Sami Rollins

Peer-to-Peer Computing

The term “peer-to-peer” (P2P) refers to a class of systems and applications that employ distributed resources to perform a critical function in a decentralized manner. With the pervasive deployment of computers, P2P is increasingly receiving attention in research, product development, and investment circles. This interest ranges from enthusiasm, through hype, to disbelief in its potential. Some of the benefits of a P2P approach include: improving scalability by avoiding dependency on centralized points; eliminating the need for costly infrastructure by enabling direct communication among clients; and enabling resource aggregation. This survey reviews the field of P2P systems and applications by summarizing the key concepts and giving an overview of the most important systems. Design and implementation issues of P2P systems are analyzed in general, and then revisited for each of the case studies described in Section 6. This survey will help people understand the potential benefits of P2P in the research community and industry. For people unfamiliar with the field it provides a general overview, as well as detailed case studies. It is also intended for users, developers, and information technologies maintaining systems, in particular comparison of P2P solutions with alternative architectures andPeer-to-peer (P2P) technology, or peer computing, is a paradigm that is viewed as a potential technology for redesigning distributed architectures and, consequently, distributed processing. Yet the scale and dynamism that characterize P2P systems demand that we reexamine traditional distributed technologies. A paradigm shift that includes self-reorganization, adaptation and resilience is called for. On the other hand, the increased computational power of such networks opens up completely new applications, such as in digital content sharing, scientific computation, gaming, or collaborative work environments. In this book, Vu, Lupu and Ooi present the technical challenges offered by P2P systems, and the means that have been proposed to address them. They provide a thorough and comprehensive review of recent advances on routing and discovery methods; load balancing and replication techniques; security, accountability and anonymity, as well as trust and reputation schemes; programming models and P2P systems and projects. Besides surveying existing methods and systems, they also compare and evaluate some of the more promising schemes. The need for such a book is evident. It provides a single source for practitioners, researchers and students on the state of the art. For practitioners, this book explains best practice, guiding selection of appropriate techniques for each application. For researchers, this book provides a foundation for the development of new and more effective methods. For students, it is an overview of the wide range of advanced techniques for realizing effective P2P systems, and it can easily be used as a text for an advanced course on Peer-to-Peer Computing and Technologies, or as a companion text for courses on various subjects, such as distributed systems, and grid and cluster computing.

Turducken: hierarchical power management for mobile devices

Maintaining optimal consistency in a distributed system requires that nodes be always-on to synchronize information. Unfortunately, mobile devices such as laptops do not have adequate battery capacity for constant processing and communication. Even by powering off unnecessary components, such as the screen and disk, current laptops only have a lifetime of a few hours. Although PDAs and sensors are similarly limited in lifetime, a PDAs power requirement is an order-of-magnitude smaller than a laptops, and a sensors is an order-of-magnitude smaller than a PDAs. By combining these diverse platforms into a single integrated laptop, we can reduce the power cost of always-on operation. This paper presents the design, implementation, and evaluation of Turducken, a Hierarchical Power Management architecture for mobile systems. We focus on a particular instantiation of HPM, which provides high levels of consistency in a laptop by integrating two additional low power processors. We demonstrate that a Turducken system can provide battery lifetimes of up to ten times that of a standard laptop for always-on operation and three times for a system that periodically sleeps.

Users and batteries: interactions and adaptive energy management in mobile systems

Battery lifetime has become one of the top usability concerns of mobile systems. While many endeavors have been devoted to improving battery lifetime, they have fallen short in understanding how users interact with batteries. In response, we have conducted a systematic user study on battery use and recharge behavior, an important aspect of user-battery interaction, on both laptop computers and mobile phones. Based on this study, we present three important findings: 1) most recharges happen when the battery has substantial energy left, 2) a considerable portion of the recharges are driven by context (location and time), and those driven by battery levels usually occur when the battery level is high, and 3) there is great variation among users and systems. These findings indicate that there is substantial opportunity to enhance existing energy management policies, which solely focus on extending battery lifetime and often lead to excess battery energy upon recharge, by adapting the aggressiveness of the policy to match the usage and recharge patterns of the device. We have designed, deployed, and evaluated a user- and statistics-driven energy management system, Llama, to exploit the battery energy in a user-adaptive and user-friendly fashion to better serve the user. We also conducted a user study after the deployment that shows Llama effectively harvests excess battery energy for a better user experience (brighter display) or higher quality of service (more application data) without a noticeable change in battery lifetime.

Javelin++: scalability issues in global computing

Javelin is a Java-based infrastructure for global computing. This paper presents Javelin++, an extension of Javelin, intended to support a much larger set of computational hosts. First, Javelin++‘s switch from Java applets to Java applications is explained. Then, two scheduling schemes are presented: a probabilistic work-stealing scheduler and a deterministic scheduler. The deterministic scheduler also implements eager scheduling, as well as another fault-tolerance mechanism for hosts that have failed or retreated. A Javelin++ API is sketched, then illustrated on a raytracing application. Performance results for the two schedulers are reported, indicating that Javelin++, with its broker network, scales better than the original Javelin.

Automating energy management in green homes

Homes powered fully or partially by renewable sources such as solar are becoming more widely adopted, however energy management strategies in these environments are lacking. This paper presents the first results of a study that explores home automation techniques for achieving better utilization of energy generated by renewable technologies. First, using a network of off-the-shelf sensing devices, we observe that energy generation and consumption in an off-grid home is both variable and predictable. Moreover, we find that reactive energy management techniques are insufficient to prevent critical battery situations. We then present a recommendation based system for helping users to achieve better utilization of resources. Our study demonstrates the feasibility of three recommendation components: an early warning system that allows users of renewable technologies to make more conservative decisions when energy harvested is predicted to be low; a task rescheduling system that advises users when high-power appliances such as clothes dryers should be run to optimize overall energy utilization; and an energy conservation system that identifies sources of energy waste and recommends more conservative usage.

Pixie: a jukebox architecture to support efficient peer content exchange

Peer-to-peer (P2P) content exchange has recently gained attention from both the research and industrial communities. The dynamic nature of peer networks and the resource constraints of peer hosts have introduced a number of unique technical challenges that must be addressed to make large-scale P2P content exchange applications viable. In this work, we present and evaluate Pixie, an architecture that integrates one-to-many distribution of content and peer networks. Pixie provides a valuable data location service as well as a number of scalability properties both in terms of data location and content distribution. Our results indicate that, using a one-to-many scheme, we can significantly reduce the resources consumed in searching for and distributing content across peer networks. These scalability properties will become increasingly important as peer content exchange is extended to support more advanced applications.

Using rule mining to understand appliance energy consumption patterns

Managing energy in the home is key to creating a sustainable future for our society. More tools are increasingly available to measure home energy usage, however these tools provide little insight into questions such as why an appliance consumes more energy than normal or what kinds of behavioral changes might be most likely to reduce energy usage in the home. To answer these questions, a deeper understanding of the causal factors that influence energy usage is necessary. In this work, we conduct a broad study of factors that influence energy consumption of individual devices in the home. Our first contribution is collection of a context-rich data set from six homes across the United States. The second contribution of this work is a set of insights into key factors influencing energy usage derived by the novel application of a rule mining algorithm to identify significant associations between energy usage and four key features: hour of the day, day of the week, use of other appliances in the home, and user-supplied annotations of activities such as working or cooking. Our analysis confirms our hypothesis that, though most devices show a regular pattern of daily or weekly use, this is not true for all devices. Associations that relate use of two different devices in the same home are often stronger, and are observed for nearly 25% of device uses. Overall, we observe that the associations derived from the first five weeks of data in our data set are sufficient to explain nearly 70% of the device uses in the subsequent five weeks of data, and over 90% of the associations identified during the first five weeks recur in the latter portion of the data set. The associations identified by our approach may be used to to aid in end-user applications that heighten awareness and encourage energy savings, improve energy disaggregation algorithms, or even detect anomalous uses that may signal problems in aging-in-place homes.

Triage: balancing energy and quality of service in a microserver

The ease of deployment of battery-powered and mobile systems is pushing the network edge far from powered infrastructures. A primary challenge in building untethered systems is offering powerful aggregation points and gateways between heterogeneous end-points---a role traditionally played by powered servers. Microservers are battery-powered in-network nodes that play a number of roles: processing data fromclients, aggregating data, providing responses to queries, and actingas a network gateway. Providing QoS guarantees for theseservices can be extremely energy intensive. Since increasedenergy consumption translates to a shorter lifetime, there is a need for a new way to provide these QoS guarantees at minimal energy consumption.n This paper presents Triage, a tiered hardware and softwarearchitecture for microservers. Triage extends the lifetime of a microserver by combining two independent, but connected platforms: a high-power platform that provides the capability to executecomplex tasks and a low-power platform that provides high responsiveness at low energy cost. The low-power platform acts similarto a medical triage unit, examining requests to find critical ones, and scheduling tasks to optimize the use of the high-powerplatform. The scheduling decision is based on evaluating each tasks resource requirements using hardware-assisted profiling of execution time and energy usage. Using three microserver services, storage, network forwarding, and query processing, we show that Triage provides more than 300% increase in microserver lifetime over existing systems while providing probabilistic quality of service guarantees.

Power-aware data management for small devices

Pervasive computing devices such as Personal Digital Assistants (PDAs) and laptop computers are becoming increasingly ubiquitous. The future promises even more advanced devices such as digital watches, jewelry, and even clothing. However, as pervasive devices become more widely used for more advanced applications, their resource limitations are becoming more apparent. In this work, we focus on data management and power limitations. We investigate the benefit of using power-aware schemes to automatically manage content across a collection of devices and prolong data availability. We monitor the available energy supply on each device and migrate content from devices that are in danger of dying. In our simulated environment, we have found that, using intelligent techniques for data management can increase the amount of time a collection of devices remains usable by over 2 times. Furthermore, our techniques can perform autonomously, independent of user intervention.

AVoN calling: AXL for voice-enabled Web navigation

Abstract The World Wide Web is a rich source of information that has become a universal means of communication. XML promises to be the future of the World Wide Web. However, as HTML is replaced by its more powerful counterpart, traditional browsers are not sufficient to display the information communicated in an XML document. Todays browsers are capable of showing only a textual version of an XML document. This is limiting not only for a viewer in a traditional scenario, but is a barrier for a user who wishes to access the information without having access to a traditional keyboard, mouse, and monitor. This paper presents a framework for developing non-traditional, schema-driven, customizable interfaces used to navigate and modify XML documents that may be served over the Web. Our system, Audio XmL (AXL) focuses on developing a speech-based component within that framework. At the most basic level, we provide a Speech DOM, a spoken equivalent to the Document Object Model. Beyond that, we provide an intuitive set of commands based upon schema as well as a customization language. AXL enables voice-based Web browsing, but without requiring extra effort on the part of the Web page designer. Given any XML document, AXL allows the user to navigate, modify and traverse the structure and links of the document entirely by voice. To illustrate, we focus on how AXL can enable a user to browse the Web via a cellular phone.