Transiently-powered systems featuring non-volatile memory as well as external peripherals enable the development of new low-power sensor applications. However, as programmers, we are ill-equipped to reason about systems where power failures are the norm rather than the exception. A first challenge consists in being able to capture all the volatile state of the application –external peripherals included– to ensure progress. A second, more fundamental, challenge consists in specifying how power failures may interact with peripheral operations. In this paper, we propose a formal specification of intermittent computing with peripherals, an axiomatic model of interrupt-based checkpointing as well as its proof of correctness, machine-checked in the Coq proof assistant. We state the correctness of the checkpointing mechanism as a trace refinement property between the model and the specification, which accounts for peripheral device operations replays due to power failures. Our proof methodology relies on intermediate oracle semantics to tame the non-determinism of power failures scenarios.

Body area networks (BANs), cloud computing, and machine learning are platforms that can potentially enable advanced healthcare outside the hospital. By applying distributed sensors and drug delivery devices on/in our body and connecting to such communication and decision-making technology, a system for remote diagnostics and therapy is achieved with additional autoregulation capabilities. Challenges with such autarchic on-body healthcare schemes relate to integrity and safety, and interfacing and transduction of electronic signals into biochemical signals, and vice versa. Here, we report a BAN, comprising flexible on-body organic bioelectronic sensors and actuators utilizing two parallel pathways for communication and decision-making. Data, recorded from strain sensors detecting body motion, are both securely transferred to the cloud for machine learning and improved decision-making, and sent through the body using a secure body-coupled communication protocol to auto-actuate delivery of neurotransmitters, all within seconds. We conclude that both highly stable and accurate sensing-from multiple sensors-are needed to enable robust decision making and limit the frequency of retraining. The holistic platform resembles the self-regulatory properties of the nervous system, i.e., the ability to sense, communicate, decide, and react accordingly, thus operating as a digital nervous system.

Transiently-powered devices are a class of small devices powered by energy harvesting. Because such devices are subject to frequent power outages, many recent works propose to checkpoint data residing in volatile RAM into non-volatile RAM. In this article, we propose a new incremental checkpointing mechanism supported by a common hardware component, namely a Memory Protection Unit (MPU). This mechanism leverages the hardware interrupts of the MPU: volatile RAM is read-only on boot and is progressively unlocked as soon as protection violations occur. The MPU interrupt handler is designed to flag the corresponding volatile RAM blocks as dirty, i.e., modified. When a power outage is foreseen to be imminent, the software simply has to copy the dirty blocks from volatile RAM into the non-volatile RAM to ensure application progress over power outages. We validate our approach analytically and in cycle-accurate simulation, and we show that the proposed solution can be easily implemented on real hardware.

Transiently-powered systems featuring non-volatile memory as well as external peripherals enable the development of new low-power sensor applications. However, as programmers, we are ill-equipped to reason about systems where power failures are the norm rather than the exception. A first challenge consists in being able to capture all the volatile state of the application -- external peripherals included -- to ensure progress. A second, more fundamental, challenge consists in specifying how power failures may interact with peripheral operations. In this paper, we propose a formal specification of intermittent computing with peripherals, an axiomatic model of interrupt-based checkpointing as well as its proof of correctness, machine-checked in the Coq proof assistant. We also illustrate our model with several systems proposed in the literature.

We propose a methodology to measure, model and simulate power consumption of peripheral devices of a lowpower embedded micro-controller, while keeping a reasonable development cost.This methodology is experimented against the low-power MSP-EXP430FR5739 platform that includes nonvolatile RAM for intermittent computing purposes and a handful of peripherals.The experimental measurements enable the characterization of the consumption of the peripherals, while many existing comparable studies do not provide power consumption for peripherals.These measurements are integrated into a simulator that targets low-power peripheral-intensive applications, as are most of IoT embedded programs.The accuracy of the power consumption estimation is within a 5% error on intermittent embedded computing using peripherals.

Object stores have become the de-facto platform for storage in the cloud due to their scalability, high availability, and low cost. However, they provide weaker metadata semantics and lower performance compared to distributed hierarchical file systems. In this paper, we introduce HopsFS-S3, a hybrid distributed hierarchical file system backed by an object store while preserving the file system's strong consistency semantics. We base our implementation on HopsFS, a next-generation distribution of HDFS with distributed metadata. We redesigned HopsFS' block storage layer to transparently use an object store to store the file's blocks without sacrificing the file system's semantics. We also introduced a new block caching service to leverage faster NVMe storage for hot blocks. In our experiments, we show that HopsFS-S3 outperforms EMRFS for IO-bound workloads, with up to 20% higher performance and delivers up to 3 AX the aggregated read throughput of EMRFS. Moreover, we demonstrate that metadata operations on HopsFS-S3 (such as directory rename) are up to two orders of magnitude faster than EMRFS. Finally, HopsFS-S3 opens up the currently closed metadata in object stores, enabling correctly-ordered change notifications with HopsFS' change data capture (CDC) API and customized extensions to metadata.

Recently has emerged the concept of transiently powered systems powered by harvesting power and being able to retain information between power failures using non-volatile RAM. While existing solutions focus on purely computing systems, this article presents Sytare, a software layer designed to allow the use of non-trivial peripherals such as timers, serial interface or radio devices in transiently powered systems.

In a near future, energy harvesting is expected to replace batteries in ultra-low-power embedded systems. Research prototypes of such systems have recently been proposed. As the power harvested in the environment is very low, such systems need to cope with frequent power outages. They are referred to as transiently-powered systems (TPS). In order to execute non-trivial applications, TPS need to retain information between power losses. To achieve this goal, emerging non-volatile memory (NVM) technologies are a key enabler: they provide a lightweight solution to retain, between power outages, the state of an application and of its peripheral devices. These include sensors, serial interface or radio devices for instance. Existing works have described various checkpointing mechanisms to adapt embedded applications to TPS but the use of peripherals was not yet handled. in these works. This paper proposes a solution for embedded applications using any peripheral device to run despite transient power. We follow a kernel-oriented approach resulting in minimal impact on the programming model of the application. We implement the new concepts in our lightweight kernel called Sytare, running on an MSP430FR5739 micro-controller and we analyze the cost of the proposed solution.

Les systemes dits a alimentation intermittente sont de petits systemes embarques recuperant l'energie dans leur environnement. A cause de contraintes de taille et de cout, ils subissent de frequentes coupures de courant, mais sont neanmoins capables d'executer un programme logiciel, en sauvegardant les donnees necessaires au calcul dans une memoire non-volatile. Cet article presente une technique permettant a ces systemes d'utiliser des peripheriques non triviaux tels qu'un convertisseur analogique-numerique, une interface serie ou une radio.

Recently has emerged the concept of transiently-powered systems: tiny battery-less embedded systems which harvest energy from their environment. To retain information despite frequent and unpredictable power failures, a transiently-powered system can use non-volatile memory to store checkpoints of program state. However current checkpointing techniques only consider the state of computation (processor, memory) and disregard peripheral state completely. This paper presents a software framework that allows for the use of non-trivial peripherals such as analog-to-digital converters, serial interfaces or radio devices, in transiently-powered systems.

Hadoop is a popular system for storing, managing, and processing large volumes of data, but it has bare-bones internal support for metadata, as metadata is a bottleneck and less means more scalability. The result is a scalable platform with rudimentary access control that is neither user-nor developer friendly. Also, metadata services that are built on Hadoop, such as SQL-on-Hadoop, access control, data provenance, and data governance are necessarily implemented as eventually consistent services, resulting in increased development effort and more brittle software. In this paper, we present a new project-based multi-tenancy model for Hadoop, built on a new distribution of Hadoop that provides a distributed database backend for the Hadoop Distributed Filesystem's (HDFS) metadata layer. We extend Hadoop's metadata model to introduce projects, datasets, and project-users as new core concepts that enable a user-friendly, UI-driven Hadoop experience. As our metadata service is backed by a transactional database, developers can easily extend metadata by adding new tables and ensure the strong consistency of extended metadata using both transactions and foreign keys.

Leader election protocols are a fundamental building block for replicated distributed services. They ease the design of leader-based coordination protocols that tolerate failures. In partially synchronous systems, designing a leader election algorithm, that does not permit multiple leaders while the system is unstable, is a complex task. As a result many production systems use third-party distributed coordination services, such as ZooKeeper and Chubby, to provide a reliable leader election service. However, adding a third-party service such as ZooKeeper to a distributed system incurs additional operational costs and complexity. ZooKeeper instances must be kept running on at least three machines to ensure its high availability. In this paper, we present a novel leader election protocol using NewSQL databases for partially synchronous systems, that ensures at most one leader at any given time. The leader election protocol uses the database as distributed shared memory. Our work enables distributed systems that already use NewSQL databases to save the operational overhead of managing an additional third-party service for leader election. Our main contribution is the design, implementation and validation of a practical leader election algorithm, based on NewSQL databases, that has performance comparable to a leader election implementation using a state-of-the-art distributed coordination service, ZooKeeper.

Enabling anonymous communication over the Internet is crucial. The first protocols that have been devised for anonymous communication are subject to freeriding. Recent protocols have thus been proposed to deal with this issue. However, these protocols do not scale to large systems, and some of them further assume the existence of trusted servers. In this paper, we present RAC, the first anonymous communication protocol that tolerates freeriders and that scales to large systems. Scalability comes from the fact that the complexity of RAC in terms of the number of message exchanges is independent from the number of nodes in the system. Another important aspect of RAC is that it does not rely on any trusted third party. We theoretically prove, using game theory, that our protocol is a Nash equilibrium, i.e, that freeriders have no interest in deviating from the protocol. Further, we experimentally evaluate RAC using simulations. Our evaluation shows that, whatever the size of the system (up to 100.000 nodes), the nodes participating in the system observe the same throughput.

Many uniform total order broadcast protocols have been designed in the last 30 years. Unfortunately, none of them achieves both optimal throughput and low latency. Indeed, protocols achieving optimal throughput rely on a ring dissemination pattern, which induces high latencies. Protocols achieving low latency rely on IP multicast and fail to achieve good throughput because of message losses. In this paper, we describe FastCast, the first protocol that achieves both optimal throughput and low latency. To achieve low latency, FastCast relies on IP multicast. To achieve optimal throughput, FastCast defines a protocol responsible for dynamically computing the throughput at which processes can send IP multicast messages. Thanks to this dynamic bandwidth allocation protocol, FastCast allows multiple processes to simultaneously send messages, while avoiding message losses. An evaluation of FastCast on a cluster of 8 machines shows that it indeed achieves optimal throughput and a very low latency.

Peer-to-peer (P2P) video-on-demand (VoD) requires that nodes collaborate in the downloading of video files as a number of file pieces. In general for VoD, a node is only interested in another node's video file pieces if its download position in the video file precedes the download position of the other node. In this paper, we capture this neighbour relation using the Gradient overlay network, a gossip-generated P2P topology. The Gradient network topology self-organizes into logical concentric rings, such that nodes at earlier download positions in the file are found at increasing distances from the centre, while nodes that have downloaded the whole file are located in the centre of the topology. We build a P2PVoD protocol that uses nodes sampled from the Gradient overlay, and we evaluate its performance in simulation. We also present the layered gossiping architecture and discuss the role of self-organizing mechanisms in the Gradient overlay network's architecture, including positive and negative feedback, decay, and external events from the environment.