隋著移動數據網絡的迅速發展，世界各地部署著不同制式的3G，HSPA和LTE網絡，這給移動互聯網應用帶來新的機遇和挑戰。不像他們的對手, 有線網絡，移動數據網絡有較長的延遲，較高的包率，以及急劇波動的頻寬。為了應付這些挑戰，移動無線基站往往配備較大的緩衝區（KBs以至數百多MBs），以吸收短期的頻寬波動，並促進鏈路層重傳。然而，由於大部分互聯網協議，並特別是TCP，全都假設路由器只擁有較小緩衝區，所以在移動網絡的大型緩衝區下，往往導致TCP展現出次優的性能。本論文解決兩個根本性的問題來克服這一項挑戰。首先，我們開發了一種新算法 (SoD）專門來估計網絡的鏈路緩衝區的大小和隊列長度。在今天的3G移動數據網絡，本文以大量網絡數據驅動的模擬結果證明 SoD 表現優於現有的算法，如 Max-min 和 Loss-pair。另外，我們採用 SoD算法來創造新的TCP擁塞控制模塊以解決移動網絡的大型緩衝區以及的頻寬波動問題。這個新的TCP變種，稱為 TCP- QA，在模擬情況以及移動數據網絡下均表現出大大優於現有的TCP變種，包括TCP CUBIC ，TCP Vegas，TCP Westwood，和FAST TCP。 / The rapidly emerging mobile data networks fueled by the world-wide deployment of 3G, HSPA, and LTE networks created new opportunities and challenges for developing mobile Internet applications. Unlike their wired counterpart, mobile data networks are known to exhibit longer delay, higher packet loss rate, and rapidly fluctuating bandwidth. To tackle these challenges mobile radio base stations are often equipped with large buffers (from hundreds of KBs to multi-MBs) to absorb short-term bandwidth fluctuations and to facilitate link-layer retransmissions. However as most Internet protocols in general, and TCP in particular, were designed with the assumption of small router buffer size, the large buffer in mobile networks can and do interact, often negatively, with the Internet protocols, leading to sub-optimal performance. This work tackles two fundamental problems in overcoming this challenge. First, we developed a novel Sum-of-Delays (SoD) algorithm specifically designed to estimate the link buffer size and queue length of bandwidth-varying networks. Extensive trace-driven simulation results showed that SoD outperforms existing algorithms such as max-min and loss-pair by orders of magnitude in today’s 3G mobile data networks. Second, we apply the SoD algorithm to TCP’s congestion control module to incorporate and compensate for the mobile network’s large buffer size. This new TCP variant, called TCP-Queue-length-Adaptive (TCP-QA), substantially outperforms existing TCP variants including TCP CUBIC, TCP Vegas, TCP Westwood, and FAST TCP. / Detailed summary in vernacular field only. / Chan, Chi Fung. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2012. / Includes bibliographical references. / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts also in Chinese. / Abstract --- p.ii / 摘要 --- p.iii / Acknowledgements --- p.iv / Contents --- p.v / Chapter Chapter 1 --- Introduction --- p.1 / Chapter Chapter 2 --- Background and Related Work --- p.5 / Chapter 2.1 --- Link Buffer Size Estimation --- p.6 / Chapter 2.2 --- Queue Length Estimation --- p.8 / Chapter 2.3 --- Performance over Fixed Networks --- p.10 / Chapter 2.4 --- Performance over Bandwidth-Varying Networks --- p.15 / Chapter Chapter 3 --- Sum-Of-Delay Method --- p.19 / Chapter Chapter 4 --- Performance Evaluation --- p.22 / Chapter 4.1 --- Link Buffer Size Estimation in Fixed Bandwidth Networks --- p.23 / Chapter 4.2 --- Link Buffer Size Estimation in Variable Bandwidth Networks --- p.25 / Chapter 4.3 --- Queue Length Estimation in Fixed Bandwidth Networks --- p.28 / Chapter 4.4 --- Queue Length Estimation in Variable Bandwidth Networks --- p.31 / Chapter 4.5 --- Convergence --- p.34 / Chapter 4.6 --- Limitations in Passive Estimation --- p.40 / Chapter Chapter 5 --- Impact of Competing Traffic --- p.42 / Chapter 5.1 --- Simulation Setup --- p.42 / Chapter 5.2 --- Link Buffer Size Estimation --- p.44 / Chapter 5.3 --- Queue Length Estimation --- p.50 / Chapter 5.4 --- Discussions --- p.54 / Chapter Chapter 6 --- TCP-Queue-Length-Adaptive (TCP-QA) --- p.56 / Chapter 6.1 --- Challenges in Mobile Networks --- p.57 / Chapter 6.2 --- Queue-Length-Adaptive Congestion Control --- p.59 / Chapter 6.3 --- Performance Comparisons --- p.65 / Chapter Chapter 7 --- Summary and Future Work --- p.76 / Bibliography --- p.78

Today`s continuously growing Internet requires users and network applications to have knowledge of network metrics. This knowledge is critical for decision making during the usage of network applications. This thesis studies application related network metrics. The major approach in this work is to examine the traffic between a simulated user and network applications. We use the historical data collected from previous usage of network applications to make predictions for future usage of those applications. We also use the historical data obtained from a given application to make predictions about another application. Prediction mechanisms require us to make parameter choices so that certain weights can be placed on historical data versus current data. We study these different choices and use the values from our best experimental results. From these studies we conclude that our data prediction is quite accurate and remains stable over a range of parameter choices. The use of shared routing paths between users and network applications are explored in the performance prediction of applications. Only some servers at the same locations show similar prediction results. The network applications studied are also varied, including web, streaming, DNS, etc. We see whether sharing information obtained from different applications can be used to make predictions of application performance. However, we observe limited success in predictions across applications.

The deployment of heterogeneous networks (HetNets), in which low power nodes (LPNs) and high power nodes (HPNs) coexist, has become a promising solution for extending coverage and increasing capacity in wireless networks. Meanwhile, several advanced technologies such as massive multi-input multi-output (MIMO), cloud radio access networks (C-RAN) and device-to-device (D2D) communications have been proposed as competent candidates for supporting the next generation (5G) network. Since single technology cannot solely achieve the envisioned 5G requirements, the e ect of integrating multiple technologies in one system is worth to be investigated. In this thesis, a thoroughly theoretical analysis is conducted to evaluate the network performance in di erent scenarios, where two or more 5G techniques are employed. First, the downlink performance of massive MIMO enabled HetNets is fully evaluated. The exact and asymptotic expressions for the probability of a user being associated with a macro cell or a small cell are presented. The analytical expressions for the spectrum e ciency (SE) and energy e ciency (EE) in the K-tier network are also derived. The analysis reveals that the implementation of massive MIMO in the macro cell can considerably improve the network performance and decrease the demands for small cells in HetNets, which simpli es the network deployment. Then, the downlink performance of a massive MIMO enabled heterogeneous C-RAN is investigated. The exact expressions for the SE and EE of the remote radio heads (RRHs) tier and a tractable approximation approach for evaluating the SE and EE of the macrocell tier are obtained. Numerical results collaborate the analysis and prove that massive MIMO with dense deployment of RRHs can signi cantly enhance the performance of heterogeneous C-RAN theoretically. Next, the uplink performance of massive MIMO enabled HetNets is exploited with interference management via derived SE and EE expressions. The numerical results show that the uplink performance in the massive MIMO macrocells can be signi cantly improved through uplink power control in the small cells, while more uplink transmissions in the macrocells have mild adverse e ect on the uplink performance of the small cells. In addition, the SE and EE of the massive MIMO macrocells with heavier load can be improved by expanding the small cell range. Lastly, the uplink performance of the D2D underlaid massive MIMO network is investigated and a novel D2D power control scheme is proposed. The average uplink achievable SE and EE expressions for the cellular and D2D are derived and results demonstrate that the proposed power control can e ciently mitigate the interference from the D2D. Moreover, the D2D scale properties are obtained, which provide the su cient conditions for achieving the anticipated SE. The results demonstrate that there exists the optimal D2D density for maximizing the area SE of D2D tier. In addition, the achievable EE of a cellular user can be comparable to that of a D2D user. Stochastic geometry is applied to model all of the systems mentioned above. Monte Carlo simulations are also developed and conducted to validate the derived expressions and the theoretical analysis.

A complex adaptive systems approach has been permeating organizational studies and the field of supply network management helping to describe and explain supply network dynamics and emergent inter-firm structures. This has improved our theoretical knowledge of the nature of supply networks transforming raw materials into products, within a constantly changing environment. From the early days of simple structures, describing bi-lateral, local arrangements between firms for the creation of relatively simple products, we are now in an environment of various supply network archetypes, describing different global sourcing regimes of highly integrated, sophisticated products within multi-tier networks. This thesis is a study of the coevolution of the firm and supply network in the commercial aerospace manufacturing sector producing jetliners of 100 or more seats. One of the contributions of this research is to demonstrate how the holistic approach of complexity science can be applied to describe, understand and gain new insight into the coevolution of the firm and the supply network. Based on the findings of multiple interviews and questionnaires in eight global aerospace firms across multiple supply chain tiers, this research finds high-performing clusters of inter-firm characteristics, plus the aspects of structure and integration which deliver the supply network performance. Practitioners can use these specific results to examine their own firms and the new coevolutionary conceptual framework developed in the thesis may aid future research studies of complex adaptive systems in practice. The simple survey design and analysis method used in the final research stage of this research, has the potential for use in other industries, markets and other complex adaptive systems generally to examine performance outcomes and the effects of having or adopting new inter-firm characteristics. Finally, implications for policy include the potential to legitimize supply networks in order to stimulate competition and innovation in the economy.

The popularity of the World Wide Web (WWW) in providing a vast array of information has drawn a large number of users in the past few years. The dramatic increase in the number of Internet users, however, has brought undesirable impacts on users, such as long response time and service unavailability. The utilization of multiple servers can be used to reduce adverse impacts. The challenge is to identify a good resource site to allocate to the user given a group of servers from which to select. In this project, a performance-directed site selection system was developed for a web-based application called AADMLSS (African American Distributed Multiple Learning Styles System). Four different sets of experiments were conducted in this study. In order to evaluate the effectiveness of the test system, two other server selection methods, Load-based and Random-based methods, were implemented for comparative purposes. The experiments were also run during daytime and nighttime to see the impact of network load on the response time. Experimental results indicate that the performance-directed site selection system outperforms the Load-based and Random-based methods consistently. The response time is typically high during daytime and low during nighttime, indicating that the network load has an impact on the response time delivered. The results also show that server performance contributes to the overall response time, and network performance is the more dominating factor in determining a good resource site for the user.

Thesis (M.S.)--Mississippi State University. Department of Computer Science. / Title from title screen. Includes bibliographical references.

Cloud computing is a new computing paradigm that aims to transform computing services into a utility, just as providing electricity in a “pay-as-you-go” manner. Data centers are increasingly adopting virtualization technology for the purpose of server consolidation, flexible resource management and better fault tolerance. Virtualization-based cloud services host networked applications in virtual machines (VMs), with each VM provided the desired amount of resources using resource isolation mechanisms. Effective network performance isolation is fundamental to data centers, which offers significant benefit of performance predictability for applications. This research is application-driven. We study how network performance isolation can be achieved for latency-sensitive cloud applications. For media streaming applications, network performance isolation means both predicable network bandwidth and low-jittered network latency. The current resource sharing methods for VMs mainly focus on resource proportional share, whereas ignore the fact that I/O latency in VM-hosted platforms is mostly related to resource provisioning rate. The resource isolation with only quantitative promise does not sufficiently guarantee performance isolation. Even the VM is allocated with adequate resources such as CPU time and network bandwidth, problems such as network jitter (variation in packet delays) can still happen if the resources are provisioned at inappropriate moments. So in order to achieve performance isolation, the problem is not only how many/much resources each VM gets, but more importantly whether the resources are provisioned in a timely manner. How to guarantee both requirements to be achieved in resource allocation is challenging. This thesis systematically analyzes the causes of unpredictable network latency in VM-hosted platforms, with both technical discussion and experimental illustration. We identify that the varied network latency is jointly caused by VMM CPU scheduler and network traffic shaper, and then address the problem in these two parts. In our solutions, we consider the design goals of resource provisioning rate and resource proportionality as two orthogonal dimensions. In the hypervisor, a proportional share CPU scheduler with soft real-time support is proposed to guarantee predictable scheduling delay; in network traffic shaper, we introduce the concept of smooth window to smooth packet delay and apply closed-loop feedback control to maintain network bandwidth consumption. The solutions are implemented in Xen 4.1.0 and Linux 2.6.32.13, which are both the latest versions when this research was conducted. Extensive experiments have been carried out using both real-life applications and low-level benchmarks. Testing results show that the proposed solutions can effectively guarantee network performance isolation, by achieving both predefined network bandwidth and low-jittered network latency. / published_or_final_version / Computer Science / Master / Master of Philosophy

The research work presents an enhanced cross-layer routing solution for Wireless Mesh Networks (WMN) based on Received Signal Strength. WMN is an emerging technology with varied applications due to inherent advantages ranging from self-organisation to auto-con guration. Routing in WMN is fundamen- tally achieved by hop counts which have been proven to be de cient in terms of network performance. The realistic need to enhance the link quality metric to improve network performance has been a growing concern in recent times. The cross-Layer routing approach is one of the identi ed methods of improving routing process in Wireless technology. This work presents an RSSI-aware routing metric implemented on Optimized Link-State Routing (OLSR) for WMN. The embedded Received Signal Strength Information (RSSI) from the mesh nodes on the network is extracted, processed, transformed and incorporated into the routing process. This is to estimate efficiently the link quality for network path selections to improved network performance. The measured RSSI data is filtered by an Exponentially Weighted Moving Average (EWMA) filter. This novel routing metric method is called RSSI-aware ETT (rETT). The performance of rETT is then optimised and the results compared with the fundamental hop count metric and the link quality metric by Expected Transmission Counts (ETX). The results reveal some characteristics of RSSI samples and link conditions through the analysis of the statistical data. The divergence or variability of the samples is a function of interference and multi-path e effect on the link. The implementation results show that the routing metric with rETT is more intelligent at choosing better network paths for the packets than hop count and ETX estimations. rETT improvement on network throughput is more than double (120%) compared to hop counts and 21% improvement compared to ETX. Also, an improvement of 33% was achieved in network delay compared to hop counts and 28% better than ETX. This work brings another perspective into link-quality metric solutions for WMN by using RSSI to drive the metric of the wireless routing protocol. It was carried out on test-beds and the results obtained are more realistic and practical. The proposed metric has shown improvement in performance over the classical hop counts metric and ETX link quality metric.

M. Tech. Electrical Engineering. / High-speed transmission network requires adequate synchronization planning for quality of service. Based on ITU-T synchronization network guidelines, this dissertation proposes a Synchronous Digital Hierachy (SDH) synchronization model catering for loop of synchronization and robust to communications links faults. From candidate schemes for providing synchronization information, we have adopted the master-slave strategy. Heuristics have been developed to assist in the allocation of primary and back-up reference clocks. The contribution lies in the systematic selection of master node placement, the distribution of the reference clock from master node to all slave nodes using shortest path and the allocation of system redundancy by means of clock priority table for each nodal clock. To cater for clock stability and accuracy, the use of atomic clock (Cesium, Rubidium) as frequency standard, in long term basis, has proven that clock s characteristics namely stability and accuracy may be controlled in the midst of jitter/wander. Fiber optics transmission medium has proven to be adequate for optimal clock dissemination with very trivial frequency deviation from the nominal positional instant, traceable to Primary Reference Clock.

M. Tech. Electrical Engineering. / This study proposes a distributive joint power and rate adaptation algorithm (JRPA) in ad hoc networks based on coupled interference minimisation. In the proposed method, the influence of coupled interference was controlled by dynamically adjusting network users' transmit power choices. The users are therefore aware of the current link status while determining their data rates. In addition, every maximize utility of other users as it maximizes its utility due to the inevitable cooperation, hence, improving a collective network performance. Solving this network utility maximization problem results in a supermodular game equivalence where users cooperate to maximise both local and global utility, hence the supermodular game theory concept was used to analyse the optimality and convergence of the proposed solution.