Shared Data Structures Research Articles

Emergent Functional Organization of Gut Microbiomes in Health and Diseases.

Continuous and significant progress in sequencing technologies and bioinformatics pipelines has revolutionized our comprehension of microbial communities, especially for human microbiomes. However, most studies have focused on studying the taxonomic composition of the microbiomes and we are still not able to characterize dysbiosis and unveil the underlying ecological consequences. This study explores the emergent organization of functional abundances and correlations of gut microbiomes in health and disease. Leveraging metagenomic sequences, taxonomic and functional tables are constructed, enabling comparative analysis. First, we show that emergent taxonomic and functional patterns are not useful to characterize dysbiosis. Then, through differential abundance analyses applied to functions, we reveal distinct functional compositions in healthy versus unhealthy microbiomes. In addition, we inquire into the functional correlation structure, revealing significant differences between the healthy and unhealthy groups, which may significantly contribute to understanding dysbiosis. Our study demonstrates that scrutinizing the functional organization in the microbiome provides novel insights into the underlying state of the microbiome. The shared data structure underlying the functional and taxonomic compositions allows for a comprehensive macroecological examination. Our findings not only shed light on dysbiosis, but also underscore the importance of studying functional interrelationships for a nuanced understanding of the dynamics of the microbial community. This research proposes a novel approach, bridging the gap between microbial ecology and functional analyses, promising a deeper understanding of the intricate world of the gut microbiota and its implications for human health.

Breaking Up a Digital Monopoly

How to decompose a vertically integrated digital monopoly to enable competitive services based on a shared data structure.

Managing Concurrent Queues for Efficient In- Vehicle Gateways

In modern vehicles, electrical control units communicate over multiple in-vehicle networks via domainor zone-oriented gateway architectures. This work examines efficient frame transfer across incoming and outgoing Controller Area Network (CAN) interfaces at a gateway. In our embedded platform prototype, each CAN interface of a CAN-to-CAN gateway is controlled independently by a corresponding Portable Operating System Interface (POSIX) thread. Inter-thread communication and synchronization are implemented using shared data structures, either lock-free concurrent queues or traditional lock-based circular queues that embody single-producer single-consumer (SPSC) principles. Our experimental framework provides a realistic open automotive platform that integrates multiple CAN interfaces. An Odroid XU3 device acts as a gateway, providing two Universal Asynchronous Receiver- Transmitter (UART) to CAN interfaces that lead to Raspberry Pi 3B nodes (connected to CAN via Serial Peripheral Interface (SPI)). To increase communication reliability and performance by minimizing frame loss and improving the egress CAN frame rate, low-level parameters at the gateway are optimized, especially the inter-character delay of the UART-to-CAN interfaces. Our experiments show that the frame waiting time in the queue is shorter for the concurrent queue, while relaying messages is equally fast, reaching a maximum output rate of ~380 frames/s for large enqueue rates near saturation. Power consumption is almost four times higher for the concurrent queue implementation.

Probabilistic Approximate Computing at Nanoscales: From data structures to memories

The slowdown of CMOS technology scaling has placed architectures and algorithms on focus for future performance improvements in nanoscale computing systems. Two promising approaches at algorithmic level are approximate computing (AC) and probabilistic data structures (PDSs) that employ the tolerance of an application to small deviations in the results for reducing the complexity of the hardware implementation. AC focuses on applications that process numerical data and relies mostly on approximate (or inexact) low-level arithmetic operations. Instead, PDSs target categorical data and rely on shared data structures and other higher-level simplifications that introduce probabilistic deviations even when all operations are exact. Both AC and PDSs have been able to dramatically reduce the cost in some applications, but they are so far completely disconnected in the application domains, the abstraction levels, and the research communities. In this article, we introduce probabilistic approximate computing (PAC), a new paradigm to use application tolerance for small deviations to reduce the implementation complexity of data structures and hardware when implemented with nanoscale memory technologies. Its goal is to have data structures on which both AC and probabilistic techniques are used in a synergetic way to improve efficiency, while keeping deviations within acceptable margins.

ADT2FHIR - A Tool for Converting ADT/GEKID Oncology Data to HL7 FHIR Resources.

Harmonized and interoperable data management is a core requirement for federated infrastructures in clinical research. Institutions participating in such infrastructures often have to invest large degrees of time and resources in implementing necessary data integration processes to convert their local data to the required target structure. If the data is already available in an alternative shared data structure, the transformation from source to the desired target structure can be implemented once and then be distributed to all participants to reduce effort and harmonize results. The HL7® FHIR® standard is used as a basis for the shared data model of several medical consortia like DKTK and GBA. It is based on so-called resources which can be represented in XML. Oncological data in German university hospitals is commonly available in the ADT/GEKID format. From this common basis we conceptualized and implemented a transformation which accepts ADT/GEKID XML files and returns FHIR resources. We identified several problems with using the general ADT/GEKID structure in federated research infrastructures, as well as some possible pitfalls relating to the FHIR need for resource ids and focus on semantic coding which differs from the approach in the ADT/GEKID standard. To facilitate participation in federated infrastructures, we propose the ADT2FHIR transformation tool for partners with oncological data in the ADT/GEKID format.

Distributed Data Storage and Fusion for Collective Perception in Resource-Limited Mobile Robot Swarms

In this letter, we propose an approach to the distributed storage and fusion of data for collective perception in resource-limited robot swarms. We demonstrate our approach in a distributed semantic classification scenario. We consider a team of mobile robots, in which each robot runs a pre-trained classifier of known accuracy to annotate objects in the environment. We provide two main contributions: (i) a decentralized, shared data structure for efficient storage and retrieval of the semantic annotations, specifically designed for low-resource mobile robots; and (ii) a voting-based, decentralized algorithm to reduce the variance of the calculated annotations in presence of imperfect classification. We discuss theory and implementation of both contributions, and perform an extensive set of realistic simulated experiments to evaluate the performance of our approach.

Detecting Cases of Monoclonal Gammopathy of Undetermined Significance in Electronic Health Data

Introduction: Multiple myeloma (MM) is known to be preceded by the largely asymptomatic and incidentally diagnosed condition monoclonal gammopathy of undetermined significance (MGUS). Retrospective studies estimate that 3% of all US adults aged ≥50 years have this premalignant condition, and that MGUS patients have a lifelong risk of progression to MM of about 1% per year. This persistent uncertainty results in heightened anxiety and clinical surveillance, and the associated financial burden. Recent population-based studies of MGUS have identified cases through testing stored blood specimens, an expensive and inefficient process that hinders initiation of large-scale studies of MGUS etiology and outcome. Thus, to facilitate population-based MGUS research, we are developing an algorithm to efficiently and accurately identify patients diagnosed with MGUS in longitudinally-collected automated healthcare claims and electronic health record (EHR) data. Our aim is to achieve maximum positive predictive value (PPV) with an algorithm that can be applied to large electronic health datasets to accurately identify MGUS cases for longitudinal studies of MGUS etiology, natural history, and health service utilization, minimizing the need for individual chart review.Methods: Men and women aged ≥50 years who sought care at a large, community-based healthcare provider group in central Massachusetts at least once between 2007 and 2015 and were enrolled for ≥12 months were eligible for this analysis. Patients diagnosed with MM prior to, or ≤3 months after the first MGUS diagnosis (ICD-9 code 273.1) were excluded. The MGUS case-finding algorithm was constructed using the Virtual Data Warehouse (VDW) at the Meyers Primary Care Institute. The VDW, which has been adopted by healthcare systems participating in the NCI-funded Cancer Research Network, is a distributed, standardized resource comprised of clinical and administrative data, and populated with linked demographic, administrative, outpatient laboratory, and healthcare utilization data.The first step of the algorithm identified all eligible patients with ≥2 MGUS diagnosis codes entered on different dates within a 12 month period. Next, we selected those patients who, within 90 days of MGUS diagnosis, had: A) ≥1 code for a serum protein electrophoresis test (CPT code 84155 or 84156); B) a serum or urine immunofixation test (CPT 86334 or 86335); and C) an office visit with an oncologist. Relevant protein electrophoresis test results, which typically appear as text in the EHR, were not extractable from available data, and thus patients met the algorithm criteria if the corresponding CPT code was present in their EHR. A targeted manual chart review, conducted by three nurse abstractors and adjudicated by two clinicians, validated the algorithm and constituted the gold standard for subsequent calculations of PPV.Results: The algorithm identified 833 subjects with ≥2 MGUS diagnosis codes in their EHR within a 12 month period; 429 (52%) patients met all four algorithm criteria. Records for 250 of the latter patients (58%) were selected for review. Potential MGUS cases meeting all four algorithm criteria were 85% white, 51% female, had a median of 10 MGUS diagnosis codes in their EHR, and a mean age of 74 years at diagnosis. In the underlying patient population, immunofixation was routinely ordered for patients with abnormal protein electrophoresis test results; thus, 98% of patients who had a protein electrophoresis test also had an immunofixation test, reducing the discriminatory capacity of this algorithm step. Review of the first 2% of charts suggests a preliminary PPV of 75% or greater for the four-step algorithm; chart review is ongoing.Conclusions: Early results from this single-site study suggest that MGUS cases can be identified with reasonable accuracy in electronic health data through a four-step algorithm. However, in light of the frequent use of immunofixation noted above, the incorporation of specific lab test results may be required to improve the PPV. Following completion of ongoing chart review, the algorithm will be applied to a second Cancer Research Network healthcare system with a shared data structure for further refinement, and to evaluate its performance in an independent population. DisclosuresNo relevant conflicts of interest to declare.

Digitization of Assembly Line for Complex Products ‒ The Digital Nursery of Workpiece Digital Twins

Abstract During the last decade the use of digital twins has been spreading to various complex systems, such as assembly lines, and complex products such as aircraft, drones, satellites, vehicles and machinery. This paper discusses the considerations related to the interaction between the assembly-line digital twin and the digital twins of the evolving complex products (workpieces). The paper proposes hierarchical framework including the assembly-line digital twin at the top, that employs the workstation digital twins which are temporarily connected to their current workpiece digital twins. The result is a system that can accurately and efficiently synchronize these simultaneous digital twins. A shared data structure is proposed to facilitate the control and tracking of the assembly progress of each workpiece. This data structure is carried with the workpiece and is shared with the current station digital twin and if necessary the line digital twin.

ADAPT: An auxiliary storage data path toolkit

The legacy storage data path is largely structured in black-box layers and has four major limitations: (1) functional redundancies across layers, (2) poor cross-layer coordination and data tracking, (3) presupposition of high-latency storage devices, and (4) poor support for new storage data models. While addressing all these limitations is a daunting challenge, we introduce ADAPT, an auxiliary storage data path toolkit that complements the legacy storage data path to help mitigate these limitations. This toolkit enables all storage layers to coordinate and track data using shared data structures constructed through the ADAPT API. Our case studies have shown that we can directly support applications such as a key-value store without going through the file system. We also built an ADAPT-based file system and prioritized caching to demonstrate the usability, extensibility, and robustness of ADAPT. In addition, we built per-file secure deletion via our ADAPT-based file system to demonstrate data-path-wide coordination and data tracking.

The parallelism motifs of genomic data analysis.

Genomic datasets are growing dramatically as the cost of sequencing continues to decline and small sequencing devices become available. Enormous community databases store and share these data with the research community, but some of these genomic data analysis problems require large-scale computational platforms to meet both the memory and computational requirements. These applications differ from scientific simulations that dominate the workload on high-end parallel systems today and place different requirements on programming support, software libraries and parallel architectural design. For example, they involve irregular communication patterns such as asynchronous updates to shared data structures. We consider several problems in high-performance genomics analysis, including alignment, profiling, clustering and assembly for both single genomes and metagenomes. We identify some of the common computational patterns or ‘motifs’ that help inform parallelization strategies and compare our motifs to some of the established lists, arguing that at least two key patterns, sorting and hashing, are missing.This article is part of a discussion meeting issue ‘Numerical algorithms for high-performance computational science’.

Hyperqueues

The hyperqueue is a programming abstraction for queues that results in deterministic and scale-free parallel programs. Hyperqueues extend the concept of Cilk++ hyperobjects to provide thread-local views on a shared data structure. While hyperobjects are organized around private local views, hyperqueues provide a shared view on a queue data structure. Hereby, hyperqueues guarantee determinism for programs using concurrent queues. We define the programming API and semantics of two instances of the hyperqueue concept. These hyperqueues differ in their API and the degree of concurrency that is extracted. We describe the implementation of the hyperqueues in a work-stealing scheduler and demonstrate scalable performance on pipeline-parallel benchmarks from PARSEC and StreamIt.

Recoverable mutual exclusion

Mutex locks have traditionally been the most common mechanism for protecting shared data structures in concurrent programs. However, the robustness of such locks against process failures has not been studied thoroughly. The vast majority of mutex algorithms are designed around the assumption that processes are reliable, meaning that a process may not fail while executing the lock acquisition and release code, or while inside the critical section. If such a failure does occur, then the liveness properties of a conventional mutex lock may cease to hold until the application or operating system intervenes by cleaning up the internal structure of the lock. For example, a process that is attempting to acquire an otherwise starvation-free mutex may be blocked forever waiting for a failed process to release the critical section. Adding to the difficulty, if the failed process recovers and attempts to acquire the same mutex again without appropriate cleanup, then the mutex may become corrupted to the point where it loses safety, notably the mutual exclusion property. We address this challenge by formalizing the problem of recoverable mutual exclusion, and proposing several solutions that vary both in their assumptions regarding hardware support for synchronization, and in their efficiency. Compared to known solutions, our algorithms are more robust as they do not restrict where or when a process may crash, and provide stricter guarantees in terms of efficiency, which we define in terms of remote memory references.

Sundials/ML: Connecting OCaml to the Sundials Numeric Solvers

This paper describes the design and implementation of a comprehensive OCaml interface to the Sundials library of numeric solvers for ordinary differential equations, differential algebraic equations, and non-linear equations. The interface provides a convenient and memory-safe alternative to using Sundials directly from C and facilitates application development by integrating with higher-level language features, like garbage-collected memory management, algebraic data types, and exceptions. Our benchmark results suggest that the interface overhead is acceptable: the standard examples are rarely twice as slow in OCaml than in C, and often less than 50% slower. The challenges in interfacing with Sundials are to efficiently and safely share data structures between OCaml and C, to support multiple implementations of vector operations and linear solvers through a common interface, and to manage calls and error signalling to and from OCaml. We explain how we overcame these difficulties using a combination of standard techniques such as phantom types and polymorphic variants, and carefully crafted data representations.

Anomalies and similarities among consensus numbers of variously-relaxed queues

Shared data structures are a basic building block in distributed computing, but can be expensive to implement. One way to circumvent the high implementation cost of linearizability is to relax the sequential specification of the data type. This gives up some guarantees, for instance on the ordering of data elements, as a tradeoff against performance. We want to explore the effects of this tradeoff on the computational power of the shared data structures. In this paper, we characterize the effects of three different types of relaxation, chosen from the literature, on the computational power of FIFO queues. By parametrically relaxing each of the three operations on a queue (Enqueue, Dequeue, Peek), we obtain an infinite 3-dimensional space for each type of relaxation. We find the consensus number, a standard measure of the computational power of shared data types, of each point in these spaces, completely describing the effect of these three types of relaxation on the computational power of queues.

FRC: a high-performance concurrent parallel deferred reference counter for C++

We present FRC, a high-performance concurrent parallel reference counter for unmanaged languages. It is well known that high-performance garbage collectors help developers write memory-safe, highly concurrent systems and data structures. While C++, C, and other unmanaged languages are used in high-performance applications, adding concurrent memory management to these languages has proven to be difficult. Unmanaged languages like C++ use pointers instead of references, and have uncooperative mutators which do not pause easily at a safe point. Thus, scanning mutator stack root references is challenging. FRC only defers decrements and does not require mutator threads to pause during collection. By deferring only decrements, FRC avoids much of the synchronization overhead of a fully-deferred implementation. Root references are scanned without interrupting the mutator by publishing these references to a thread-local array. FRC's performance can exceed that of the C++ standard library's shared pointer by orders of magnitude. FRC's thread-safety guarantees and low synchronization overhead enable significant throughput gains for concurrently-readable shared data structures. We describe the components of FRC, including our static tree router data structure: a novel barrier which improves the scalability of parallel collection workers. FRC's performance is evaluated on several concurrent data structures. We release FRC and our tests as open-source code and expect FRC will be useful for many concurrent C++ software systems.

Relaxed Data Types as Consistency Conditions †

In the quest for higher-performance shared data structures, weakening consistency conditions and relaxing the sequential specifications of data types are two of the primary tools available in the literature today. In this paper, we show that these two approaches are in many cases different ways to specify the same sets of allowed concurrent behaviors of a given shared data object. This equivalence allows us to use whichever description is clearer, simpler, or easier to achieve equivalent guarantees. Specifically, for three common data type relaxations, we define consistency conditions such that the combination of the new consistency condition and an unrelaxed type allows the same behaviors as Linearizability and the relaxed version of the data type. Conversely, for the consistency condition k-Atomicity, we define a new data type relaxation such that the behaviors allowed by the relaxed version of a data type, combined with Linearizability, are the same as those allowed by k-Atomicity and the original type. As an example of the possibilities opened by our new equivalence, we use standard techniques from the literature on consistency conditions to prove that the three data type relaxations we consider are not comparable to one another or to several similar known conditions. Finally, we show a particular class of data types where one of our newly-defined consistency conditions is comparable to, and stronger than, one of the known consistency conditions we consider.

Inter-Controller Traffic to Support Consistency in ONOS Clusters

In distributed SDN architectures, the network is controlled by a cluster of multiple controllers. This distributed approach permits to meet the scalability and reliability requirements of large operational networks. Despite that, a logical centralized view of the network state should be guaranteed, enabling the simple development of network applications. Achieving a consistent network state requires a consensus protocol, which generates control traffic among the controllers whose timely delivery is crucial for network performance. We focus on the state-of-art ONOS controller, designed to scale to large networks, based on a cluster of self-coordinating controllers. In particular, we study the inter-controller control traffic due to the adopted consistency protocols. Based on real traffic measurements and the analysis of the adopted consistency protocols, we develop some empirical models to quantify the traffic exchanged among the controllers, depending on the considered shared data structures, the current network state (e.g., topology) and the occurring network events (e.g., flow or host addition). Our models provide a formal tool to be integrated into the design and dimension the control network interconnecting the controllers. Our results are of paramount importance for the proper design of large SDN networks, in which the control plane is implemented in-band and cannot exploit dedicated network resources.

The role of the inter-controller consensus in the placement of distributed SDN controllers

We consider a distributed Software Defined Networking (SDN) architecture adopting a cluster of controllers to improve network scalability and reliability. Different from previous works that focused solely on the control traffic exchanged between controllers and switches, we additionally consider the control traffic exchanged among the controllers (e.g., the control traffic to keep the shared data structures synchronized). We advocate a careful placement of the controllers, which should take into account both the two kinds of control traffic. We evaluate, for some real ISP network topologies, the delay tradeoffs for the controller placement problem and we propose a novel evolutionary algorithm to find the corresponding Pareto frontier. Furthermore, we develop a simple model to estimate the reaction time perceived by the switches, which is accurately validated in an operational Software Defined WAN (SDWAN). We also formalize the optimization problem to minimize the reaction time and devise a novel approximation algorithm, whose performance is assessed against the optimal solver in real ISP network topologies.Our work provides novel quantitative tools to optimize the planning and the design of the network supporting the control plane of SDN networks, especially when the network is very large and in-band control plane is adopted. We also show that for operationally distributed controllers (e.g., OpenDaylight and ONOS), the location of the controller that acts as leader in the consensus algorithm has a strong impact on the reactivity perceived by the switches.

Parallelization of the multi-level [formula omitted]-adaptive finite cell method

The multi-level hp-refinement scheme is a powerful extension of the finite element method that allows local mesh adaptation without the trouble of constraining hanging nodes. This is achieved through hierarchical high-order overlay meshes, a hp-scheme based on spatial refinement by superposition. An efficient parallelization of this method using standard domain decomposition approaches in combination with ghost elements faces the challenge of a large basis function support resulting from the overlay structure and is in many cases not feasible. In this contribution, a parallelization strategy for the multi-level hp-scheme is presented that is adapted to the scheme’s simple hierarchical structure. By distributing the computational domain among processes on the granularity of the active leaf elements and utilizing shared mesh data structures, good parallel performance is achieved, as redundant computations on ghost elements are avoided. We show the scheme’s parallel scalability for problems with a few hundred elements per process. Furthermore, the scheme is used in conjunction with the finite cell method to perform numerical simulations on domains of complex shape.

Scaling Synchronization in Multicore Programs

Designing software for modern multicore processors poses a dilemma. Traditional software designs, in which threads manipulate shared data, have limited scalability because synchronization of updates to shared data serializes threads and limits parallelism. Alternative distributed software designs, in which threads do not share mutable data, eliminate synchronization and offer better scalability. But distributed designs make it challenging to implement features that shared data structures naturally provide, such as dynamic load balancing and strong consistency guarantees, and are simply not a good fit for every program. Often, however, the performance of shared mutable data structures is limited by the synchronization methods in use today, whether lock-based or lock-free. To help readers make informed design decisions, this article describes advanced (and practical) synchronization methods that can push the performance of designs using shared mutable data to levels that are acceptable to many applications.