In the rapidly evolving domain of the Internet of Things (IoT), this study focuses on enhancing the energy efficiency of Deep Learning accelerators implemented on FPGA-based heterogeneous platforms aligned with the principles of sustainable computing. Diverging from the conventional focus on the inference phase, this research introduces innovative optimizations aimed at minimizing the overhead associated with the FPGA configuration phase. Our investigation achieved a remarkable 40.13-fold reduction in configuration energy each time the FPGA is powered on through precise fine-tuning of configuration parameters. Furthermore, the implementation of our Idle-Waiting strategy significantly reduced the overall energy consumption across requests. Under scenarios with regular request periods, the enhanced Idle-Waiting strategy augmented with power-saving methods, outperforms the traditional On-Off strategy in duty-cycle mode for request periods extending up to 499.06 ms. This enhancement is most pronounced at a 40 ms request period, where it increases the system's operational lifetime by a factor of 12.39 within a 4147 J energy budget. Additionally, this paper introduces an adaptive strategy switching approach to manage scenarios with irregular request periods, employing both predefined and learnable threshold methods. This approach is not only more consistently stable than single-strategy methods but also generally outperforms them. Within this approach, our learnable threshold experiences only a 10% performance drop compared to the future-aware strategy and is at least 6% better than using single-strategy methods. These results underscore the significant potential for deploying more energy-efficient and sustainable systems within IoT applications. Future research will explore the application of these power-saving techniques to a broader spectrum of tasks on diverse FPGA platforms.

Long Short-Term Memory Networks (LSTMs) are pivotal in on-device time series analysis for embedded systems, particularly for managing sensor data streams. Yet, their deployment on resource-constrained embedded devices presents notable challenges. In response, we introduce a novel parameterized architecture for LSTM accelerators designed explicitly for embedded Field-Programmable Gate Arrays (FPGAs). Our approach involves strategic design choices, such as employing computationally efficient activation functions and optimizing clock frequency with a pipelined Arithmetic Logic Unit (ALU). These decisions drive our architecture towards enhanced energy efficiency while maintaining adaptability across diverse application scenarios. A key feature of our architecture is its configurable parameters, which allow for tailored optimization through the optional use of Digital Signal Processor Slices for ALUs and the selective implementation of activation functions. Our empirical evaluations conducted on the Spartan-7 XC7S15 FPGA demonstrate the robustness of our methodology, achieving a 2.33× improvement in energy efficiency over previous solutions. Furthermore, our study examines the correlation between memory resource types and energy efficiency across various LSTM model sizes. Impressively, even with a 9× increase in the hidden size of the LSTM cell, our accelerator maintains an energy efficiency of 10.03 GOP/s/W, with only a minor decrease of 14.65%. However, it is critical to note that our current design is not yet optimized for larger FPGA models such as the Spartan-7 XC7S25 and XC7S50. For these models, timing constraints, rather than resource limitations, pose challenges to scaling, highlighting a potential area for future optimization.

In recent years, many studies have been published on blood pressure (BP) estimation from photoplethysmogram (PPG) signals to overcome limitations of current BP devices, such as high risk of complications in invasive methods or limited accuracy, temporal resolution and comfort for cuff-based systems. However, most of these studies suffer from methodological drawbacks regarding data handling, leading to overly positive evaluation of such methods. We present an approach on beat-to-beat blood pressure estimation from raw PPG signals that is evaluated extensively on data of unseen test subjects. The proposed method obtains an mean absolute error ± standard deviation of 8.07 ± 6.86 mmHg for DBP and 8.73 ± 7.36 mmHg for SBP. To increase transparency of the model's decision-making, we examine layer activation of the employed convolutional neural network. Moreover, we analyze the impact of fine tuning for personalization of the model and derive strategies to enhance the personalization process.

The increase in computational and communication power of pervasive devices is also enabling to embed increasing intelligence in devices, there included the capability to act in autonomy, and possibly interacting with each other, in order to achieve specific goals, thus leaving humans out of the decision loop. In this context, the goal of the panel is thus reasoning about the possible implications (technical, ethical, and lega) of assigning great and often critical decision power to pervasive autonomous systems.

This short paper reports the summary of the panel held in the context of the PerCom 2022 conference.

Deep learning (DL) is quickly becoming a core technology to process sensor data from IoT devices. Nowadays, data is usually sent to remote Cloud services where GPU-based ML platforms process it. In contrast, our vision is to have IoT devices with local, on-device DL. This can increase privacy and accessibility as well as reduce processing costs. To do so, specialized hardware accelerators are needed to perform ML operations efficiently without high performance CPUs. In this paper we introduce our approach for a generator for ML hardware accelerators in the IoT. We focus specifically on the inference of ML algorithms for a typical class of IoT use cases, i.e., processing time-series data under real-time constraints. Additionally our optimization techniques rely on the co-design of hardware and ML algorithms to explore the most specialized and efficient corners of the design space. In this work we analyze the requirements for our generator, discuss optimization stages and techniques and show a case study based on an LSTM model for traffic flow prediction.

Abstract Artificial intelligence (AI) is an important part of today’s pervasive computing systems. Still, there is no end-to-end system platform that allows to deploy, update, manage and execute AI models in pervasive systems. We propose such a system platform in this paper. Most importantly, we reuse concepts and techniques from twenty years of pervasive computing research on how to enable runtime adaptation and apply it to AI. This allows to specify adaptive AI models that are able to react to a multitude of dynamic changes, e.g. with respect to available devices, networking conditions, but also application requirements and sensor data sources. Developers can optimise their applications iteratively, starting with a generic setup and refining it step by step towards their specific pervasive computing scenario. To show the applicability of our platform, we apply it to two pervasive use cases and evaluate them, achieving up to four times faster inference and three times lower energy consumption compared to a classical AI deployment.

The size of sensor networks supporting smart cities is ever increasing. Sensor network resiliency becomes vital for critical networks such as emergency response and waste water treatment. One approach is to engineer “self-aware” sensors that can proactively change their component composition in response to changes in work load when critical devices fail. By extension, these devices could anticipate their own termination, such as battery depletion, and offload current tasks onto connected devices. These neighboring devices can then reconfigure themselves to process these tasks, thus avoiding catastrophic network failure. In this article, we compare and contrast two types of self-aware sensors. One set uses Q-learning to develop a policy that guides device reaction to various environmental stimuli, whereas the others use a set of shallow neural networks to select an appropriate reaction. The novelty lies in the use of field programmable gate arrays embedded on the sensors that take into account internal system state, configuration, and learned state-action pairs, which guide device decisions to meet system demands. Experiments show that even relatively simple reward functions develop both Q-learning policies and shallow neural networks that yield positive device behaviors in dynamic environments.

Distributed Hash Tables (DHT) are a common architecture for decentralized data management. However, currently existing DHTs allow every participant to access any index. To build secure applications, we need to control this access. In this paper, we present k-rAC, a fine-grained access control for DHTs. It allows to specify and enforce different access rights for each individual data item in the presence of attackers and up to k subverted participants. Our approach can be tailored towards different thread levels and usage scenarios by adapting k and by choosing one of three authentication mechanisms that we provide.

This demonstration will showcase mixed reality technologies that we developed for a series of public art performances in Vienna in October 2015 in a collaboration of performance artists and researchers. The focus of the demonstration is on natural interaction techniques that can be used intuitively to control an avatar in a virtual 3D world. We combine virtual reality devices with optical location tracking, hand gesture recognition and smart devices. Conference attendees will be able to walk around in a Minecraft world by physically moving in the real world and to perform actions on virtual world items using hand gestures. They can also test our initial system for shared avatar control, in which a user in the real world cooperates with a user in the virtual world. Finally, attendees will have the opportunity to give us feedback about their experience with our system.

Experimental evaluation of MMVE research is mostly based on simulation. To convince industry of the advantages of alternate MMVE architectures requires experimental results obtained from using a commercially deployed MMVE, such as Minecraft. In this paper we describe the transformation of Minecraft into a flexible research platform, that allows researchers to modify Minecraft on different layers and functional areas. As proof-of-concept we replace Minecraft's client/server architecture with a distributed server cluster. The Minecraft virtual environment is distributed amongst the server cluster, allowing for migration of users between nodes of the cluster. This is achieved without modifying the Minecraft client and with minimal modification of the server.

Pervasive applications are designed to support users in their daily lives. In order to provide their services, these applications interact with the environment, i.e. their context. They either adapt themselves as a reaction to context changes, or adapt the context via actuators according to their needs. If multiple applications are executed in the same context, interferences are likely to occur. In this paper, we present COMITY-a framework for interference management in multi-platform pervasive systems. Based on contracts specifying an application's interaction with the context, the framework automatically detects interferences and resolves them through a coordinated application adaptation. We analyze the problem of interference resolution, discuss respective algorithms and extensively evaluate our prototype.

A key challenge for Peer-to-Peer-based Massively Multi-user Virtual Environments is efficient delivery of updates. In contrast to centralized approaches, which can efficiently calculate changes and deliver these, peer-based approaches have to generate updates from arbitrary participating peers. In this paper we present an approach to reduce update messages in Peer-to-Peer-based Massively Multi-user Virtual Environments introducing a distinct class of update events, which we call continuous events. A continuous event results in a series of updates that can be pre-computed. By distributing the event series as a change function, affected peers can calculate and perform updates after receiving a continuous event without further update messages. We present a system design enabling the distribution and computation of continuous events and discuss basic algorithms for continuous event management. We also show simulation results which confirm a substantial reduction of update messages.

Distributed virtual worlds encompass a number of different possible user interactions. These are realized as interaction events that need to be reflected in all user's views in a timely and synchronous fashion. This is difficult because of two aspects: A) user requirements vary for different events and due to B) restricted system resources, mainly network delay. In this paper we present a framework that allows the processing of events using different synchronization protocols depending on the user requirements. We integrate existing protocols into the framework allowing different handling of events.

A major characteristic of pervasive computing applications is their ability to adapt themselves to changing execution environments and physical contexts. In this article, we analyze different kinds of adaptations and introduce a multidimensional classification for them. On this basis, we propose a novel approach for peer-to-peer-based pervasive computing that provides support for the identified classes and integrates them in a multilevel architecture. We give a comprehensive overview of this architecture and its current realization in the Peer-to-Peer Pervasive Computing (3PC) project, discussing what adaptation is realized on each level, how the levels interact with each other, and how the overall system benefits from the integrated treatment of adaptation.

Peer-to-Peer-based Massively Multi-user Virtual Environments (P2P-MMVEs) often use coordinator peers to enhance the system's capabilities. The success of this approach depends on selecting suitable peers as coordinators. However, existing approaches either do not select coordinators based on their suitability or simply assume that techniques for such a selection exist. In this paper we argue that these techniques are not currently available for P2P- MMVEs. We believe that developing such techniques is one of the few remaining challenges for the real- world feasibility of P2P-MMVEs. As a first step towards this goal, we identify and discuss the challenges associated with this research in detail.

Many peer-to-peer (P2P) systems require accurate information about their peer's available bandwidth, e.g., for load balancing. Determining this information is difficult, as a suitable approach must address two challenges. First, it must be able to deal with fluctuating bandwidth. Second, it must incur low overhead to prevent interference with the operation of the P2P system. In this paper we present an approach to estimate the available bandwidth of peers in a P2P system, based on a combination of traffic observation and the strategic injection of traffic into the system. We evaluate our approach and show that it is accurate and responsive in settings with variable bandwidth while resulting in limited interference with the system.

Pervasive applications have been designed to support users in their everyday life. For this purpose, they are able to interact with their physical environment, their context. They are aware of their context and use this information for configuration decisions. Furthermore, they can actively modify the context to meet their user's needs. This leads to new challenges in multi-user environments as applications which are executed simultaneously share a common context and thus can directly impact each other. In this paper we present a framework to coordinate multiple pervasive applications explicitly considering their context-interactivity. We show how application coordination can be integrated in an existing component-based system exemplified by our system PCOM. We conduct measurements for the extended system and discuss the obtained results.

Pervasive computing systems are highly complex distributed systems. Due to their intrinsic heterogeneity and dynamism, applications must reconfigure themselves continuously to adapt to changes in their execution context. In addition, applications influence their context themselves, e.g., by outputting audio. This may lead to conflicts, when applications interfere with each other or disturb users in the vicinity. To handle such conflicts, we propose a dynamic conflict management system that detects and resolves conflicts at runtime. In this paper we focus on how the system can be used by end-users in order to tailor a pervasive system to the userpsilas preferences. By defining individual conflicts in a conflict specification our system ensures that system states which are unacceptable for the user are automatically detected and resolved. We present our approach, discuss our proposed conflict notation and evaluate the users' acceptability of our notation with a short user study.

Massively multi-user virtual environments (MMVEs) are highly interactive systems. They require the propagation of state updates to users with little delay. In this paper we propose a novel update propagation approach for MMVEs that enables such low-latency propagation while offering the scalability needed to support MMVEs with massive user numbers. Our approach combines peer-based and server-based update propagation into a hybrid system. It adapts itself dynamically to the available system resources and the current situation in the virtual world. We describe our approach in detail, evaluate it and discuss further steps towards a low-latency update propagation system.

Peer-to-Peer systems have become a substantial element in computer networking. Distributing the load and splitting complex tasks are only some reasons why many developers have come to adopt this technology. However, all of them face a severe problem at the very beginning: setting up an overlay network, such that other clients can easily join it. With an empty peer cache common bootstrapping methods require some manually triggered actions for discovering a peer on the overlay. We therefore introduce an approach for an automated bootstrapping based on DDNS. In this paper we give detailed information about our protocol and document its efficiency and scalability.

In this demo proposal we present Planet Π4, a Massively Multiplayer Online Game (MMOG) developed to evaluate and compare peer-to-peer-based MMOG systems with scalability in mind. The game requires low network latency and creates frequent game state updates. It has a modular architecture that can be adapted and extended with new functionality. Using this modular design we have developed different peer-to-peer infrastructures. Workshop participants will be able to play the game and compare the versions with each other.

In this demo we present our approach for establishing a communication channel between hosts behind a NAT-based router. To do so, we developed a peer-to-peer based variant of the STUN protocol using so-called superpeers. Using this protocol we determine the used NAT types for the hosts and select a suitable NAT traversal technique dynamically.

Energy efficiency in pervasive computing is crucial for devices operated by battery. To provide energy efficiency we created an energy efficient middleware, called SANDMAN. It selects energy-efficient protocol stacks dynamically and switches idle devices in an energy-efficient sleep mode. In this paper we present an overview of the challenges we met when realizing this approach, possible solutions and lessons learned.

Massively multiplayer online games have become increasingly popular. However, their operation is costly, as game servers must be maintained. To reduce these costs, we aim at providing a communication engine to develop massively multiplayer online games based on a peer-to-peer system. In this paper we analyze the requirements of such a system and present an overview of our current work.

Applications in the pervasive computing domain are challenged by the dynamism in which their execution environment changes, e.g. due to user mobility. As a result, applications have to adapt to changes regarding their required resources. In this paper we present PCOM, a component system for pervasive computing. PCOM offers application programmers a high-level programming abstraction which captures the dependencies between components using contracts. The resulting application architecture is a tree formed by components and their dependencies. PCOM supports automatic adaptation in cases where the execution environment changes to the better or to the worse. User supplied as well as system provided strategies take users out of the control loop while offering flexible adaptation control.

Pervasive computing environments add a multitude of additional devices to our current computing landscapes. Specialized embedded systems provide sensor information about the real world or offer a distinct functionality, e.g. presentation on a "smart wall". Spontaneous networking leads to constantly changing availability of services. This requires middleware support to ease application development. Additionally, we argue that an extensible middleware platform covering small embedded systems to fill-fledged desktop computers is needed. Such a middleware should provide easy-to-use abstractions to access remote services and device-specific capabilities. We present a micro-broker-based approach which meets these requirements by allowing uniform access to device capabilities and services through proxies and the integration of different interoperability protocols. A minimum configuration of the middleware can be executed on embedded systems. Resource-rich execution environments are supported by the extensibility of the middleware.