Critical Analysis on Route Reservation in Ad Hoc Networks - Research Paper Example

Contents Contents 2 Introduction 3 The scalable resource reservation scheme 5 The basic structure of SRRS 6 Adaptable reservation 7 Route Reservation9 Unsuccessful Reservation 10 Reservation based ad hoc networks. 11 Working of Reservation-Based (RB) Switching ad hoc networks 12 References 13 Introduction Internet traffic has been growing exponentially for the last few years due to the rapid increase in user popula­tion and the proliferation of new applications. In the near future, the traffic volume is expected to become unprece­dentedly high, fueled by the extraordinarily large number of future mobile Internet handsets, the bandwidth require­ments of emerging multimedia applications, and the avail-ability of high-speed home access networks. Among vari­ous advances in networking technologies, wavelength divi­sion multiplexing (WDM) has the capability to provide very high-bandwidth transmissions and is thus expected to dom­inate the future Internet core. In addition to providing high bandwidth to network appli­cations, another important emerging requirement for the fu­ture Internet is the provision of guaranteed quality of service (QoS) for real-time multimedia applications. The resource reservation protocol (RSVP) is a signaling protocol that reserves the required resources for time-critical applications so that they can be served with a guaranteed minimum band-width and bounded delay and jitter, in order to meet the re­quired QoS. Various extensions to RSVP are currently being standardized, while many other reservation protocols have also been proposed. In most of the previous reservation protocols, a connection has to be fully estab­lished before the transmission of any data packet can start. A problem with such protocols in future broadband networks is that the round-trip or one-way propagation delay required for establishing a connection is non-negligible compared to the time required to transmit the data, which leads to low link utilization since the required capacity is reserved long be-fore it is actually used, considerably affecting the maximum achievable throughput (Fig. 1a). Another problem with such protocols is that the time required to set up the connec­tion considerably increases the latency, which is undesirable and may be unacceptable to some applications, especially those involving latency-sensitive bursty traffic. Figure 1. Comparison of SRRS and ordinary reserva­tion schemes (ORS). The light shaded parts represent time segments during which capacity is reserved but not locked, while the dark shaded parts represent time segments during which capacity is both reserved and locked. (a) Ordinary reservation schemes. (b) SRRS with small äS to reduce la­tency. (c) SRRS with a medium value for äS to provide a tradeoff between latency and the probability that a session is not successfully established in time (in which case data packets may have to be dropped and the throughput may de-crease). (d) SRRS with larger äS to increase throughput for latency-insensitive sessions or latency-sensitive messages with aggressive reservation. When conservative transmis­sion is used, the acknowledgement has to be received at the source node before the transmission of any data packet can begin so that no packets are dropped due to buffer overflow. Note that in cases (b) and (c) the latency is smaller than in (d), but packets may be dropped unless an appropriate flow control mechanism, such as credit-first flow control, is used.[Source:- ieee.com] . In this document I present the scalable resource reser­vation scheme (SRRS) and the constraint-based deflection routing with backtracking and bifurcation (CD/BB) for QoS management over the next-generation Internet. SRRS and CD/BB routing can be viewed as an evolution of our pre­vious protocols, including the conflict-sense routing (CSR) protocol for multiprocessors, the ready-to-go virtual circuit (RGVC) protocol and the efficient reservation virtual circuit (ERVC) protocol for the Thunder and Lightening ATM network testbed, the virtual circuit deflec­tion (VCD) protocol to be used in the MOST all-optical Tera Switch under development at UCSB. The objective of this work is to develop scalable, flexible, and aggres­sive reservation mechanisms and routing algorithms that will enable the next-generation Internet (with multiproto­col label switching (MPLS)) to efficiently support ex­isting and emerging multimedia and data applications with strict QoS requirements. SRRS has three categories of fea­tures: (1) the aggressive/optimistic/conservative transmis­sion disciplines, (2) timed reservation and activated lock­ing, and (3) adaptable reservation. The timed reservation feature and the aggressive transmission discipline of SRRS can solve the aforementioned problems and guarantee very small call-blocking rate and packet-loss ratio without rely­ing on buffering. Constraint-based routing and the resource reservation mechanisms also facilitate traffic engineering in the Internet. The features of SRRS and CD/BB routing are tightly integrated together to efficiently support each other in order to achieve the objective. The scalable resource reservation scheme In this part I put forward the scalable resource reser­vation scheme (SRRS) for efficient signaling and resource reservation. In the next section, we will then propose routing algorithms that are particularly suitable to be used in com­bination with SRRS. The basic structure of SRRS SRRS consists of seven phases. These phases are “pipelined” in the sense that a latter phase can start at a node before a former phase is completed in the network. Phase 1 (Signaling Phase): In order to establish a con­nection (i.e., a route with reserved capacity) for the transmission of data packets, the source node S (for source-initiated reservation) sends a setup packet for signaling. Phase 2 (Reservation Phase): Node i on the estab­lished part of a connection reserves the capacity re­quired by the connection at reservation starting time ti (according to its local clock). Phase 3 (Locking Phase): Node i on the established part of a connection locks the reserved capacity for tr he connection. Phase 4 (Transmission Phase): The source node S starts transmission at transmission starting time tS (of its local clock) or when it receives an acknowledge­ment for binding (e.g., in MPLS networks) or the establishment of the connection. Node i forwards the data packets after they are received. If the data packets catch up with the setup packet or ar­rive at node i before the reserved capacity is available, node i per-forms one of the following actions: (1) dropping, (2) buffering, or (3) deflection. Phase 5 (Renegotiation/Rerouting Phase): If a ses­sion’s transmission rate or duration changes, or if the session has to be rerouted to accommodate other con­nections or to increase its transmission bandwidth, the source node or an intermediate node sends a control packet to adjust the resources and/or the path occupied by the connection. Phase 6 (Release Phase): Node i on the connection re-lease the reserved, allocated, and/or locked resources. Go to Phase 2 if there are reservation subintervals re­maining for the session. Phase 7 (Tear-down Phase): Node i on the connec­tion tears down the connection. More details concerning these phases and several scal­able reservation mechanisms can be found in. In the following subsections, we present several advanced features that are particularly suitable for inclusion in SRRS. Most of them are extensible features, which are advanced features that do not have to be implemented at all network nodes in order to work correctly and efficiently. Adaptable reservation The reservation made by an SRRS connection may con­tain more than one set of QoS requirements, one for the nor­mal mode and others for degraded modes. A node may then adapt to the traffic conditions and provide acceptable service quality to all the connections maximizing user satisfaction and minimizing aggregate penalty. This feature will be re­ferred to as adaptable reservation. When adaptable reservation is used for a connection, its setup packet is required to carry with it the resource require­ments under the degraded modes, and the session has to be able to cope with such a degraded bandwidth enforced by the network when there are no sufficient network resources. Network nodes (including base stations) with such an ad­vanced feature have to record the requirements of each con­nection or forwarding equivalence class (FEC), including the maximum tolerable delay and delay variation, the peak, average, and/or minimum bandwidths without degradation, and the bandwidth requirements under degradation level 1, level 2, and so on, along with the priority of the session/FEC and the penalty for each degradation level. When there is sufficient bandwidth available, a network node allocates the required bandwidths without degrada­tion to all the connections as ordinary communication pro­tocols; when there is not sufficient bandwidth left, a net-work node first holds a certain amount of traffic that is time-noncritical. If this is still insufficient to accommodate the traffic for new connections, handoffs, and/or reactivat­ing/renegotiating connections, the network node reallocates the required bandwidth eunder degradation to some connec­tions with lower priority. The network node should try to provide the required bandwidths without degradation to high-priority connec­tions if possible, while reducing the bandwidths assigned to some (or even all) of the connections when necessary. A goal of the bandwidth reallocation algorithm is to minimize the aggregate penalty. Note that the possibility of such a degradation is within the agreement made between the net-work node and the sessions when they established the con­nections or when they renegotiated for new bandwidth. De-grading some services while still at a tolerable level can achieve higher satisfaction to all users as a whole, as com­pared to degrading, for example, real-time multimedia ses­sions without delay guarantees as in current best-effort net-works. Also, it helps prevent resource wasting when the traffic is heavy and the resources are most needed. In some networks it may be advantageous for an appli­cation to specify the circumstances for it to use a reduced bandwidth even when there is still bandwidth available in a wireless cell or at a network node. For example, a future mo­bile telephone company may offer lower price per Kbps in lightly-loaded cells, while asking for higher price in heavily-loaded cells. A user may then specify the ideal and/or de-graded bandwidths for each of the price ranges. Other cri­teria such as power consumption may also be taken into ac-count for determining the bandwidth to be used. We refer to this type of techniques as economic reservation. We have incorporated the adaptable reservation feature in a medium access control (MAC) protocol proposed in. The resultant MAC protocol can provision an accept-able QoS even when there are many handoffs and/or new mobile stations arrive at a cell and suddenly increase the ag­gregate bandwidth requirement considerably. We have also proposed an architecture for implementing the adaptation feature at the application layer for network/distributed ap­plications. Route Reservation When the destination node receives one RREQ packet from the source node, it returns a route rely (RREP) packet by unicasting back to the source following the route recorded in the route_list. Our pro­tocol uses symmetric links between neighboring nodes. It does not attempt to follow paths between nodes when one of the nodes cannot hear the other one; however we may include the use of such links in future enhancements. As a RREQ travels from the source to the des­tination, it automatically sets up the reverse path from the destination back to the source. To set up a reverse path, a node records the address of the neighbor from which it received the copy of the RREQ. From the RREQ packets, we can obtain the state of the data slots. According to the information recorded within the RREQ, the destination can set up a QoS route and reserves resources (slots) hop­-by-hop backward to the source. Using the source routing algorithm, we copy the fields from RREQ to RREP. As the RREP traverses reverse to the basis, each node all along the path reserves those free slots which were calculated in advance. When the source receives a RREP, the end-to-end bandwidth reservation is successful, and the virtual circuit (VC) is established. Then, the source node can begin transmitting datagram. Unsuccessful Reservation When the RREP travels back to the source, the reservation operation may not be successful. This can result from the fact that the slots which we want to reserve are occupied by another VC or the path breaks. If this is the case, we must give up the route. The interrupted node sends a NACK (i.e., RESERVE_FAIL_NACK) back to the des­tination, and the destination re-starts the reservation process again along the next feasible path (note that in the route discovery process, each RREQ which arrives the destination piggybacks a feasible QoS route). If there is no VC can be setup along all feasible QoS routes, the destination broadcasts another NACK (i.e., NO_ROUTE) to notify the source. Upon receiving NO_ROUTE, the source can re‑start the discovery process if it still requires a route to the destina­tion. If there is no any response back to the source before the timeout occurs, the source will also re-perform the route discover operation. Once a VC is established, the source can begin sending data-grams in the data phase. At the end of the session, all reserved slots must be released. These free slots will be contended by all new con­nections. However, if the last packet is lost, we will not know when the reserved slots should be released. This issue will be discussed in the next sub-section. Reservation based ad hoc networks. Plenty of the reservation-based QoS routing protocols have been planned before. However, the capable Bandwidth calculation complexity and the bandwidth condition complexity have not been address grimly. An algorithm to direct the purpose to decide the Route that is the majority likely to please the QoS obligation and an algorithm to reserve the good time slit and thus keeps additional free time period for other needs. Simulation outcomes demonstrate that our protocol can attain elevated route establishment probability and low packet loss rate. The design of well-organized direction-finding protocols is a decisive issue for all types of networks. Compared with the traditional wired network, the mobile ad hoc network (MANET) has no fixed topology. Therefore, the source-initiated on demand routing protocol, which establishes the route Between the source and destination only when the source demands that, becomes the most popular routing protocol in the MANET. However, these on-demand routing protocols Use the best-effort approach to transmit message and can not guarantee the quality of the transmission. As the bandwidth of the wireless channel increased, multimedia services can be provided in the wireless network. These services need the assurance of a convinced bandwidth or a surrounded delay, or else, the excellence of these services will be intolerable. Therefore, quality of check (QoS) becomes a significant matter in the MANET and the QoS routing is the majority significant subject. In the MANET, since the radio signals inside two hops may hinder with each other, canal corollary is one of the best move toward to avoid intrusion. It is shown in that the best technique to guarantee QoS is attained only with appropriate resource reservation. Therefore, several works provide QoS guaranteed transmissions by reserving resources for the Time Division Multiple Access (TDMA) based MANETs. These protocols can keep away from contention and collision and therefore are more well-organized than the other QoS protocols, but they necessitate the MANET to be synchronized. The main disadvantage of these protocols is that they do not give a clear guide to preserve correct time slot and thus lower downward the route organization Probability. To improve previous work, we propose a novel ondemand bandwidth reservation QoS routing protocol for TDMA based MANETs. The aim is to study a even route, which satisfy the bandwidth obligation. The main aid of this paper are: first, we suggest an algorithm to keep the correct time slot and thus raises the route organization likelihood. Second, we suggest an algorithm to guide the purpose host to decide the route that is the majority likely to make happy the QoS obligation. Our QoS direction-finding protocol consists of two phases: The route detection and bandwidth condition phases. In the route detection phase, we suggest a novel algorithm to work out the obtainable bandwidth of the likely QoS routes, so that the purpose host can want the way that is nearly everyone likely to please the QoS obligation. After the QoS route has been select, the bandwidth condition stage follows. In the bandwidth reservation stage, i suggest a new algorithm to work out the mass of each accessible time slot. The hosts in the QoS route will set sideways the time opening with the lowly heaviness, which will accumulation the smallest numeral of free time slots of contiguous hosts, and thus raise the route association likelihood. Simulation results show that our procedure can attain much higher route organization probability and lower packet loss rate than those of the Forwarding Algorithm(FA). Circuit switching provide restricted admission to the resources by resources of reservation. Working of Reservation-Based (RB) Switching ad hoc networks The standard of operation of an RB scheme is quite easy. Prior to data broadcast, a source node reserves a multi hop route to the target through a route detection phase. We suppose that route detection messages are sent on a detach control channel. Once an middle node agrees to communicate traffic for a exacting source in the network, it cannot start a session or communicate messages for any supplementary source until the on-going session is over. The resource node releases the route following the session ends. We highlight that this reservation pertains to node dispensation but not to the communal ordinary radio channel. In other words, the middle nodes offer their dealing out time only to the source which reserved the way; but, reservation of a multi hop route does not provide any node a restricted right of entry to the common radio channel (in terms of frequency bands, time slots, or spreading codes). Fig. below explains an example of reserved routes in a network where an RB system is used. To arrange to References 1. Ali, I.A., H.T. Mouftah, and A.H. El-Sawi, “A dynamic rout­ing protocol for broadband networks,” IEEE Symp. Comput­ers and Communications, Jul. 1997, pp. 495-500. Read More