Congestion Control Mechanisms For Ad-hoc Social Networks

Traditional ad-hoc networks provide an excessive advantage in the scenarios where an infrastructure network is difficult to create (e.g., battle field, disaster environment) and where the users require a rapid access of communication. An Ad-hoc Social Network (ASNET) is a branch of traditional ad-hoc networks, which operates in an infrastructure-less distributed environment using social properties of its users (such as social graph, community, centrality, similarity, social tie and human mobility pattern, etc.) for communication. ASNETs provide an easier communication and efficient utilization of the resources by using the above mentioned social properties of nodes. In the communication processes, ASNETs usually exchange two types of data, i.e., social metadata and application data. In order to search similar interests or share social popularity levels of nodes, there are exchanges of social metadata between existing nodes. Additionally, to transfer files or data, ASNETs use application data exchanging. However, due to the usage of wireless environments, the bandwidth in ASNETs is quite scarce. This leads to the congestion created by social metadata and application data in the network. Furthermore, when ASNET users run multiple social applications in a single source node, some other scenarios can also create congestion loss in ASNETs, these include:(i) mismatch between source and an intermediate node interest, (ii) limited queue capacity of the intermediate nodes, (iii) improper scheduling of data packets, (iv) less link capacity of destination node, and (v) unnecessary acknowledgments.ASNETs, that explore social connectivity among users of mobile devices, are also becoming one of the most important forms of mobile networks. However, increment in number of users due to social networks and a single user running several social applications at the same time paves the way for congestion in ASNETs. Therefore, maximum bandwidth utilization of ASNET in resource scarce environments is one of the major challenging tasks. To overcome the congestion issues, traditional Transmission Control Protocol (TCP) uses the round trip time mechanism for sharing bandwidth resources between users. However, it does not explore the social properties among nodes and it cannot differentiate effectively between various types of packet losses in wireless networks. On the other hand, the usage of various applications and the dynamic nature of the network also create unnecessary congestion, which might degrade the quality of service of an intermediate node dramatically. Therefore, there exist some dropping approaches using drop-tail or Random Early Discard (RED) techniques to drop data packets from the intermediate node queue for circumventing congestion. Furthermore, in order to control congestion in an intermediate node and acquire the most important data initially, a technique such as data scheduling through node-to-node communication can be helpful in ASNETs. Usually, ad-hoc network packet scheduling schemes behave in First-In-First-Out (FIFO) order. Finally, in multi-hop ASNETs, reliable social connectivity and transmission of data for popular nodes are vital with congestion control solution. The reason for unreliability in multi-hop ASNETs is the usage of the same path between data and acknowledgment packets which paves the way for wireless contentions among nodes and thus creates the collision on paths within the network. Unnecessary acknowledgments and sharing of large amount of data with single node creates congestion in multi-hop ASNETs.In ASNETs, lot of users are involved in communicating with one another. The goal of every user/node in an ASNET is to save resources for efficient communication with similar interested nodes. Therefore, in such scenarios, social properties such as similarity, social tie and social graph have the greatest advantages. Furthermore, for efficient utilization of resources, usage of social property centrality has a vital factor. Therefore, sharing of resources based on social-awareness and minimizing the wastage of resources through congestion control scheme are a significant research direction. The consideration of congestion problems is necessary for resolving the wastage of resources in ASNETs so that users can acquire the necessary and exact data packets within time. Due to non-social behavior of traditional ad-hoc networks, existing schemes cannot fulfill the requirements of ASNETs and cannot perform efficiently in congested environments.To improve congestion control in ASNETs, this dissertation focuses on controlling data packets of nodes by integrating the network (e.g., available bandwidth and size of node queue length) and social properties (e.g., similarity, closeness and degree centrality). We addressed four main problems as follows:(a) How can we partition bandwidth among senders when an intermediate node has limited bandwidth, and it only wants to share the resources with those sender nodes that have similar interests?(b) Which dropping method and social property is useful in ASNETs when an intermediate node receives lots of data from senders, and it has less queue capacity in comparison to these data packets?(c) How can we transfer social prioritized data packets from an intermediate node earlier and provide fairness in service? Furthermore, which social property is helpful for efficient employment of resources among nodes and scheduling data packets accordingly?(d) Which method is helpful in solving the reliability issue associated with a congestion control scheme? How can such a method provide full utilization of resources when data packets transfer through multi-hop nodes but the capacity of the end node and intermediate nodes are limited due to hidden/exposed node problem and unnecessary acknowledgment respectively?The first problem of this dissertation is solved by proposing a socially-aware congestion avoidance protocol, namely TIBIAS, which takes advantage of similarity matching among intermediate nodes, to improve the resource efficiency of ASNETs. TIBIAS performs efficient data transfer. During the course of bandwidth resource allocation, it gives high priority for maximally matched interest between different TCP connections on ASNET links. TIBIAS does not require any modification at lower layers or on the receiver nodes. Experimental results show that TIBIAS perform better in comparison to other relevant state-of-the-art protocols, in terms of link utilization, unnecessary reduction of the congestion window, throughput and retransmission ratio.The solution to the second question is a Bio-inspired Packet Dropping (BPD) algorithm for ASNETs. BPD imitates the matching procedure of receptors and epitopes in Immune Systems (IS) to detect congestions. The drop probability settings depend on the selection of data packets, which is based on a node's priority level. BPD selects the most prioritized node through social properties, which is inspired by the B-cell stimulation in IS. In order to fairly prioritize data packets, two social properties are used:1) similarity and 2) closeness centrality between nodes. Extensive simulations are carried out to evaluate and compare BPD to be other existing schemes in terms of mean good put, mean loss rate, throughput, delay, attained bandwidth and overhead ratio. The results show that BPD outperforms these existing schemes.Thirdly, in order to solve the congestion problems in ASNETs, we proposed a user popularity-based packet scheduling method called Pop-aware for congestion control in ASNETs. Pop-aware initially calculates the traffic load of an intermediate node. Secondly, it computes the social popularity of sender nodes using degree centrality (social property) and then prioritizes all incoming flows. Pop-aware also provides fairness of service received for each flow after calculating active service rate. We illustrate the performance of Pop-aware through a series of simulations. In comparison to existing scheduling algorithms, it performs better in terms of control overhead, total overhead, average throughput, packet loss rate, packet delivery rate and average delay.The final focus of this dissertation involves the proposal of a Reliable TCP for Popular Data in Socially-aware Ad-hoc Networks called RTPS. It computes the social popularity of sender nodes using degree centrality (social property). Therefore, it improves the transmission reliability by assigning bandwidth to users based on their popularity levels:extra bandwidth is assigned to those nodes that are more popular and their acknowledgments are sent with higher priority. In addition, RTPS further reduces contentions and packet losses by delaying acknowledgment packet transmissions. The working of RTPS is based on the receiver side that provides solution without gathering explicit information of intermediated nodes. Our detailed investigations demonstrate the excellent performance of RTPS in terms of bandwidth division, throughput, latency and overhead with different hop-distance, number of concurrent TCP flows and variable packet loss rate.