Congestion Control And Resource Allocation In Named-Data Wireless Network

NDN(Named Data Networking)changes the communication pattern of ex-isting IP(Internet Protocol)networks.Because of the explosion of the amount and types of content on the Internet and one of the main behaviors of using the network is retrieving information,NDN's communication pattern is requesting and obtaining information.Interest packets are sent to the network to retrieve Data packets,instead of implementing an end-to-end connection in an IP net-work,so that efficient content distribution is achieved,and redundant content transmission in the network is greatly reduced.This thesis refers to the wireless network built with NDN as Named-Data Wireless Network.Compared with IP,the use of NDN to build a wireless net-work has many advantages:1.The NDN communication process is only related to the content name and has nothing to do with the location of the content,so that the data cache in the network can be conveniently used.Therefore,NDN naturally supports asynchronous data exchange between parties of communica-tion.There is no longer a need to maintain a stable end-to-end communication path between communicating parties,which means that the communication is"connectionless".This is obviously an advantage in wireless networks because the connection is usually intermittent due to node movement or poor signal transmission conditions in wireless networks.2.NDN can easily support node mobility without configuring network parameters(such as IP address,subnet mask,etc.).This is especially beneficial for dynamic environments,where Mo-bile IP-based solutions are not effective.3.Since each Data packet in NDN is identified only by name,regardless of the location of the Data packet and the application/user that uses the Data packet,the Data packet can be cached at any network node,which provides the possibility of the user to use a nearer data cache,thereby reducing the content retrieving delay and the occupation of wire-less link capacity.4.In heterogeneous wireless networks where multiple radio access technologies coexist,if you can combine multiple network interfaces,you can use the advantage of each technology and avoid its disadvantage.In an IP network,an IP address is both a terminal identifier and a location identifier for routing.This makes it difficult to use multiple wireless access technologies simultaneously in an IP network.The NDN communication does not need to establish a connection,and can freely use a variety of network interfaces to forward Interest packets to obtain Data packets.It is necessary to study the congestion control in Named-Data Wireless Net-work for the following reasons:1.Congestion control in NDN is necessary:On the one hand,if the user does not limit the rate of sending Interest packets,then the rate at which the content provider returns Data packets will be too high ac-cordingly;on the other hand,if the nodes in the network do not limit the rate at which Interest packets are forwarded,the rate at which the upstream neighbor nodes forward Data packets will be too high.When the sum of the rate of In-terest packets and Data packets transmitted on a link is higher than its capacity,the link is congested,which may result in packet loss and retransmission,which reduces the user's QoE and network throughput.Therefore,congestion control in NDN is still necessary.2.The traditional congestion control algorithm is not suitable for NDN:NDN users can retrieve Data packets from multiple caches in the network,and the direction of Interest packet forwarding is decided by the network nodes in hop-by-hop fashion,so NDN communication is characterized by multiple-sources and flexible multipath.Due to the above characteristics of NDN,traditional congestion control algorithms cannot guarantee good perfor-mance.3.In the Named-Data Wireless Network,since the wireless resources are valuable and scarce,the capacity of the wireless link is low,so it is more necessary to study the congestion control in Named-Data Wireless Network.In Named-Data Wireless Network,congestion control needs to be jointly designed with wireless resource allocation.On the one hand,congestion con-trol and wireless resource allocation are closely related:Congestion control determines users' demand for links' capacities,and wireless resource alloca-tion determines links' capacities.On the other hand,if congestion control and wireless resource allocation do not work well together,the following two per-formance degradation situations may occur:the network has a high demand for the transmission rate of a certain link,but the resource allocation does not provide enough capacity,so the network throughput is limited;or the network has a low demand for the transmission rate of a link,but the resources allocat-ed to this link provide too much capacity,which reduces resource utilization efficiency.Therefore,this thesis mainly studies congestion control and wireless re-source allocation in Named-Data Wireless Network,which includes:how to jointly optimize congestion control and link transmission power allocation in Named-Data Multihop Wireless Network;how to achieve jointly optimal n-ode scheduling and congestion control in Named-Data Wireless Network with broadcast forwarding;how to jointly design congestion control and heteroge-neous wireless resource allocation when NDN is applied to Heterogeneous Wireless Networks.Specific research content and major innovations include the following:1.A jointly optimized congestion control and power control algorithmNDN is a network protocol that is very suitable for building multihop wire-less network.This thesis refers to multihop wireless network built with NDN as Named-Data Multihop Wireless Network.In Named-Data Multihop Wire-less Network,congestion control determines users' need for links' capacities in the network,and transmission power control affects capacities that links can provide.Therefore,congestion control and power control are closely related,they jointly affect the performance of the network.Therefore,this thesis stud-ies the joint optimization of congestion control and power control,and uses NUM(Network Utility Maximization)idea to formulate the problem as an op-timization problem.The objective of the formulated optimization problem is to increase the utility function that is positively related to users' Data packets receiving rates,and the constraints related to the communication characteristics of NDN are taken into account.Then,an iterative algorithm is proposed based on the subgradient algorithm,and it is proved that the proposed algorithm can converge to the optimal solution of the formulated problem.One can control the rate at which the user sends Interest packets,the rate at which the node forwards Interest packets/Data packets,and the transmission power of the link according to the result obtained by the algorithm.Our proposed algorithm re-alizes the cooperation between congestion control and power control,and can obtain greater network throughput and better fairness.2.A jointly optimal node scheduling and congestion control algorithmIn Named-Data Multihop Wireless Network,according to many research results,the use of "broadcast forwarding" can improve the success rate of con-tent discovery and the efficiency of resource utilization.Therefore,this thesis further studies Named-Data Multihop Wireless Network using broadcast for-warding.In this kind of network,node scheduling selects nodes to activate and forward packets in each time slot,so determines the long-term average trans-mission rate of each node,and congestion control determines the users' need for transmission rate of each node,so node scheduling is closely linked with congestion control and they jointly affects the performance of the network.Therefore,this thesis regards joint optimization of node scheduling and con-gestion control as a research focus,and uses the idea of network utility max-imization to model the joint optimization of node scheduling and congestion control as an optimization problem.In this paper,an iterative approximate sub-gradient algorithm is proposed to solve the formulated problem.By introducing virtual queues and using Lyapunov optimization method,the difficulty due to the NDN flow balance rule is solved.This paper also proves that the proposed algorithm can converge to the optimal solution of the formulated problem.The user's Interest packets sending rate,the rate at which the node forwards Interest packets/Data packets,and node scheduling can be controlled according to the proposed algorithm.Our proposed algorithm uses the cooperation between n-ode scheduling and congestion control to increase the network throughput,and can improve the fairness among users.3.Jointly designed hop-by-hop congestion control and heterogeneous wireless resource allocation algorithmThe use of NDN to build heterogeneous wireless networks can help users flexibly utilize multiple access networks.This thesis refers to heterogeneous wireless networks built with NDN as Named-Data Heterogeneous Wireless Networks.In Named-Data Heterogeneous Wireless Networks,wireless re-source allocation determines the link capacity between the user and the access point.Hop-by-hop congestion control determines the user's demand for link capacity(between the access point).Only by jointly designing wireless re-source allocation and hop-by-hop congestion control can the wireless resource utilization efficiency and throughput of the network be improved.In this the-sis,the concept of network capacity region is firstly introduced to analyze the upper bound of network throughput of Named-Data Heterogeneous Wireless Networks,and the network capacity region is represented by four kinds of in-equality constraints.Then virtual queues are introduced,the problem of satis-fying the inequality constraints is transformed into the stability of the virtual queues,and the jointly designed hop-by-hop congestion control and heteroge-neous wireless resource allocation algorithm is proposed by using the method of minimizing Lyapunov drift.We proved that the proposed algorithm can achieve maximum network throughput.The proposed algorithm can provide greater network throughput,faster content download speed,and better fairness.