One of the key challenges facing network administrators in securing an enterprise network is the anonymity of the traffic on the network. Although current research has taken steps forward addressing the issue of identifying the application layer protocols (e.g., SSH, HTTP, or FTP) a more fine-grained identification is required for a variety of applications that run over these established application layer protocols. We are specifically interested in disambiguating traffic that is carried by the HTTP application layer protocol. In this paper, we investigate representatives of classes of applications, namely social networking (Facebook), web-mail (Gmail), and streaming video applications (YouTube), all of which communicate via the HTTP protocol. We use specific features derived from network traffic (i.e., the TCP/IP packet headers) that can be used to classify the flows as belonging to each application. An important aspect of our work is to classify the applications based on any segment of the traffic flow. We consider different signals that can be derived from the network flow such as the packet sizes and inter-arrival times and apply simple statistical and spectral analysis to identify distinguishing features of the applications. Our classification system yields a classification rate of 93% or better using only packet size statistics. We evaluate our system on network flows collected from the backbone of the UC Davis campus network. Furthermore, we consider two types of noise an adversary may inject to evade detection: packet padding and altering the inter-packet delays. Despite these two types of noise, using our classification method we are still able to achieve a reasonable classification rate.

Internet users and their emerging applications require high-data-rate access networks with mobility and untethered convenience. This is creating an increasing demand for wireless and optical access networks—to bring the high capacity of fiber and untethered access of wireless closer to the user. Hence, in this dissertation we explore novel architectures and algorithms for Wireless-Optical Broadband Access Network (WOBAN), an access network with the cost-effectiveness of a wireless access network and high performance of an optical backhaul network. The main topics of the dissertation are the following. (1) We propose a capacity- and delay-aware routing algorithm, called CaDAR, for WOBAN. This algorithm addresses two major challenges in WOBAN: (i) it efficiently utilizes the finite radio capacity at each wireless node by intelligently allocating the capacity on each outgoing wireless link of a WOBAN, and (ii) it performs shortest-delay routing across both wireless front-end and optical backhaul. (2) To improve the performance of a WOBAN, it is essential to enhance the capacity for wireless access using a low-cost solution to match the high capacity offered by the optical backhaul. We design a mixed-capacity wireless access (MCWA) architecture for WOBAN that enhances the capacity of a few wireless nodes using multiple radios to improve the overall performance of WOBAN. We also develop an intelligent channel and radio assignment (ICRA) for WOBAN that utilizes the MCWA architecture to distribute the network load over different radios and orthogonal wireless channels to reduce interference and contention in the wireless front-end. (3) Access networks are increasingly shaped by the services they provide to the users. We design a service-oriented network architecture that integrates a cloud with WOBAN, called Cloud-Integrated WOBAN (CIW). It provides cloud components in the wireless front-end of a WOBAN to handle some cloud requests locally. By bringing the cloud services closer to users, CIW offloads traffic from the wireless backhaul and improves the performance of a WOBAN. We also present a novel energy-saving routing mechanism, called Green Routing for CIW (GRC), that allows CIW to self-manage the activation of network components to minimize the overall energy consumption of CIW. (4) We also apply our knowledge obtained from the studies on the improvement of wireless backhaul in WOBAN for the enhancement of 3G and 4G backhaul. We discuss the performance bottleneck of 3G and 4G backhaul imposed by legacy copper lines and design a wireless overlay network (WON) using high-capacity point-to-point wireless links that increases the capacity of 3G and 4G backhaul, reduces the performance bottleneck, utilizes the bandwidth of deployed fibers more efficiently, and provisions backup paths during failures.

Wireless networks provide flexible and ubiquitous access to the telephone networks or the Internet. Multiple technologies have been developed to provide wireless access including cellular/3G/LTE, WiFi, and WiMAX. The cellular network is the most popular network, which provides the stable and constant service under most types of mobility. Its future version, LTE Advanced, is the most advanced wireless access technology, and supports the current or future bandwidth-sensitive and/or delay-sensitive applications, such as Voice over IP and real-time video streaming. The WiFi network has been aggressively deployed in many areas and provide access to laptops, PDAs and smartphones. These are referred to as WiFi “hotspots”. The basic WiFi infrastructure usually offers flexible and easy to deploy wireless access inside a small area at a low cost. The WiMAX network has its own advantages to provide higher transmission speed for a point-to-point communication over longer transmission distances. In many areas, these wireless networks co-exist, overlap and interlace with one another to create a heterogeneous wireless network. Instead of functioning independently, significant benefits can be accrued through cooperation and coordination among these networks by leveraging their unique advantages. This is possible as user devices come with multiple network interfaces to connect to each of these networks. In order to exploit the advantages of these heterogeneous networks, it is important to implement an efficient resource allocation algorithm to coordinate the resources of multiple wireless networks and also have a good design of the heterogeneous network. The dissertation makes contributions in both the above areas.In current WiFi networks, the overall spectrum is divided among multiple overlapping channels. The adjacent access points need to operate on orthogonal channels to avoid the interference. In Chapter II, we give a precise analysis of the interference among different channels and discuss the potential possibility of utilizing partially over-lapping channels in the multi-hop mesh networks. It is possible to optimize different network performance metric, such as throughput, by balancing parallel transmissions and partially received transmission power offers the best system performance, such as throughput.In the multi-hop mesh networks, some mesh nodes are highly congested either due to the interferences from multiple neighbors, or when they are located at the intersection of multiple routing paths. These highly congested mesh nodes significantly degrade the network performance, since the throughput of a path is limited by the node with minimum capacity. In Chapter III, we discuss the efficient cooperation between aWiFibased mesh network and a WiMAX network to mitigate the impact of congested nodes. The WiMAX network, with its longer transmission range, can be leveraged to bypass the traffic from some highly congested mesh nodes. The load balancing makes the throughput of the heterogeneous network higher than the sum of the throughput from the WiFi and WiMAX networks if they operate independently.In the widely deployed wireless networks, the infrastructure mode is used, where customer devices only communicate with the base stations. This mode is suitable for normal Internet access. However, some emerging applications, such as P2P file sharing, teleconferencing, network games, require frequent communications among terminals that may be in the coverage of the same base station. The exiting infrastructure mode of the network architecture results in high resource waste due to the unnecessary transmissions via the base station even when both the user could directly communicate. In Chapter IV, we propose a novel network architecture, Local-Interest-Group (LIG), in which all nodes can communicate in any ways according to the application requirements. The real-time algorithm and protocol minimize the interferences among co-existing LIGs and maximize the bandwidth utilization, which greatly improves the overall system performance under multiple performance metrics.In network planning, it is difficult to efficiently locate base stations due to the inaccuracies in the prediction of the traffic density. The movement of traffic to different parts in the city during different times of a day makes fixed base stations either operate at very low load or become highly congested at different time periods. Fortunately, the detailed analysis based on network measurement shows that the movement of traffic density is predicable. In Chapter V, we propose a new network component, Traffic-Tracing Gateway (TTG), which works as the base station but traces the traffic movement taking the advantage of the heterogeneous wireless networks. By following the optimal trajectories, TTGs cover the maximum traffic and provide much better system performance in both single-hop or multi-hop networks.This dissertation proposes efficient resource allocation methods in heterogenous wireless access network with partially overlapping channels and the cooperation between WiFi and WiMAX networks. In the dissertation, we also propose the novel network designs based on local-interest-groups and traffic-tracing gateways, to augment existing wireless access networks and making them more resource efficient while providing higher end-to-end performance.

Covert timing channels aim at transmitting hidden messages by controlling the time between transmissions of consecutive payload packets in overt network communication. Previous results used encoding mechanisms that are either easy to detect with statistical analysis, thus spoiling the purpose of a covert channel, and/or are highly sensitive to channel noise, rendering them useless in practice. In this paper, we introduce a novel covert timing channel which allows to balance undetectability and robustness: i) the encoded message is modulated in the inter-packet delay of the underlying overt communication channel such that the statistical properties of regular traffic can be closely approximated and ii) the underlying encoding employs spreading techniques to provide robustness. We experimentally validate the effectiveness of our approach by establishing covert channels over on-line gaming traffic. The experimental results show that our covert timing channel can achieve strong robustness and undetectability, by varying the data transmission rate.

While Internet Telephony (IP Telephony) encompasses many different architectures and services, the key idea is the transport of real-time voice traffic over the Internet. IP Telephony architecture [2] allows the entire end-to-end path or a portion thereof to be routed over the Internet. The endpoints are regular personal computers (PCs) that are equipped with IP Telephony software which allows them to interface with the existing Public Switched Telephone Network (PSTN) through an Internet Telephony Gateway (ITG). The only delay suffered by the voice traffic in the PSTN network is the propagation delay which is fixed once the circuit has been established. On the other hand, the Internet is still inherently a best-effort network and provides no end-to-end bandwidth guarantees. Thus, transporting packetized voice over the Internet can result not only in variable delays but also losses which may cause poor audio quality at the receiver. Furthermore, these losses and delays can increase as the number of hops traversed by the voice packets increase. As a result, the quality of the received audio can degrade as the distance traversed by voice packets over the Internet increases.

Modern society is becoming increasingly dependent on high-speed communication networks for instant access to information. People now make banking transactions, monitor news and weather, play interactive games, listen to music, and watch real-time video and sports over the Internet. High-speed broadband access is required to facilitate many of the above applications. Similarly, scientific experiments have now become more dependent on high-speed networks as they move towards the paradigm of distributed computing. As an example, large data sets such as DNA information of micro-organisms may be transferred from remote data warehouses to a computation facility over a backbone grid network (lambda grid) where the data is processed, visualized, and delivered to the scientist. This dissertation investigates how emerging network technologies, namely optical broadband access networks and lambda grids, coupled with efficient protocols at the edge of the network---the end-system---may deliver such emerging applications to the respective users.The dissertation begins by introducing relevant and emerging high-speed network technologies, namely the Passive Optical Network (PON) and lambda grids in Chapter 1. In Chapter 2, the problem of aggregating files from distributed databases at a (grid) computing node over a lambda grid is considered. The challenge is (i) to identify routes (i.e., circuits) in the lambda-grid network along which files should be transmitted and (ii) to schedule the transfers of these files over their respective circuits.In Chapter 3, the broadband access network is considered. We study the challenge of enabling open access in a broadband access network. Open access implies the ability of multiple service providers to share the access network infrastructure to make services available to the end users. This requires fairness in terms of throughput, delay, jitter, and other network parameters in the access channel among the sharing entities, namely service providers and end users. The idea of having two different Service-Level Agreements (SLAs) called Dual SLAs to meet the above fairness requirements is introduced and demonstrated in this chapter.Traditionally, the network speed has been slow compared to the processor speed and therefore the network has generally been the bottleneck for transferring data. However, high-speed networks such as lambda grids and PONs can offer line rates of several Gigabits per second. In these circumstances, the congestion moves to the edge of the network---namely the end-system. The ability of a system to receive data from the network and deliver it to the application may often be the bottleneck for seamless data transfer. Therefore, performance matching the network performance with the end-system performance is an important research challenge.In this context a lightweight, end-system performance aware rate-adaptive transport protocol called RAPID, is introduced in Chapter 4. RAPID attempts to maximize the end-to-end throughput while minimizing packet loss over a lambda grid. In Chapter 5, an analytical model of the end-system based on a Stochastic Reward Net (SRN) model is investigated. The objective is to estimate the optimal rate for the best data transfer. The dissertation is concluded in Chapter 6 with a discussion of important open research problems.This dissertation makes important contributions by introducing and investigating novel architectures, algorithms, protocols and systems knowledge that will help meet the demands of next-generation applications.

While Internet Telephony (IP Telephony) encompasses many different architectures and services, the key idea is the transport of real-time voice traffic over the Internet. IP Telephony architecture [2] allows the entire end-to-end path or a portion thereof to be routed over the Internet. The endpoints are regular personal computers (PCs) that are equipped with IP Telephony software which allows them to interface with the existing Public Switched Telephone Network (PSTN) through an Internet Telephony Gateway (ITG). The only delay suffered by the voice traffic in the PSTN network is the propagation delay which is fixed once the circuit has been established. On the other hand, the Internet is still inherently a best-effort network and provides no end-to-end bandwidth guarantees. Thus, transporting packetized voice over the Internet can result not only in variable delays but also losses which may cause poor audio quality at the receiver. Furthermore, these losses and delays can increase as the number of hops traversed by the voice packets increase. As a result, the quality of the received audio can degrade as the distance traversed by voice packets over the Internet increases.

Unlike the bandwidth in a wired network, radio spectrum is a significantly scarce resource. This resource needs to be efficiently shared and reused by a number of users who may be simultaneously accessing a variety of mobile services. Thus, careful planning and management of the radio spectrum is required to maximize its value across all users and all accessed services. The evolving need for supporting differentiated services for novel multimedia applications in the wireless network adds a new dimension to this complex problem. Network researchers and architects are investigating the use of pricing to efficiently handle many of the corresponding network design and management issues. In this paper, we make two important contributions on this topic. First, we review the existing body of literature that attempts to utilize pricing in the design of next-generation wireless networks. Specifically, we classify the existing works into three categories: (a) pricing-based resource provisioning; (b) pricing-based static planning; and (c) pricing-based adaptive resource management. Second, we propose a unified pricing scheme which attempts to encompass the various design issues into a single comprehensive framework that can potentially lead to a scalable, differentiated-services wireless network architecture for the future.

A crucial aspect of Peer-to-Peer (P2P) systems is that of providing incentives for users to contribute their resources to the system. Without such incentives, empirical data show that a majority of the participants act as free riders. As a result, a substantial amount of resource goes untapped, and, frequently, P2P systems devolve into client–server systems with attendant issues of performance under high load. We propose to address the free rider problem by introducing the notion of a P2P contract. In it, peers are made aware of the benefits they receive from the system as a function of their contributions. In this paper, we first describe a utility-based framework to determine the components of the contract and formulate the associated resource allocation problem. We consider the resource allocation problem for a flash crowd scenario and show how the contract mechanism implemented using a centralized server can be used to quickly create pseudoservers that can serve out the requests. We then study a decentralized implementation of the P2P contract scheme in which each node implements the contract based on local demand. We show that in such a system, other than contributing storage and bandwidth to serve out requests, it is also important that peer nodes function as application-level routers to connect pools of available pseudoservers. We study the performance of the distributed implementation with respect to the various parameters including the terms of the contract and the triggers to create pseudoservers and routers.

Future broadband wireless access systems plan to inte grate various classes of MTs Mobile Terminals each class with a di erent type of quality of service QoS requirement When the load on a wireless network is high the guarantee of QoS for each class of MTs is a challenging task This study considers two classes of MT pro led MTs and non pro led or regular MTs It is assumed that pro led users require a guaranteed QoS The measure of QoS is the probability of forced termination of a call that was allowed to access the network Two previous hando prioritization schemes i pre request scheme and ii guard channel scheme decrease hando failure and hence forced termination In this work we compare and contrast both the schemes through exten sive simulation it is found that neither method can guar antee a desired level of QoS for the pro led MTs We then propose a novel Call Admission Control Algorithm CACA that can maintain any desired level of QoS while successful call completion rate is very high In the proposed algorithm the new call arrival rate is estimated continuously and when the estimated arrival rate is higher than a predetermined level some new calls are blocked irrespective of the avail ability of channels The objective of this pre blocking of calls is to maintain the wireless network system s observed new call arrival rate at no more than a predetermined rate We show that the proposed method can guarantee any desired level of QoS for pro led users

Layered video is a video-compression technique to encode video data in multiple layers. It typically consists of a base layer and some additional layers that provide enhanced video quality. The multicasting operation of layered video consists of many receivers dynamically joining and leaving different multicast sessions of different layers depending on their network condition. A layered video multicasting system needs to satisfy: (i) bounded end-to-end delay from the video source to each receiver; (ii) minimum total cost; and (iii) minimum delay jitter between the various video streams received by each receiver. The problem of computing such data distribution paths is NP-complete. This paper presents a new heuristic algorithm, called layered video multicast super-tree routing algorithm, with O(Rn2) time complexity and O(R2) message complexity, where n is the number of nodes in the network and R is the receiver group size. Our investigation shows that the multicast data paths computed by our algorithm can always satisfy the delay constraint with reasonably low total cost.

In this paper we study two major sources of packets delay in the GSM/GPRS wireless network, at Base Station and at GGSN node. Fisrt, for the former one, we present an analytical model to study the performance of channel sharing schemes to support both circuit switched voice and packet data services in a GSM/GPRS network. We study three channel sharing schemes: 1) fixed sharing in which cell channels are statically partitioned into two sets one for voice calls and the other for data traffic; 2) partial sharing in which ndatachannels are reserved for data while the remaining N i ndatachannels are shared by voice and data with preemptive priority for voice calls; and 3) complete sharing in which all the channels shared by voice and data with preemptive priority to voice calls. We investigate several key issues such as call blocking rate and mean packet delay for different cell loads with the data source modeled by a Markov Modulated Poisson Process (MMPP). We validate the mathematical model through simulations and quantify the impact of the data source model and the voice call load on the mean packet delay for different channel sharing schemes. Secondly, for another souce causing delay, we present the analytical model to quantify the benefit of replicated GGSN with load balancing architecture.Our results show that replicated GGSN architecture with load balancing policy can significantly reduce the packet delay even when the total GGSN capacity remains constant.

The two key resources in an IP telephony network are the Internet telephony gateways (ITGs) and the IP network. These resources must be effectively managed to simultaneously provide good QoS to calls and maximize network resource utilization. This paper presents two main contributions. First, we design a call admission policy based on congestion sensitive pricing. As the load increases, this policy preferentially admits users who place a higher value on making a call while simultaneously maintaining a high utilization of network resources. We derive the function mapping congestion to price for the admission policy that maximizes revenue. Second, we design a call redirection policy to select the best ITG to serve the call. The policy balances load to improve network efficiency and incorporates QoS sensitivity to improve call quality. Simulation results show the following: (i) congestion pricing based admission control lowers call blocking probability, increases provider revenue, and improves economic efficiency over a static flat-rate admission control scheme; (ii) congestion sensitivity in the redirection policy balances load across all the ITGs while QoS sensitivity improves call audio quality; and (iii) incorporating price sensitivity in the redirection policy improves the economic efficiency, i.e., ensures that users who pay more get higher QoS. The techniques studied in this paper can be combined into a single resource management solution that can improve network resource utilization, provide differentiated service, and maximize provider revenue.

Server-side congestion arises when a large number of users wish to retrieve files from a server over a short period of time. Under such conditions, users are in a unique position to benefit enormously by sharing retrieved files. Pseudoserving, a new paradigm for Internet access, provides incentives for users to contribute to the speedy dissemination of server files through a contract set by a "superserver". Under this contract, the superserver grants a user a referral to where a copy of the requested file may be retrieved in exchange for the user's assurance to serve other users for a specified period of time. Simulations that consider only network congestion occurring near the server show that: (1) pseudoserving is effective because it self-scales to handle very high request rates; (2) pseudoserving is feasible because a user who participates as a pseudoserver benefits enormously in return for a relatively small contribution of the user's resources; (3) pseudoserving is robust under realistic user behavior because it can tolerate a large percentage of contract breaches; and (4) pseudoserving can exploit locality to reduce usage of network resources. Experiments performed on a local area network that account for the processing of additional layers of protocols and the finite processing and storage capacities of the server and the clients, corroborate the simulation results. They also demonstrate the benefits of exploiting network locality in reducing download times and network traffic while making referrals to a pseudoserver. Limitations of pseudoserving and potential solutions to them are also discussed.