Wireless networks are a fixture of present day computing. We are seeing a simultaneous increase in network density and throughput demand as the clients of these networks grow accustomed to more data hungry applications. Contention-based channel access methods take bigger performance hits and waste more energy as network density and load increases. It is therefore clear that the future of wireless networking will need to exploit some form of schedule based channel access in order to simultaneously solve the problems of energy consumption and maximization of channel utilization.

The focus of this work is on leveraging implicit properties of network traffic to benefit the performance of schedule based medium access mechanisms. We focus on one of these properties: the packet arrival behavior of the traffic. We chose to start our work by trying to answer the following question: "If we use predictions of the behaviors of flows in the network, can we decrease the delay in schedule-based medium access control?" The main idea is to use traffic forecasting to anticipate transmission schedules instead of establishing them reactively, i.e., as traffic arrives at the MAC layer. Although not all applications generate forecastable traffic, we contend that many applications do. Examples of predictable network traffic include Voice-over-IP (VoIP) applications such as Skype, iChat, and Google talk. Video streaming applications have lower QoS demands but also contain many predictable patterns. All of these applications are becoming increasingly commonplace in the home networks of today.

An experimental method was used to evaluate the benefit that accurate traffic prediction could have on the performance of a schedule based MAC protocol (DYNAMMA). Comparing the performance of DYNAMMA to our modified version of it (DYNAMMA-PRED) in simulations showed that prediction does improve delay performance of the schedule based protocol significantly, particularly at lower network loads.

The next step was to address the topic of extracting patterns out of packet arrival times of each flow with more mathematical rigor. We did this by measuring the entropy of packet arrivals in a network flow. Given that entropy is defined as the "measure of information", its value in this context signifies the amount of pattern in the packet arrival times of a flow - the less information each arrival holds, the more pattern there is overall.

During our investigation of the entropy of the packet arrival times, our research produced the concept of an "entropy fingerprint" - a plot of the entropy of the packet arrival times of a flow over a range of time scales. Each entropy fingerprint has numerous characteristics that are related to the packet arrival behavior of the flow that generated it. These fingerprints can be used in many ways, such as identifying what application generated the flow or whether the flow's packet arrivals are likely to be regular or irregular at a given time scale. In addition to the entropy fingerprints, the entropy estimator that we developed turned out to be usable as forecaster as well, able to predict the chances of a packet arrival in the next slot.

Analyzing the entropy fingerprints of various types of traffic confirmed that there was useful information in the packet arrival times of network flows that could be leveraged in a schedule based MAC protocol. Furthermore, the work presented us with a traffic forecaster that we could use to extract this information for use in such a MAC protocol.

Following this work, we designed a medium access control protocol to embody the fusion of traffic forecasting with a schedule based access control mechanism. We called the protocol TRANSFORMA, which stands for TRAffic FORecasting Medium Access. TRANSFORMA was designed using the principle that the MAC layer should detect the properties of each flow transparently and adapt its level of service accordingly. TRANSFORMA attempts to do that by observing an application flow, learning its pattern, and forecasting the flows future behavior based on the observed one. In its current implementation, TRANSFORMAs forecaster examines the packet arrival process of each application flow and determines the corresponding per-flow inter-packet arrival times. It then uses this information to establish the flows medium access schedule. TRANSFORMA operates under the assumption that applications that place more stringent requirements, e.g., higher data rates and delay sensitivity have forecastable network usage patterns. The simulation results show that TRANSFORMA significantly improves on the delay performance of its predecessor, DYNAMMA.

The final contribution of this work is a real-system implementation of TRANSFORMA. This fully functional network link enables experimentation with real application traffic, serves to validate our simulation results and demonstrates that the concepts embodied in TRANSFORMA are practical.

Outdoor IoT networks are used for a range of applications including surveillance and environmental monitoring. The nodes of such networks sense environmental data and, through wireless networks that interconnect them, transmit these measurements to data collection sites known as sinks. Much research in this area focuses on 2D deployments where nodes must remain coplanar and 3D deployments where nodes can move freely through the air or underwater. However outdoor deployments where nodes remain on the surface of terrain cannot be modeled as either 2D or 3D because the models fail to account for obstructions to communication and coverage that are caused by the terrain itself. Such “2.5D deployments” require terrain-aware deployment tools and algorithms that are not yet well developed. In this dissertation I present my work on such tools and algorithms.

First, I developed two novel terrain-aware network deployment algorithms. Grid Partition accounts for visibility over 2.5D terrain, but it also improves performance by using elevation as a visibility proxy that vastly reduces the number of more costly visibility computations. The other algorithm, EMNAglobal,steps, improves on EMNAglobal, an estimation-of-distribution algorithm, by computing the covariance matrix of the solution population’s distribution by referencing the mean of the previous generation instead of the mean of the current generation’s mean. Comprehensive experiments show that the proposed algorithms perform as well as or better than traditional optimization meta-heuristics, measured as network sensor coverage. In addition, the core idea behind one of the algorithms, Grid Partition, can serve as a meta-heuristic itself and is used in one of my recoveryalgorithms.

Next, I developed and evaluated a class of terrain-aware network recovery algorithms which guide networks in self-repair following a node failure. Terrain-Aware Recovery with Commshed Intersections (TARCI) consists of four algorithm variants. Experiments show that the variant that incorporates Grid Partition as a meta-heuristic (TARCI-GP) achieves nearly the same quality of results as a variant that includes an exhaustive step while requiring two orders of magnitude less CPU time. In addition, I show that my terrain-aware recovery algorithms perform better than recovery algorithms that are intended for 2Ddeployments.

Finally, to help with the development of my terrain-aware algorithms, I created a terrain-aware experimental framework. The Terrain-Aware Framework For IoT (TAFFI) simulates Java- and C++-based algorithms, both distributed and centralized. The framework provides a library of line-of-sight algorithms and a suite of benchmark terrains. To ensure that the benchmarks include a wide range of landform variation, another contribution of this dissertation is the design of two terrain classifiers that ensure that the benchmark terrains used in the experiments represent a wide variety of terrain landforms.

In contrast to live hardware testbeds, which are evaluated largely on performance, network simulators are primarily used to test and validate new protocols and algorithms. As such, a network simulator should be designed to be modular, extensible, re-usable, and robust. This work examines the design of TCP congestion control in the ns-3 network simulator and presents a new design, not specific to ns-3, that will be more modular and extensible. We also demonstrate that the implementation in the current ns-3 release (version 3.25) is not able to duplicate results presented in previous literature. These findings, along with the improved design, are the main contributions of this paper. To validate our design, we show how its modular approach can be used to implement existing TCP congestion control variants, e.g., TCP New Reno and TCP Westwood, as well as new ones, e.g., TCP Inigo.

Several network protocols, services, and applications adjust their operation dynamically based on current network conditions. Consequently, keeping accurate estimates of network conditions and performance as they fluctuate over time is critical.

In this thesis, we explore the use of computational intelligence, in particular machine learning techniques to estimate "near-future" network performance based on past network conditions. We call our approach to network performance estimation SENSE for Smart Experts for Network State Estimation. SENSE is able to respond to network dynamics at different time scales, i.e., long- and medium-term fluctuations as well as short-lived variations.

Then, by applying SENSE, we proposed a novel algorithm to dynamically enable and disable IEEE 802.11 DCF's RTS/CTS handshake. Our algorithm uses current packet size and transmission rate, as well as an estimate of network contention to dynamically decide whether to use RTS/CTS. To the best of our knowledge, the proposed algorithm is the first to enable and disable the RTS/CTS handshake based on a set of current network conditions, and automatically adapt as these conditions change. Simulation results using a variety of WLAN- as well as wireless multi-hop ad-hoc network scenarios, including synthetic and real traffic traces, demonstrate that the proposed approach consistently outperforms current best practices, such as never enabling RTS/CTS or using a pre-specified threshold to decide whether to switch RTS/CTS on or off.

We also propose a modified version of a simple, yet effective machine learning technique called "Fixed-Share" algorithm to optimize IEEE 802.11's backoff algorithm. To the best of our knowledge, this is the first approach that uses machine learning to dynamically set the IEEE 802.11's contention window based on past performance. Through simulations using a variety of network scenarios, we show that our method outperforms IEEE 802.11's original exponential back off algorithm as well as an approach that adapts based on a few recent data transmission events.

User privacy on the Internet has been an increasing concern in recent years. With the proliferation and sophistication of information services, data mining, and search engines, a simple network address may be used to reveal a great deal of information about a user, including location, identity, and behavior.

A new network architecture paradigm known as "Software-Defined Networking" (SDN) has recently garnered attention in both industry and academia. Defined by a separation of data and control planes, it offloads routing decisions from the switching hardware and provides an innovative framework upon which new protocols and services can be deployed.

In this thesis, we examine how SDN can be employed by service providers to offer endpoint privacy. We identify existing approaches to privacy and determine that they make unacceptable tradeoffs between performance and anonymity. We argue an acceptable level of privacy can be provided to most users, with noticeably lower latency and throughput impact, by working with the network provider; to that end, we introduce AnonyFlow, an in-network anonymization service designed to efficiently and seamlessly provide privacy to users as they communicate with other endpoints and services. We design, implement, and evaluate a prototype of AnonyFlow, based on an OpenFlow SDN deployment, that achieves endpoint anonymity without compromising on throughput or latency.

Network simulations are heavily used in the networking community to evaluate the performance of computer networks and their protocols. Simulations are often chosen over alternatives such as live experiments due to limited resources in terms of scalability, as well as reproduceability of the experiments. Many of the routing protocols designed for Mobile Ad Hoc Networks (MANETs) are evaluated solely on their performance calculated by these simulations, but the simulation environments the routing protocols are exposed to are often limited in scope. Only certain aspects of the routing protocols are tested, so the protocols are only understood in terms of the fabricated scenarios that they are subjected to.

We first investigate the current best practices in simulation-based multi-hop wireless ad-hoc network (MANET) protocol evaluation to examine how wide-spread this problem is in the networking community. We extend a prior characterization of the settings and parameters used in MANET simulations by studying the papers published in one of the premier mobile networking conferences between 2006 and 2010. We find that there are still several configuration pitfalls which many papers fall victim to, which in turn damages the integrity of the results as well as any research aimed at reproducing and extending these results. We then describe the simulation “design space” of MANET routing in terms of its basic dimensions and corresponding parameters. We then discuss the benchmark infrastructure that was created to provide an easy to use solution for testing these protocols in a wide range of scenarios. The following chapter looks extensively at the realistic scenarios provided with the benchmark that act as sample scenarios to promote modeling simulations after real-world situations, and to show the flexibility in adding new scenarios. We also propose four “auxiliary” metrics to increase simulation integrity. Next, we show results generated by the benchmarking tool and provide our concluding thoughts.

This thesis describes the design and implementation of the first version of the Twizzler network stack. In this implementation, we focused on providing Twizzler developers with the network functionality that was immediately needed, while having flexibility that promotes growth and expansion. To encourage research in creating a data-centric network stack, we implemented the FLIP protocol as Layer 3 of the network stack, moving away from the traditional TCP/IP stack. FLIP allows for custom headers and its adaptability gives application developers control over which functions the network provides for their specific application. We also describe the implementation of ARP, the first network application using the Twizzler stack, which allows for network discovery. We expect that the future directions of this work will propel the development of a full-fledged network stack for the Twizzler operating system.

With the advancement of the Internet of Things movement, ubiquitous computing is now an accelerating field for embedded system researchers to pursue. Almost all of these new connected systems are equipped with sensors to gather useful information about their surroundings, and of these sensors the visual sensor presents the richest available datasets in comparison to its scalar counterparts. Deploying wireless sensor networks in remote environments makes it difficult to acquire power to these systems and many choose a battery-supplied source. After optimizing the sensor node for long battery life, many sacrifices are made in performance to delay the need for battery replacement maintenance. Now, however, there is promising furtherance in energy harvesting and rechargeable battery capable sensor nodes, and the need for node maintenance may be obsolete in the coming years. Sensor network devices can now focus less on optimizing for battery life and instead increase the amount of on-board local processing making the device "smart". What defines a smart camera is its ability to process video footage locally instead of a central processing unit online or at a remote network sink node. Progress in the field of computer vision has brought the ability to identify complex objects and behaviors of monitored environments. In this thesis, we present SlugCam, a solar-powered wireless smart camera network platform that can be used in a variety of outdoor applications including surveillance of public spaces, habitat and environmental monitoring, wildfire prevention and detection, to name a few. The system is built with off-the-shelf components which not only keeps it modular and low cost, but also facilitates its prototyping, rapid duplication, and evolution. SlugCam's on-board processing capability allows computer vision software to run locally which contributes to the system's autonomous operation capabilities. Energy efficiency in SlugCam is accomplished both in: (1) hardware by micro-managing low-power components; as well as in (2) software by having the system's operation duty cycles automatically adapt to the current state of the battery in order to balance the trade-off between application-level requirements and power awareness.

We present PIMAP, a system framework for continuous patient monitoring with a specific focus on preventing pressure injuries (a.k.a. bed sores). Pressure injuries are classified as “never events”, meaning they should never occur in healthcare facilities and yet in the U.S. they affect 2.5 million patients a year at a cost of $2.5 billion. In addition the majority of patients affected are the most vulnerable, the elderly and/or physically disabled.

There are many proposed solutions to prevent pressure injuries, the most promising are patient monitoring based, such as monitoring the pressure of the patient against a mattress, monitoring the motion of the patient, and measuring the health of the patient’s skin. Patient monitoring has the advantage that data can be automatically collected and analyzed without healthcare intervention, providing additional insights that would otherwise have to be calculated manually or ignored.

Through the identification of the most promising techniques and through anecdotal evidence from our collaboration with UCSF we discovered a lack of a reliable way to sense, store, analyze, and visualize novel medical device data. There is no current system that is able to: (1) seamlessly, reliably, and persistently acquire patient monitoring data from various medical devices, (2) analyze acquired data, and (3) present the results to healthcare personnel in an ecient and user-friendly fashion. Instead there are many one-off solutions that will only work with a specific medical device or commercial systems that only work with a commercial medical device.

From this motivation we present Pressure Injury Monitoring And Prevention (PIMAP), a system framework that presents a standard to sense, store, analyze, and visualize medical sensor data in real-time. The framework abstracts the system so that researchers can focus their efforts on specifics, such as the medical device or analysis without having to develop an entire system.

Vehicles following each other in close-proximity in the form of a platoon reduce fuel consumption due to reduction in air drag saving up to 5% for the first vehicle and 10% for the following vehicles. Introduction of vehicle-to-vehicle communication has led to further decrease in inter-vehicular spacing, enabling further reduction in fuel consumption and improving the road capacity by up to 166%. Dynamic platoon formation and dissolution is generally governed by the lead vehicle, leading to a centralized approach. This suffers from various drawbacks such as single point of failure and performance bottlenecks resulting in serialization of maneuvers and shorter platoon length. A decentralized approach mitigates these issues and is traditionally realized using the "top-down" approach by first identifying the maneuvers and then executing the required steps and exchanging the necessary messages. Such decentralized platooning systems, which are known as “Deliberate Platooning”, suffer from higher complexity and reduced flexibility. This dissertation presents ADEPT – Adaptive Decentralized Emergent-behavior based PlaTooning, that aims at mitigating the shortcomings of "traditional" decentralized platooning systems. ADEPT uses a novel decentralized automated vehicle platooning approach inspired by nature's "emergent behavior” commonly found in biological systems. In ADEPT, each vehicle follows a set of simple rules when they need to interact in order to carry out the maneuvers. Platooning maneuvers such as join, exit, and merge “emerge” as a result of vehicles following these “emergent rules”. Through extensive simulation experiments using a platooning-enabled vehicular network simulator driven by a wide range of scenarios, we demonstrate that ADEPT yields superior performance in terms of maneuver time and communication overhead, especially in multi-vehicle multi-maneuver scenarios. Additionally, we devise mechanisms for obstacle avoidance, vehicle following, gap- and predecessor determination on curved roads and, using a well-known simulation environment equipped with a physics engine, demonstrate that our emergent platooning approach is effective when vehicle characteristics such as weight, center of mass and friction are considered.