nach oben

2019 | Buch

Kapitel lesen Erstes Kapitel lesen

Distributed Applications and Interoperable Systems

19th IFIP WG 6.1 International Conference, DAIS 2019, Held as Part of the 14th International Federated Conference on Distributed Computing Techniques, DisCoTec 2019, Kongens Lyngby, Denmark, June 17–21, 2019, Proceedings

herausgegeben von: José Pereira, Laura Ricci

Verlag: Springer International Publishing

Buchreihe : Lecture Notes in Computer Science

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This book constitutes the proceedings of the 19th IFIP International Conference on Distributed Applications and Interoperable Systems, DAIS 2019, held in Kongens Lyngby, Denmark, in June 2019, as part of the 14th International Federated Conference on Distributed Computing Techniques, DisCoTec 2019.
The 9 full papers presented together with 2 short papers were carefully reviewed and selected from 28 submissions. The papers addressed challenges in multiple application areas, such as the Internet-of-Things, cloud and edge computing, and mobile systems. Some papers focused on middleware for managing concurrency and consistency in distributed systems, including data replication and transactions.

Inhaltsverzeichnis

Frontmatter

Syncpal: A Simple and Iterative Reconciliation Algorithm for File Synchronizers

Abstract

Today file synchronizers are tools often used to facilitate collaboration scenarios and data management across multiple devices. They replicate the file system, e.g. from a cloud storage to a device disk, achieving convergence by only transmitting detected changes. A popular variant available in a plethora of products are state-based file synchronizers such as the Dropbox client. They detect operations by computing the difference between a previously persisted state and the respective current state. However, state-based synchronization is difficult because we need to detect and resolve conflicting operations as well as the propagation order of non-conflicting operations. This work presents Syncpal, an algorithm that reconciles two divergent file systems using an iterative approach. It first handles conflicts, one at a time, making sure that resolving one conflict does not negatively affect other ones, while avoiding conflicts whenever possible. It then finds order dependencies (and breaks cycles) between the remaining non-conflicting operations to avoid the violation of operation preconditions during propagation. This work is relevant for file synchronizer researchers and developers who want to improve their products with an algorithm whose iterative nature reduces the overall complexity and the probability of bugs. In addition to our proposed algorithm and a formal analysis of the underlying problem, our validation approach for the proposed algorithm includes the presentation of a full-scale implementation of an exemplary file system model.

Marius Shekow

Check-Wait-Pounce: Increasing Transactional Data Structure Throughput by Delaying Transactions

Abstract

Transactional data structures allow data structures to support transactional execution, in which a sequence of operations appears to execute atomically. We consider a paradigm in which a transaction commits its changes to the data structure only if all of its operations succeed; if one operation fails, then the transaction aborts. In this work, we introduce an optimization technique called Check-Wait-Pounce that increases performance by avoiding aborts that occur due to failed operations. Check-Wait-Pounce improves upon existing methodologies by delaying the execution of transactions until they are expected to succeed, using a thread-unsafe representation of the data structure as a heuristic. Our evaluation reveals that Check-Wait-Pounce reduces the number of aborts by an average of 49.0%. Because of this reduction in aborts, the tested transactional linked lists achieve average gains in throughput of 2.5x, while some achieve gains as high as 4x.

Lance Lebanoff, Christina Peterson, Damian Dechev

Putting Order in Strong Eventual Consistency

Abstract

Conflict-free replicated data types (CRDTs) aid programmers develop highly available and scalable distributed systems. However, the literature describes only a limited portfolio of conflict-free data types and implementing custom ones requires additional knowledge of replication and consistency techniques. As a result, programmers resort to ad hoc solutions which are error-prone and result in brittle systems. In this paper, we introduce strong eventually consistent replicated objects (SECROs), a general-purpose data type for building available data structures that guarantee strong eventual consistency (SEC) without restrictions on the operations. To evaluate our solution we compare a real-time collaborative text editor built atop SECROs with a state-of-the-art implementation that uses JSON CRDTs. This comparison quantifies various performance aspects. The results show that SECROs are truly general-purpose and memory efficient.

Kevin De Porre, Florian Myter, Christophe De Troyer, Christophe Scholliers, Wolfgang De Meuter, Elisa Gonzalez Boix

Composable Actor Behaviour

Abstract

Code reusability is the cornerstone of object-oriented programming. Reuse mechanisms such as inheritance and trait composition lay at the basis of a whole range of software engineering practices with the goal to improve software quality and reliability. In this paper we investigate code reuse mechanisms for actors, and find that it is currently difficult to specify the behaviour of an actor out of reusable parts. We discuss different kinds of code reuse mechanisms in different kinds of actor model, and we motivate why these mechanisms are currently unsatisfactory. As a possible solution we define a new reuse mechanism based on delegation-based trait composition. In a nutshell, the mechanism allows programmers to compose the behaviour of actors, and every time a compound behaviour is spawned into an actor, it will cause multiple actors to be spawned (one for each independent behaviour). Some messages will be automatically delegated to the actor that implements the desired functionality. We provide an implementation of this model in a prototype Active Object language called Stella, and we formalise a subset of Stella using a small-step operational semantics to unambiguously define the different steps involved in our reuse mechanism.

Sam Van den Vonder, Joeri De Koster, Wolfgang De Meuter

Gossip Learning as a Decentralized Alternative to Federated Learning

Abstract

Federated learning is a distributed machine learning approach for computing models over data collected by edge devices. Most importantly, the data itself is not collected centrally, but a master-worker architecture is applied where a master node performs aggregation and the edge devices are the workers, not unlike the parameter server approach. Gossip learning also assumes that the data remains at the edge devices, but it requires no aggregation server or any central component. In this empirical study, we present a thorough comparison of the two approaches. We examine the aggregated cost of machine learning in both cases, considering also a compression technique applicable in both approaches. We apply a real churn trace as well collected over mobile phones, and we also experiment with different distributions of the training data over the devices. Surprisingly, gossip learning actually outperforms federated learning in all the scenarios where the training data are distributed uniformly over the nodes, and it performs comparably to federated learning overall.

István Hegedűs, Gábor Danner, Márk Jelasity

Using Trusted Execution Environments for Secure Stream Processing of Medical Data

(Case Study Paper)

Abstract

Processing sensitive data, such as those produced by body sensors, on third-party untrusted clouds is particularly challenging without compromising the privacy of the users generating it. Typically, these sensors generate large quantities of continuous data in a streaming fashion. Such vast amount of data must be processed efficiently and securely, even under strong adversarial models. The recent introduction in the mass-market of consumer-grade processors with Trusted Execution Environments (TEEs), such as Intel SGX, paves the way to implement solutions that overcome less flexible approaches, such as those atop homomorphic encryption. We present a secure streaming processing system built on top of Intel SGX to showcase the viability of this approach with a system specifically fitted for medical data. We design and fully implement a prototype system that we evaluate with several realistic datasets. Our experimental results show that the proposed system achieves modest overhead compared to vanilla Spark while offering additional protection guarantees under powerful attackers and threat models.

Carlos Segarra, Ricard Delgado-Gonzalo, Mathieu Lemay, Pierre-Louis Aublin, Peter Pietzuch, Valerio Schiavoni

Stunner: A Smart Phone Trace for Developing Decentralized Edge Systems

Abstract

Conducting research into edge and fog computing often involves experimenting with actual deployments, which is costly and time-consuming, so we need to rely on realistic simulations at least in the early phases of research. To be able to do so we need to collect real data that allows us to perform trace-based simulation and to extract crucial statistics. To achieve this for the domain of distributed smartphone applications, for many years we have been collecting data via smartphones concerning NAT type, the availability of WiFi and cellular networks, the battery level, and many more attributes. Recently, we enhanced our data collecting Android app Stunner by taking actual P2P measurements. Here, we outline our data collection method and the technical details, including some challenges we faced with data cleansing. We present a preliminary set of statistics based on the data for illustration. We also make our new database freely available for research purposes.

Zoltán Szabó, Krisztián Téglás, Árpád Berta, Márk Jelasity, Vilmos Bilicki

FOUGERE: User-Centric Location Privacy in Mobile Crowdsourcing Apps

Abstract

Mobile crowdsourcing is being increasingly used by industrial and research communities to build realistic datasets. By leveraging the capabilities of mobile devices, mobile crowdsourcing apps can be used to track participants’ activity and to collect insightful reports from the environment (e.g., air quality, network quality). However, most of existing crowdsourced datasets systematically tag data samples with time and location stamps, which may inevitably lead to user privacy leaks by discarding sensitive information.

This paper addresses this critical limitation of the state of the art by proposing a software library that improves user privacy without compromising the overall quality of the crowdsourced datasets. We propose a decentralized approach, named Fougere, to convey data samples from user devices to third-party servers. By introducing an a priori data anonymization process, we show that Fougere defeats state-of-the-art location-based privacy attacks with little impact on the quality of crowdsourced datasets.

Lakhdar Meftah, Romain Rouvoy, Isabelle Chrisment

On the Performance of ARM TrustZone

(Practical Experience Report)

Abstract

The TrustZone technology, available in the vast majority of recent Arm processors, allows the execution of code inside a so-called secure world. It effectively provides hardware-isolated areas of the processor for sensitive data and code, i.e., a trusted execution environment (TEE). The Op-Tee framework provides a collection of toolchain, open-source libraries and secure kernel specifically geared to develop applications for TrustZone. This paper presents an in-depth performance- and energy-wise study of TrustZone using the Op-Tee framework, including secure storage and the cost of switching between secure and unsecure worlds, using emulated and hardware measurements.

Julien Amacher, Valerio Schiavoni

CapBAC in Hyperledger Sawtooth

Abstract

In the Internet of Things (IoT) context, the number of connected devices can be too large for a centralised server. This paper focuses on how to enforce authorisation in such a distributed and dynamic environment. The key idea is to use a blockchain-based technology both as a way to maintain a common distributed ledger to store and use access control information, and as a way to enforce Access Control policies in the form of smart contracts. An implementation of an access-control system is presented as a proof of concept: it corresponds to an adaptation of the Capability-based Access Control Model (CapBAC) in the form of a transaction family in Hyperledger Sawtooth. The main claim is that the features and simplicity of CapBAC magnify the usefulness of a blockchain to control the access in the IoT.

Stefano Bistarelli, Claudio Pannacci, Francesco Santini

Developing Secure Services for IoT with OP-TEE: A First Look at Performance and Usability

Abstract

The implementation, deployment and testing of secure services for Internet of Things devices is nowadays still at an early stage. Several frameworks have recently emerged to help developers realize such services, abstracting the complexity of the many types of underlying hardware platforms and software libraries. Assessing the performance and usability of a given framework remains challenging, as they are largely influenced by the application and workload considered, as well as the target hardware. Since 15 years, Arm processors are providing support for TrustZone, a set of security instructions that realize a trusted execution environment inside the processor. Op-Tee is a free-software framework to implement trusted applications and services for TrustZone. In this short paper we show how one can leverage Op-Tee for implementing a secure service (i.e., a key-value store). We deploy and evaluate the performance of this trusted service on common Raspberry Pi hardware platforms.

We report our experimental results with the data store and also compare it against Op-Tee’s built-in secure storage.

Christian Göttel, Pascal Felber, Valerio Schiavoni

Backmatter

Titel: Distributed Applications and Interoperable Systems
herausgegeben von: José Pereira
Laura Ricci
Verlag: Springer International Publishing
Electronic ISBN: 978-3-030-22496-7
Print ISBN: 978-3-030-22495-0
DOI: https://doi.org/10.1007/978-3-030-22496-7