Software Stack Research Articles

Multi-Tier GPU Virtualization for Deep Learning in Cloud-Edge Systems

Accelerator virtualization offers several advantages in the context of cloud-edge computing. Relatively weak user devices can enhance performance when running workloads by accessing virtualized accelerators available on other resources in the cloud-edge continuum. However, cloud-edge systems are heterogeneous, often leading to compatibility issues arising from various hardware and software stacks present in the system. One mechanism to alleviate this issue is using containers for deploying workloads. Containers isolate applications and their dependencies and store them as images that can run on any device. In addition, user devices may move during the course of application execution, and thus mechanisms such as container migration are required to move running workloads from one resource to another in the network. Furthermore, an optimal destination will need to be determined when migrating between virtual accelerators. Scheduling and placement strategies are incorporated to choose the best possible location depending on the workload requirements. This paper presents <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AVEC</monospace> , a framework for accelerator virtualization in cloud-edge computing. The AVEC framework enables the offloading of deep learning workloads for inference from weak user devices to computationally more powerful devices in a cloud-edge network. AVEC incorporates a mechanism that efficiently manages and schedules the virtualization of accelerators. It also supports migration between accelerators to enable stateless container migration. The experimental analysis highlights that AVEC can achieve up to 7x speedup by offloading applications to remote resources. Furthermore, AVEC features a low migration downtime that is less than 5 seconds.

IoT Based Blockchains and IPFS

Block chain, the underlying technology of crypto currency networks like Bitcoin, is designed to realize the vision of a decentralized, secure and open Internet of Things (IoT) revolution with a centralized data access model. Essential. It may be so. This article proposes a distributed access model for IoT data using a network architecture called the IoT and Blockchain Modular Consortium Architecture. The proposed architecture facilitates IoT communication via a blockchain software stack and a peer-to-peer data storage mechanism. This architecture is intended to provide privacy and be adaptable to various IoT use cases. To implement the proposed architecture, we conduct a performance analysis of existing blockchain development platforms, Ethereum and Monax.

Abisko: Deep codesign of an architecture for spiking neural networks using novel neuromorphic materials

The Abisko project aims to develop an energy-efficient spiking neural network (SNN) computing architecture and software system capable of autonomous learning and operation. The SNN architecture explores novel neuromorphic devices that are based on resistive-switching materials, such as memristors and electrochemical RAM. Equally important, Abisko uses a deep codesign approach to pursue this goal by engaging experts from across the entire range of disciplines: materials, devices and circuits, architectures and integration, software, and algorithms. The key objectives of our Abisko project are threefold. First, we are designing an energy-optimized high-performance neuromorphic accelerator based on SNNs. This architecture is being designed as a chiplet that can be deployed in contemporary computer architectures and we are investigating novel neuromorphic materials to improve its design. Second, we are concurrently developing a productive software stack for the neuromorphic accelerator that will also be portable to other architectures, such as field-programmable gate arrays and GPUs. Third, we are creating a new deep codesign methodology and framework for developing clear interfaces, requirements, and metrics between each level of abstraction to enable the system design to be explored and implemented interchangeably with execution, measurement, a model, or simulation. As a motivating application for this codesign effort, we target the use of SNNs for an analog event detector for a high-energy physics sensor.

Analysis of Fluoride-Free Content on Twitter: Topic Modeling Study.

Although social media has the potential to spread misinformation, it can also be a valuable tool for elucidating the social factors that contribute to the onset of negative beliefs. As a result, data mining has become a widely used technique in infodemiology and infoveillance research to combat misinformation effects. On the other hand, there is a lack of studies that specifically aim to investigate misinformation about fluoride on Twitter. Web-based individual concerns on the side effects of fluoridated oral care products and tap water stimulate the emergence and propagation of convictions that boost antifluoridation activism. In this sense, a previous content analysis-driven study demonstrated that the term fluoride-free was frequently associated with antifluoridation interests. This study aimed to analyze "fluoride-free" tweets regarding their topics and frequency of publication over time. A total of 21,169 tweets published in English between May 2016 and May 2022 that included the keyword "fluoride-free" were retrieved by the Twitter application programming interface. Latent Dirichlet allocation (LDA) topic modeling was applied to identify the salient terms and topics. The similarity between topics was calculated through an intertopic distance map. Moreover, an investigator manually assessed a sample of tweets depicting each of the most representative word groups that determined specific issues. Lastly, additional data visualization was performed regarding the total count of each topic of fluoride-free record and its relevance over time, using Elastic Stack software. We identified 3 issues by applying the LDA topic modeling: "healthy lifestyle" (topic 1), "consumption of natural/organic oral care products" (topic 2), and "recommendations for using fluoride-free products/measures" (topic 3). Topic 1 was related to users' concerns about leading a healthier lifestyle and the potential impacts of fluoride consumption, including its hypothetical toxicity. Complementarily, topic 2 was associated with users' personal interests and perceptions of consuming natural and organic fluoride-free oral care products, whereas topic 3 was linked to users' recommendations for using fluoride-free products (eg, switching from fluoridated toothpaste to fluoride-free alternatives) and measures (eg, consuming unfluoridated bottled water instead of fluoridated tap water), comprising the propaganda of dental products. Additionally, the count of tweets on fluoride-free content decreased between 2016 and 2019 but increased again from 2020 onward. Public concerns toward a healthy lifestyle, including the adoption of natural and organic cosmetics, seem to be the main motivation of the recent increase of "fluoride-free" tweets, which can be boosted by the propagation of fluoride falsehoods on the web. Therefore, public health authorities, health professionals, and legislators should be aware of the spread of fluoride-free content on social media to create and implement strategies against their potential health damage for the population.

A Comprehensive Performance Comparison of IEEE 802.15.4 DSME and TSCH in a Realistic IoT Scenario for Industrial Applications

In the Industrial Internet of Things (i.e., IIoT), the standardization of open technologies and protocols has achieved seamless data exchange between machines and other physical systems from different manufacturers. At the MAC sublayer, the industry-standard protocols IEEE 802.15.4 Time Slot Channel Hopping (TSCH) and Deterministic and Synchronous Multi-channel Extension (DSME) show promising properties for high adaptability and dynamically changing traffic. However, performance comparison between these MAC protocols rarely has gone beyond a simulation phase. This work presents the results of such a comparison on physically deployed networks using the facilities of the FIT-IoTLab. The evaluation includes fully implementing an IIoT protocol stack based on MQTT in Contiki-NG. It comprises the integration of DSME as part of Contiki-NG’s software stack through OpenDSME, the only publicly available implementation of DSME. Results show that both protocols suit IIoT applications, particularly for data collection. The comparison between TSCH and DSME also includes an evaluation of distributed schedulers for both MAC modes and one autonomous scheduler for TSCH within a UDP protocol stack.

On the software architectures for fog-based secure IOT deployments

In this paper, we examine architectural designs for the support of demanding ad hoc IoT applications, such as industrial and large-scale IoTs. First, we examine the traditional software stack of nodes involved in centralized sensory applications. Then, we propose a highly distributed ad hoc architecture with increased node cooperation. Finally, we propose a secure fog-based hybrid model that offers optimizations with respect to performance and security and which facilitates the development of intelligent localized end-user applications with very strict latency requirements. In the three models that we examine we highlight operations at the routing layer and at the clustering sub-layer.

KVRangeDB: Range Queries for a Hash-based Key–Value Device

Key–value (KV) software has proven useful to a wide variety of applications including analytics, time-series databases, and distributed file systems. To satisfy the requirements of diverse workloads, KV stores have been carefully tailored to best match the performance characteristics of underlying solid-state block devices. Emerging KV storage device is a promising technology for both simplifying the KV software stack and improving the performance of persistent storage-based applications. However, while providing fast, predictable put and get operations, existing KV storage devices do not natively support range queries that are critical to all three types of applications described above. In this article, we present KVRangeDB, a software layer that enables processing range queries for existing hash-based KV solid-state disks (KVSSDs). As an effort to adapt to the performance characteristics of emerging KVSSDs, KVRangeDB implements log-structured merge tree key index that reduces compaction I/O, merges keys when possible, and provides separate caches for indexes and values. We evaluated the KVRangeDB under a set of representative workloads, and compared its performance with two existing database solutions: a Rocksdb variant ported to work with the KVSSD, and Wisckey, a key–value database that is carefully tuned for conventional block devices. On filesystem aging workloads, KVRangeDB outperforms Wisckey by 23.7× in terms of throughput and reduce CPU usage and external write amplifications by 14.3× and 9.8×, respectively.

Flexible system software scheduling for asymmetric multicore systems with PMCSched: A case for Intel Alder Lake

SummaryAsymmetric multicore processors (AMPs) couple high‐performance big cores and power‐efficient small ones, all exposing a shared instruction set architecture to software, but with different microarchitectural features. The energy efficiency benefits of AMPs, together with the general‐purpose nature of the various cores, have led hardware manufacturers to build commercial AMP‐based products, first for the mobile and embedded domains, and more recently, for the desktop market segment, as with the Intel Alder Lake processor family. This trend indicates that AMPs may become a solid and more energy efficient replacement for symmetric multicores in a wide range of application domains. Previous research has demonstrated that the system software can substantially improve scheduling—critical to get the most out of heterogeneous cores—by leveraging hardware facilities that are directly managed by the OS, such as performance monitoring counters, or the recently introduced Intel Thread Director technology. Unfortunately, the OS‐level support enabling access to these hardware facilities may often take a long time to be adopted in operating systems, or may come in forms that make its utilization challenging from specific levels of the system software stack, especially in production systems. To fill this gap, we propose PMCSched, an open‐source framework enabling rapid development and evaluation of custom scheduling‐related support in the Linux kernel. PMCSched greatly simplifies the design and implementation of a wide range of scheduling policies for multicore systems that operate at different system software layers without requiring to patch the kernel. To demonstrate the potential of our framework, we conduct a set of experimental case studies on asymmetry‐aware scheduling for Intel Alder Lake processors.

The NanoFlow Repository.

Extracellular particles (EPs) are the focus of a rapidly growing area of exploration due to the widespread interest in understanding their roles in health and disease. However, despite the general need for EP data sharing and established community standards for data reporting, no standard repository for EP flow cytometry data captures rigor and minimum reporting standards such as those defined by MIFlowCyt-EV (https://doi.org/10.1080/20013078.2020.1713526). We sought to address this unmet need by developing the NanoFlow Repository. We have developed The NanoFlow Repository to provide the first implementation of the MIFlowCyt-EV framework. The NanoFlow Repository is freely available and accessible online at https://genboree.org/nano-ui/. Public datasets can be explored and downloaded at https://genboree.org/nano-ui/ld/datasets. The NanoFlow Repository's backend is built using the Genboree software stack that powers the ClinGen Resource, specifically the Linked Data Hub (LDH), a REST API framework written in Node.js, developed initially to aggregate data within ClinGen (https://ldh.clinicalgenome.org/ldh/ui/about). NanoFlow's LDH (NanoAPI) is available at https://genboree.org/nano-api/srvc. NanoAPI is supported by a Node.js Genboree authentication and authorization service (GbAuth), a graph database called ArangoDB, and an Apache Pulsar message queue (NanoMQ) to manage data inflows into NanoAPI. The website for NanoFlow Repository is built with Vue.js and Node.js (NanoUI) and supports all major browsers.

Design and implementation of I/O performance prediction scheme on HPC systems through large-scale log analysis

Large-scale high performance computing (HPC) systems typically consist of many thousands of CPUs and storage units used by hundreds to thousands of users simultaneously. Applications from large numbers of users have diverse characteristics, such as varying computation, communication, memory, and I/O intensity. A good understanding of the performance characteristics of each user application is important for job scheduling and resource provisioning. Among these performance characteristics, I/O performance is becoming increasingly important as data sizes rapidly increase and large-scale applications, such as simulation and model training, are widely adopted. However, predicting I/O performance is difficult because I/O systems are shared among all users and involve many layers of software and hardware stack, including the application, network interconnect, operating system, file system, and storage devices. Furthermore, updates to these layers and changes in system management policy can significantly alter the I/O behavior of applications and the entire system. To improve the prediction of the I/O performance on HPC systems, we propose integrating information from several different system logs and developing a regression-based approach to predict the I/O performance. Our proposed scheme can dynamically select the most relevant features from the log entries using various feature selection algorithms and scoring functions, and can automatically select the regression algorithm with the best accuracy for the prediction task. The evaluation results show that our proposed scheme can predict the write performance with up to 90% prediction accuracy and the read performance with up to 99% prediction accuracy using the real logs from the Cori supercomputer system at NERSC.

Numerical tools for burning plasmas

The software stack under development within a European coordinated effort on tools for burning plasma modelling is presented. The project is organised as a Task (TSVV Task 10) under the new E-TASC initiative (Litaudon et al 2022 Plasma Phys. Control. Fusion 64 034005). This is a continued effort within the EUROfusion inheriting from the earlier European coordination projects as well as research projects based at various European laboratories. The ongoing work of the TSVV Tasks is supported by the Advanced Computing Hubs. Major projects requiring the high performance computing (HPC) resources are global gyrokinetic codes and global hybrid particle-magnetohydrodynamics (MHD) codes. Also applications using the integrated modelling tools, such as the Energetic-Particle Workflow, based on the ITER Integrated Modelling & Analysis Suite (IMAS), or the code package for modelling radio-frequency heating and fast-ion generation may require intensive computation and a substantial memory footprint. The continual development of these codes both on the physics side and on the HPC side allows us to tackle frontier problems, such as the interaction of turbulence with MHD-type modes in the presence of fast particles. One of the important mandated outcomes of the E-TASC project is the IMAS-enabling of EUROfusion codes and release of the software stack to the EUROfusion community.

Automated Discovery of Container Executables

Linux container technologies such as Docker and Singularity offer encapsulated environments for easy execution of software. In high performance computing, this is especially important for evolving and complex software stacks with conflicting dependencies that must co-exist. Singularity Registry HPC (&ldquo;shpc&rdquo;) was created as an effort to install containers in this environment as modules, seamlessly allowing for typically hidden executables inside containers to be presented to the user as commands, and as such significantly simplifying the user experience. A remaining challenge, however, is deriving the list of important executables in the container. In this work, we present a new modular methodology that allows for discovering new containers in large community sets, deriving container entries with relevant executables therein, and fully automating both recipe generation and updates over time. As an exemplar outcome, we have employed this methodology to add to the Registry over 8,000 containers from the BioContainers community that can be maintained and updated by the software automation. All software is publicly available on the GitHub platform, and can be beneficial to container registries and infrastructure providers for automatically generating container modules, thus lowering the usage entry barrier and improving user experience.

Orchestration of materials science workflows for heterogeneous resources at large scale

In the era of big data, materials science workflows need to handle large-scale data distribution, storage, and computation. Any of these areas can become a performance bottleneck. We present a framework for analyzing internal material structures (e.g., cracks) to mitigate these bottlenecks. We demonstrate the effectiveness of our framework for a workflow performing synchrotron X-ray computed tomography reconstruction and segmentation of a silica-based structure. Our framework provides a cloud-based, cutting-edge solution to challenges such as growing intermediate and output data and heavy resource demands during image reconstruction and segmentation. Specifically, our framework efficiently manages data storage, scaling up compute resources on the cloud. The multi-layer software structure of our framework includes three layers. A top layer uses Jupyter notebooks and serves as the user interface. A middle layer uses Ansible for resource deployment and managing the execution environment. A low layer is dedicated to resource management and provides resource management and job scheduling on heterogeneous nodes (i.e., GPU and CPU). At the core of this layer, Kubernetes supports resource management, and Dask enables large-scale job scheduling for heterogeneous resources. The broader impact of our work is four-fold: through our framework, we hide the complexity of the cloud’s software stack to the user who otherwise is required to have expertise in cloud technologies; we manage job scheduling efficiently and in a scalable manner; we enable resource elasticity and workflow orchestration at a large scale; and we facilitate moving the study of nonporous structures, which has wide applications in engineering and scientific fields, to the cloud. While we demonstrate the capability of our framework for a specific materials science application, it can be adapted for other applications and domains because of its modular, multi-layer architecture.

SDU - software for high-throughput automated data collection at the Swiss Light Source.

Recent advances in automation have fostered the development of unattended data collection services at a handful of synchrotron facilities worldwide. At the Swiss Light Source, the installation of new high-throughput sample changers at all three macromolecular crystallography beamlines and the commissioning of the Fast Fragment and Compound Screening pipeline created a unique opportunity to automate data acquisition. Here, the DA+ microservice software stack upgrades, implementation of an automatic loop-centering service and deployment of the Smart Digital User (SDU) software for unattended data collection are reported. The SDU software is the decision-making software responsible for communications between services, sample and device safety, sample centering, sample alignment with grid based X-ray diffraction and, finally, data collection.

Toward resilient autonomous driving—An experience report on integrating resilience mechanisms into the Apollo autonomous driving software stack

Autonomous driver assistance systems (ADAS) have been progressively pushed to extremes. Today, increasingly sophisticated algorithms, such as deep neural networks, assume responsibility for critical driving functionality, including operating the vehicle at various levels of autonomy. Elaborate obstacle detection, classification, and prediction algorithms, mostly vision-based, trajectory planning, and smooth control algorithms, take over what humans learn until they are permitted to control vehicles and beyond. And even if humans remain in the loop (e.g., to intervene in case of error, as required by autonomy levels 3 and 4), it remains questionable whether distracted human drivers will react appropriately, given the high speed at which vehicles drive and the complex traffic situations they have to cope with. A further pitfall is trusting the whole autonomous driving stack not to fail due to accidental causes and to be robust against cyberattacks of increasing sophistication. In this experience report, we share our findings in retrofitting application-agnostic resilience mechanisms into an existing hardware-/software-stack for autonomous driving—Apollo—as well as where application knowledge helps improve existing resilience algorithms. Our goal is to ultimately decrease the vulnerability of autonomously driving vehicles to accidental faults and attacks, allowing them to absorb and tolerate both, as well as to come out of them at least as secure as before the attack has happened. We demonstrate replication and rejuvenation on the driving stack's Control module and indicate how this resilience can be extended both downwards to the hardware level, as well as upwards to the prediction and planning modules.

ShieLD: Shielding Cross-Zone Communication Within Limited-Resourced IoT Devices Running Vulnerable Software Stack

Securing IoT devices is gaining attention as the security risks associated with these devices increase rapidly. TrustZone-M, a Trusted Execution Environment (TEE) for Cortex-M processors, ensures stronger security within an IoT device by allowing isolated execution of security-critical operations, without trusting the entire software stack. However, TrustZone-M does not guarantee secure cross-world communication between applications in the Normal and Secure worlds. The cryptographic protection of the communication channel is an obvious solution; however, within a low-power IoT device, it incurs high overhead if applied to each cross-world message exchange. We present ShieLD, a framework that enables a secure communication channel between the two TrustZone-M worlds by leveraging the Memory Protection Unit (MPU). ShieLD guarantees confidentiality, integrity and authentication services without requiring any cryptographic operations. We implement and evaluate ShieLD using a Musca-A test chip board with Cortex-M33 that supports TrustZone-M. Our empirical evaluation shows, among other gains, the cross-zone communication protected with ShieLD is <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5 times</i> faster than the conventional crypto-based communication.

System integration based on packing, piping and harness routing automation using graph-based design languages

The implementation of a fully instrumented, automated and simulation-enabled engineering software platform capable of automating the currently still manual model-based systems engineering (MBSE) design process for physical systems architecture generation and optimization in an aircraft wing is presented. The software platform uses graph-based design languages to integrate and entirely automate the mainly manual packing, piping and harness routing design. This design automation and optimization is achieved by a novel software stack of an optimization software coupled with a design compiler. It is shown that through rule-based model generation by a design compiler in the form of a design graph as a central data model, a cross-domain data consistency is achieved. This allows for automated execution and coupling of engineering tasks over several different domains such as packing, piping and routing design to converge to an optimized wing physical architecture design variant in agreement with given predetermined design constraints.

Capturing Request Execution Path for Understanding Service Behavior and Detecting Anomalies Without Code Instrumentation

With the increasing scale and complexity of cloud platforms and big-data analytics platforms, it is becoming more and more challenging to understand and diagnose the processing of a service request across multi-layer software stacks of such platforms. One way that helps to deal with this problem is to accurately capture the complete end-to-end execution path of service requests among all involved components. This paper presents REPTrace, a generic methodology for capturing such execution paths in a transparent fashion. Moreover, this paper demonstrates the effectiveness of REPTrace by presenting how REPTrace can be leveraged for knowledge extraction and anomaly detection on the platforms’ request processing. Our experimental results show that, REPTrace enables capturing a holistic view of the request processing across multiple layers of the platforms (which is missing in official documentation) and discovering important undocumented features of the platforms. Fault injection experiments show execution anomalies are detected with 93% precision and 96% recall with aid of REPTrace.

Uncertainty Management in Dependable and Intelligent Embedded Software

The development of dependable and intelligent embedded systems progresses. However, integrating complex software stacks, machine learning solutions, and high-performance computing devices amplifies the functional and temporal uncertainty sources engineers must manage when developing dependable systems and software.

Dynamic Capacity Service for Improving CXL Pooled Memory Efficiency

Compute Express Link (CXL) pooled memory is gaining attention from the industry as a viable memory disaggregation solution offering memory expansion and alleviating memory overprovisioning. One essential feature for the efficient use of the pooled memory is to dynamically allocate or release memory from the pool based on hosts’ demands. We refer to this feature dynamic capacity service (DCS). This article introduces one of the industry’s first DCS implementation for CXL pooled memory. We demonstrate fully functional DCS by implementing a field-programmable gate array-based CXL pooled memory prototype and full software stacks. Our experiment shows that DCS can substantially improve system memory utilization by dynamically allocating and releasing memory resources on demand. We also present the lessons learned from the DCS implementation.