocol – AODV • Another optimization in AODV – When a link break occurs, the node upstream of the break can try to repair the link locally – If it is not possible to repair the break locally, it can send a RERR message to the source. 40
Routing Protocols for MANETs… A Reactive Protocol – AODV • Optional features in AODV to improve operation – During route discovery, if only intermediate nodes respond, the destination may not have a route to source – AODV defines a gratuitous RREP that can be sent to the destination when an intermediate node creates a route reply. – Another optional feature is RREP-ACK (route reply acknowledgement). When unidirectional links are suspected, this can be used to ensure the next hop received the RREP. If a RREPACK is not received, blacklists can be utilized to indicate unidirectional links so that these links are not used in future route discoveries. – AODV allows periodic hello messages for monitoring connectivity 41 to neighbor nodes
Routing protocols for MANETs… • Other approaches taken for designing routing protocols for MANETs – Geographical approaches • These protocols build on proactive and reactive techniques and in addition incorporate geographical information to aid in routing – Hybrid approaches • Hybrid protocols may exhibit proactive behavior under certain circumstances and reactive behavior under a different set of circumstances – Clustering and hierarchical approaches • To increase the scalability, these protocols place the nodes into groups, called clusters. The groups may be based on a number of criteria, but mostly based on either location or functionality. – Other techniques • Multipath routing, energy-conserving protocols, security-aware protocols 42
Routing protocols for MANETs… • Geographical approaches for routing – These protocols use geographical information to aid routing – The geographical information of a node can be in the form of actual geographical coordinates, as obtained through Geographical positioning system (GPS) or can be obtained through reference points on some fixed coordinate system – Using the geographical information can prevent the network -wide search for destinations, since the data or control packets can be sent in the general direction of the destination, if the recent geographical coordinates of that destination are known – Some disadvantages of this approach • Every node should have continuous access to the geographical coordinates of every other node in the system. • A route may not be found even though one exists. 43
Routing protocols for MANETs… Geographical approaches… • Location-Aided Routing (LAR) proposed by Ko an Vaidya (Mobicom 98 best paper) – A reactive protocol that uses the geographical coordinates to direct route request messages to previously known location or destination – The protocol defines two areas: expected zone and request zone • Expected zone is the area in which the destination is most likely to be present. If the location D of the destination at time t 0 and the velocity v of the destination are known, then the expected zone of the destination at time t 1 is the circle of radius |v|(t 1 - t 0 ) and center D. • The request zone is the area in which the route request for the destination should propagate. Is defined as the smallest rectangle that contains the expected zone and the source node 44
Routing protocols for MANETs… Geographical approaches… LAR • LAR route discovery – When a source needs a route to a destination, it creates a route request (RREQ) message – If the source did not have a route to the destination recently, it uses simple flooding to find a route – If the source recently had a route for the destination, it calculates the expected zone and request zone and places the coordinates of the request zone boundary in the RREQ message. • When a node receives the RREQ: – if it lies in the request zone, it processes the packet and rebroadcasts it or sends a reply depending on whether or not it has a current route to the destination; otherwise, it discards the packet. 45
Routing protocols for MANETs… Geographical approaches… LAR • Some advantages and disadvantages of LAR – Size of the request zone is a tradeoff between control overhead and probability of finding the destination – Small request zone runs the risk of not being able to find the destination – Although the destination may lie within the request zone, the path to the destination may not lie within the request zone, in which case, the route will not be discovered – If the request zone is two large, the savings in control overhead will be minimal 46
Routing protocols for MANETs… Geographical approaches… LAR • Another way of determining request zone – When a source node sends RREQ, it places its distance to the destination’s previous location DISTs in the RREQ. – When an intermediate node i receives the RREQ • It computes its distance DISTi to the destination. – If (DISTs +δ) ≥ DISTi then it processes the request, else discards the packet (here δ is some parameter. In practical implementations δ is 0 ). This δ is to prevent nodes that are close to the source rebroadcast the message. – When it forwards the request, the node replaces DISTs 47 with DISTi
Routing protocols for MANETs… Hybrid approaches – ZRP • Zone routing protocol (ZRP) (Pearlman and Haas, ICC 1998; enhanced version JSAC 2002) – It integrates both proactive and reactive components into a single protocol – Around each node ZRP defines a zone whose radius R is measured in terms of hops. – Each node utilizes proactive routing inside its zone and reactive routing outside its zone. – If the destination lies within its zone, a route will exist for the destination in its routing table (Intrazone routing). – If the destination is not within its zone, a search to find a route is needed (Interzone routing). 48
Routing protocols for MANETs… Hybrid approaches – ZRP… • Routing under ZRP – Intrazone Routing: • The routing zone of a node is defined to be the set of nodes that are R -hops away from the node (R is a parameter for the algorithm). • For sending packets to nodes that are within its routing zone, a node uses the Intrazone Routing Protocol (IARP) which is a proactive protocol which maintains up-to-date routing information to all nodes within the zone. • Neighbor discovery information is used as the basis for IARP. – IARP can be derived from globally proactive link state routing protocols that provide a complete view of network connectivity. – For example, to track the topology of R-hop routing zones, each node periodically broadcasts its link state for a depth of R-hops (controlled by the TTL field in the update message). 49
Routing protocols for MANETs… Hybrid approaches – ZRP… • Routing under ZRP… – Interzone routing: (this part is similar to AODV) • For sending packets to nodes that are not within its zone, a node uses the Interzone routing protocol (IERP), a reactive routing protocol. • For route discovery under IERP, bordercasting is used. • Bordercasting: – Once a source node determines that the destination is not within its zone, it bordercasts a query message to all its peripheral nodes (peripheral nodes of a node are nodes that are at a distance equal to the zone radius R). – The query message, is relayed towards the peripheral nodes using trees constructed within the intrazone topology. – The peripheral nodes in turn bordercast the query towards their own peripheral nodes. This process continues until the query reaches the destination. – When the query is received by the destination, it unicasts a reply message to the source node. 50
Routing protocols for MANETs… Hybrid approaches – ZRP… • Some issues that need to be considered with respect to ZRP – High mobility and/or low call rates favor a smaller zone radius; low mobility and high call rates favor larger zone radius. – Different parts of the network may experience different call arrival rate and different mobility pattern – Moreover, ZRP requires that every node in the network have the same zone radius R. This means, before the network becomes operational, all nodes should come to a consensus on the R. – This motivates the development of a zone routing capability with independent zones capability. 51
Routing protocols for MANETs… Hybrid approaches – ZRP… • IZR (independent zone routing) – a solution that addresses the above issues of ZRP (it assumes that a node knows the position of its neighbors) – Each node has two zones • Routing zone or receive zone of a node: the neighborhood around which a node proactively maintains routing information. This consists of all nodes whose minimum distance in hops is ≤ zone radius R. • Send zone of a node: all nodes that require proactive updates from the node in question. A node is expected to broadcast proactive updates to the members of its zone. 52
Routing protocols for MANETs… Hybrid approaches – ZRP… • Route maintenance for IARP (intrazone routing protocol) for IZR – A node needs to find the sending zone for sending route updates to all nodes to whose receiving zone it belongs to (note that the sending zones need not be regular in shape because the zone radius for each node could be different even if the nodes are uniformly distributed). This is accomplished as follows. • A node broadcasts a zone building packet to all members of its routing zone. The zone building packet includes the zone radius R. The TTL field in the packet is set to R so that it will propagate only to nodes within its routing zone. • Thus, a node may receive zone building packets with different zone radii from different directions. 53
Routing protocols for MANETs… Hybrid approaches – ZRP… • Route maintenance for IARP (intrazone routing protocol) for IZR… • After receiving the zone building packets from all directions, a node determines the maximum of the zone radii received in the zone building packets and uses it as the TTL value in the periodic update messages it sends in the future. – This may cause a proactive update message to propagate to nodes that are not in the send zone of a node because the zone radii are not same for all nodes. – To prevent the propagation of proactive update messages sent by a node A to nodes that are not in its send zone, each peripheral node of A also maintains a list of nodes, called expecting_nodes_list for A, in the send zone of A that lie further away from itself so that they do not forward the update packets if 54 they do not have any such nodes in their list.
Routing protocols for MANETs… Hybrid approaches – ZRP… • Route discovery under IERP (interzone routing protocol) for IZR… – First some definitions: • Rebordercast node: A node closest to the source node on the bordercast path from the source node to a peripheral node, such that its routing zone extends beyond the source node’s routing zone, is called the rebordercast node of the source node corresponding to that peripheral node. • Forwarding node: Nodes lying on the borderscast path between the source node and the rebordercast node. 55
Routing protocols for MANETs… Hybrid approaches – ZRP… • Working of bordercast resolution protocol for IZR: – Step 1: A source node S constructs the bordercast tree to uncovered peripheral nodes – Step 2: S chooses the rebordcast nodes corresponding to each of its uncovered peripheral nodes – Step 3: S then sends the query packet to each of the rebroadcast nodes via the forwarding nodes, if any. – Step 4: When the rebodercast nodes receive the packets, they become the bordercast nodes and start from step 1. 56
Routing protocols for MANETs… Hybrid approaches – ZRP… • Query control under bordercasting in IZR: – A forwarding node marks all the members of its zone as covered. – A rebodercasting node marks: • The nodes lying in the intersection of its zone with the zone of the bordercasting node as covered, if the bordercasting node is a member of its zone. • The nodes lying in the intersection of its zone with the zone of the last forwarding node as covered, if the bordercasting zone does not lie in the zone. The above method of marking covered nodes help in a query being bordercasted by nodes whose routing zones cover newer unexplored zones. 57
Routing protocols for MANETs… Hybrid approaches – ZRP… • Zone radius determination under IZR – For determining the zone radius of each node, they use a hybrid version of Min searching and adaptive traffic estimation used in their JSAC paper 58
Clustering and hierarchical approaches for routing • Motivation for this approach: – IP addresses are hierarchical, in the sense that they identify the location of the end device within the global Internet. – Routing protocols in the Internet take advantage of this hierarchy when determining routes between networks. – Ad hoc networks are not necessarily able to take advantage of such a structure because nodes joining the network have addresses that are random, and hence hierarchical routing based on addresses is not possible. – However, flat routing protocols such as AODV and DSR do not scale. So designing hierarchical protocols that use clustering of nodes can help in designing scalable routing protocols for ad hoc networks. 59
Clustering and hierarchical… • Some characteristics of algorithms for clustering are: – Clusters can be overlapping or totally disjoint – A one level hierarchy of clusters can be created or a recursive multiple hierarchies can be created. – Control within a cluster can be centralized (i. e. , a cluster head controls communication) or fully distributed with no cluster leader – Cluster leaders are typically initialized through some distributed algorithm (some existing leader election algorithms can be finetuned and used in this context) – It is possible that the cluster heads form a routing backbone within the network. • This can be useful if the cluster heads are pre-selected as nodes with greater resources. • Nodes that are located within the boundaries of two overlapping clusters are called gateways. • Gateways serve as routers between the overlapping clusters 60
Clustering and hierarchical… • Some issues related to clustering algorithms are: – When should a new leader be elected for a cluster? – What happens if a node moves outside of a cluster? • Should it elect itself as a leader or should it join another existing cluster – When should a leader be revoked? • When one leader comes within the range of another, one should give up from being a leader – who should give up, etc. 61
Clustering and hierarchical… • Advantages and disadvantages of using clustering approach – They enable hierarchical routing – Hierarchy can by used for implementing hierarchical addresses. • For example, the address of a node z in cluster y is y. z. • If this node is within another cluster x, then its address would be x. y. z. – Routes are recorded between clusters rather than nodes and hence is flexible and can be repaired more easily because there will be more routing options. – Increasing route robustness means increase in route lifetimes, fewer route reconstructions, less control traffic for route repair, etc – Disadvantages are: creating and maintaining clusters require periodic control message overhead • Some approaches taken to reduce overhead: clusters are created on 62 demand, only when needed
Other approaches to routing • Multipath routing – Approaches in this category use route cache to cache and maintain multiple routes to destinations; the routes could be node-disjoint or edge-disjoint or neither – Many of the proposals for multipath routing are on-demand approaches: extension of AODV, DSR, and others • Energy-aware routing – The goal of these protocols is to conserve energy of the nodes by using techniques such as powering down unutilized nodes, load balancing, and dynamic transmission power adjustment. – Some Power-aware MAC protocols use information obtained from the RTS/CTS packets to turn off nodes when they are 63 not receiving packets.
Other approaches to routing… • Security-aware protocols – Secure routing: • During route discovery, the following attacks are possible : a malicious node – Can lie about the existence of a path; the user cannot even be sure of the existence of a path to the destination. – Can modify the information in the routing messages to influence path selection; nodes can collude for an attack. • Many secure routing protocols have been proposed for ad hoc networks 64
Other approaches to routing… • Security-aware protocols… – Intrusion detection: process of detecting malicious nodes and triggering appropriate actions • This problem is somewhat less straightforward because membership in the network is virtually open to anyone. • However, in networks with well defined membership such as military networks, and collaborative networks comprised of a team of individuals, it is possible. • Some approaches proposed for detecting intrusion: – A set of nodes cooperatively detect anomalies and trigger global intrusion detection – Each node promiscuously monitors packets forwarded to its neighbors to check if the neighbors forward the packets to the next-hop 65
• Possible Projects: • Survey of – on-demand routing protocols – geographic routing protocols – hierarchical routing protocols – secure routing protocols – energy-efficient routing protocols 66
0615a12906a97c4a3612f9fbd6d408da.ppt