Access Patterns Research Articles

PriRanGe: Privacy-Preserving Range-Constrained Intersection Query Over Genomic Data

Genomic data is being produced rapidly by both individuals and enterprises, and outsourcing this ever-increasing data into clouds is promising for cutting the cost of data owners and mining the wealth of genomic data at a larger scale. However, genome carries sensitive information about individuals, and it is challenging to securely and efficiently perform analysis on remotely hosted genomic databases. In this paper, we present a privacy-preserving range-constrained intersection query scheme on genomic data. To achieve security and efficiency, we propose a protocol to fulfill range-constrained intersection query, named PriRanGe. With PriRanGe, a client can securely query genomic data in a specific range in a database while keeping this whole process private. The security of our design targets genomic database confidentiality, query range/result confidentiality, and access pattern protection, and the advantage in efficiency is due to most employed primitives are symmetric. We thoroughly evaluated our design by security proof, experimental analysis and comparison to the state-of-the-art works, all of which support the conclusion that this design is both secure and fast.

Highly-scalable GPU-accelerated compressible reacting flow solver for modeling high-speed flows

Emerging supercomputing systems utilize a combination of central processing units (CPUs) and graphics processing units (GPUs) in an effort to reach exascale capabilities while minimizing the energy footprint of operating such systems. Such heterogeneous machines introduce new challenges for fluids solvers because the hardware architecture and operation of a GPU are fundamentally different from conventional CPUs. In this work, a general approach for efficient implementation of finite-volume based reacting flow solvers on such heterogeneous systems is presented. Three main challenges, namely, data access pattern, thread divergence, and thread safety, are addressed. Since compressible reacting flows require special methods to deal with chemical reactions, hyperbolic and nonlinear convection terms, and the presence of turbulence, specific algorithms that ensure GPU-based efficiency are developed. The approach is demonstrated on the widely available OpenFOAM open source software by modifying core algorithms for GPU accessibility. The scalability of the resulting solver, is demonstrated using practical test cases, including flow through a scramjet engine and the dynamics of a rotating detonation engine. The solver provides near-ideal scaleup on a large number of GPUs (>3000), and extremely efficient use of the GPUs, with throughput nearly a constant even when processing a large number of control volumes.

Accurate Open-Set Recognition for Memory Workload

How can we accurately identify new memory workloads while classifying known memory workloads? Verifying DRAM (Dynamic Random Access Memory) using various workloads is an important task to guarantee the quality of DRAM. A crucial component in the process is open-set recognition which aims to detect new workloads not seen in the training phase. Despite its importance, however, existing open-set recognition methods are unsatisfactory in terms of accuracy since they fail to exploit the characteristics of workload sequences. In this article, we propose Acorn , an accurate open-set recognition method capturing the characteristics of workload sequences. Acorn extracts two types of feature vectors to capture sequential patterns and spatial locality patterns in memory access. Acorn then uses the feature vectors to accurately classify a subsequence into one of the known classes or identify it as the unknown class. Experiments show that Acorn achieves state-of-the-art accuracy, giving up to 37% points higher unknown class detection accuracy while achieving comparable known class classification accuracy than existing methods.

Disaggregating RocksDB: A Production Experience

As in the general industry, there is a trend in Meta's data centers to migrate data from locally attached SSDs to cloud storage. We extended RocksDB [26], a widely used open-source storage engine designed and built for local SSDs, to leverage disaggregated storage. RocksDB's design, such as its data and log files' access patterns, makes an append-only distributed file system a desirable underlying storage. At Meta, we built disaggregated RocksDB using Tectonic File System [35], which so far had mainly been used for our data warehouse and blob storage stacks. We identified that metadata overhead and tail latencies were Tectonic's major performance gaps and addressed them accordingly. We improved the reliability, performance and other requirements with both general and customized optimizations to the core engine in RocksDB. We also took the time to deeply understand the common challenges presented by applications running on RocksDB and implemented enhancements to address them. This architecture enabled RocksDB to adapt to a more distributed architecture for performance enhancements.

FEC: Efficient Deep Recommendation Model Training with Flexible Embedding Communication

Embedding-based deep recommendation models (EDRMs), which contain small dense models and large embedding tables, are widely used in industry. Embedding communication constitutes the main cost for the distributed training of EDRMs, and thus we propose two strategies to improve its efficiency, i.e.,embedding tiering andpre-fetching. In particular, embedding tiering uses AllReduce to communicate popular embeddings that are accessed frequently. This is counter-intuitive as embeddings belong to the sparse embedding tables, but reasonable because the access pattern of popular embeddings resembles dense models. Pre-fetching starts communication early for embeddings that receive no updates such that they are removed from the critical path of training. We implement embedding tiering and pre-fetching in a system called FEC and compare it with the state-of-the-art systems on real datasets. The results show that FEC consistently outperforms the existing methods on all datasets, and its speed can be up to 6.65x and 2.42x in terms of embedding communication time and training throughput compared with the best performing baseline.

Ad Hoc Transactions: What They Are and Why We Should Care

Many transactions in web applications are constructed ad hoc in the application code. For example, developers might explicitly use locking primitives or validation procedures to coordinate critical code fragments. We refer to database operations coordinated by application code as ad hoc transactions. Until now, little is known about them. This paper presents the first comprehensive study on ad hoc transactions. By studying 91 ad hoc transactions among 8 popular open-source web applications, we find that (i) every studied application uses ad hoc transactions (up to 16 per application), 71 of which play critical roles; (ii) compared with database transactions, concurrency control of ad hoc transactions is much more flexible; (iii) ad hoc transactions are error-prone-53 of them have correctness issues, and 33 of them are confirmed by developers; and (iv) ad hoc transactions have the potential to improve performance in contentious workloads by utilizing application semantics such as access patterns. Finally, implications of ad hoc transactions to the database research community are discussed.

Interval Parsing Grammars for File Format Parsing

File formats specify how data is encoded for persistent storage. They cannot be formalized as context-free grammars since their specifications include context-sensitive patterns such as the random access pattern and the type-length-value pattern. We propose a new grammar mechanism called Interval Parsing Grammars IPGs) for file format specifications. An IPG attaches to every nonterminal/terminal an interval, which specifies the range of input the nonterminal/terminal consumes. By connecting intervals and attributes, the context-sensitive patterns in file formats can be well handled. In this paper, we formalize IPGs' syntax as well as its semantics, and its semantics naturally leads to a parser generator that generates a recursive-descent parser from an IPG. In general, IPGs are declarative, modular, and enable termination checking. We have used IPGs to specify a number of file formats including ZIP, ELF, GIF, PE, and part of PDF; we have also evaluated the performance of the generated parsers.

Distributed large-scale graph processing on FPGAs

Processing large-scale graphs is challenging due to the nature of the computation that causes irregular memory access patterns. Managing such irregular accesses may cause significant performance degradation on both CPUs and GPUs. Thus, recent research trends propose graph processing acceleration with Field-Programmable Gate Arrays (FPGA). FPGAs are programmable hardware devices that can be fully customised to perform specific tasks in a highly parallel and efficient manner. However, FPGAs have a limited amount of on-chip memory that cannot fit the entire graph. Due to the limited device memory size, data needs to be repeatedly transferred to and from the FPGA on-chip memory, which makes data transfer time dominate over the computation time. A possible way to overcome the FPGA accelerators’ resource limitation is to engage a multi-FPGA distributed architecture and use an efficient partitioning scheme. Such a scheme aims to increase data locality and minimise communication between different partitions. This work proposes an FPGA processing engine that overlaps, hides and customises all data transfers so that the FPGA accelerator is fully utilised. This engine is integrated into a framework for using FPGA clusters and is able to use an offline partitioning method to facilitate the distribution of large-scale graphs. The proposed framework uses Hadoop at a higher level to map a graph to the underlying hardware platform. The higher layer of computation is responsible for gathering the blocks of data that have been pre-processed and stored on the host’s file system and distribute to a lower layer of computation made of FPGAs. We show how graph partitioning combined with an FPGA architecture will lead to high performance, even when the graph has Millions of vertices and Billions of edges. In the case of the PageRank algorithm, widely used for ranking the importance of nodes in a graph, compared to state-of-the-art CPU and GPU solutions, our implementation is the fastest, achieving a speedup of 13 compared to 8 and 3 respectively. Moreover, in the case of the large-scale graphs, the GPU solution fails due to memory limitations while the CPU solution achieves a speedup of 12 compared to the 26x achieved by our FPGA solution. Other state-of-the-art FPGA solutions are 28 times slower than our proposed solution. When the size of a graph limits the performance of a single FPGA device, our performance model shows that using multi-FPGAs in a distributed system can further improve the performance by about 12x. This highlights our implementation efficiency for large datasets not fitting in the on-chip memory of a hardware device.

Data Distribution for Heterogeneous Storage Systems

The exponential growth of data in many science and engineering domains poses significant challenges to storage systems. Data distribution is a critical component in large-scale distributed storage systems and plays a vital role in placing petabytes of data and beyond, among tens to hundreds of thousands of storage devices. Meantime, heterogeneous storage systems, such as those having devices with hard disk drives (HDDs) and storage class memories (SCMs), have become increasingly popular for massive data storage due to their distinct and complement characteristics. This paper presents a new data distribution algorithm called SUORA (Scalable and Uniform storage via Optimally-adaptive and Random number Addressing) specifically for heterogeneous devices to maximize the benefits of them. SUORA provides a fully symmetric, highly efficient methodology to distribute data across a hybrid and tiered storage cluster. It divides heterogeneous devices into different buckets and segments, and adopts pseudo-random functions to map data onto them with the balanced consideration of capacity, performance and life-time. By analyzing hotness and access patterns, SUORA gradually moves hot data from HDDs to SCMs to optimize the throughput, and moves cold data reversely for load balance. It combines data replication with migration to significantly reduce movement overhead while making data placement more adaptive to different workloads. Extensive evaluations on simulation and Sheepdog storage system show that, with considering distinct characteristics of various devices thoroughly, SUORA improves the overall performance efficiency of heterogeneous storage systems.

On the Effects of Transaction Data Access Patterns on Performance in Lock-Based Concurrency Control

Transaction Processing (TP) plays a primary role in the design and implementation of IT applications and services. Many TP systems exploit lock-based concurrency control to guarantee atomicity and isolation of transactions that access shared data. In this article, we show that transaction data access patterns, in particular the order of data accesses along the transaction execution, have a noticeable impact on how lock-based concurrency control affects performance. We show that the performance can remarkably change depending on whether transactions, or a percentage of them, access data items following some common ordering rule or not. We investigate on this aspect and its root causes through an analytical modeling approach, and with the evidence of data gathered through both simulation and the execution of real transactional workloads. Finally, we show how the findings of our study can be easily exploited for improving the performance of common transactional workloads.

Olive: Oblivious Federated Learning on Trusted Execution Environment against the Risk of Sparsification

Combining Federated Learning (FL) with a Trusted Execution Environment (TEE) is a promising approach for realizing privacy-preserving FL, which has garnered significant academic attention in recent years. Implementing the TEE on the server side enables each round of FL to proceed without exposing the client's gradient information to untrusted servers. This addresses usability gaps in existing secure aggregation schemes as well as utility gaps in differentially private FL. However, to address the issue using a TEE, the vulnerabilities of server-side TEEs need to be considered---this has not been sufficiently investigated in the context of FL. The main technical contribution of this study is the analysis of the vulnerabilities of TEE in FL and the defense. First, we theoretically analyze the leakage of memory access patterns, revealing the risk of sparsified gradients, which are commonly used in FL to enhance communication efficiency and model accuracy. Second, we devise an inference attack to link memory access patterns to sensitive information in the training dataset. Finally, we propose an oblivious yet efficient aggregation algorithm to prevent memory access pattern leakage. Our experiments on real-world data demonstrate that the proposed method functions efficiently in practical scales.

Liberator: A Data Reuse Framework for Out-of-Memory Graph Computing on GPUs

Graph analytics are widely used including recommender systems, scientific computing, and data mining. Meanwhile, GPU has become the major accelerator for such applications. However, the graph size increases rapidly and often exceeds the GPU memory, incurring severe performance degradation due to frequent data transfers between the main memory and GPUs. To relieve this problem, we focus on the utilization of data in GPUs by taking advantage of the data reuse across iterations. In our studies, we deeply analyze the memory access patterns of graph applications at different granularities. We have found that the memory footprint is accessed with a roughly sequential scan without a hotspot, which infers an extremely long reuse distance. Based on our observation, we propose a novel framework, called <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Liberator</i> , to exploit the data reuse within GPU memory. In <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Liberator</i> , GPU memory is reserved for the data potentially accessed across iterations to avoid excessive data transfer between the main memory and GPUs. For the data not existing in GPU memory, a Merged and Aligned memory access manner is employed to improve the transmission efficiency. We also further optimize the framework by parallel processing of data in GPU memory and data in the main memory. We have implemented a prototype of the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Liberator</i> framework and conducted a series of experiments on performance evaluation. The experimental results show that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Liberator</i> can significantly reduce the data transfer overhead, which achieves an average of 2.7x speedup over a state-of-the-art approach.

Malware Detection Using Machine Learning

Abstract: The sophistication of malicious software, known as malware, continues to advance. Previous approaches to detecting malware have predominantly focused on software-based detectors, which are susceptible to compromise. Consequently, recent efforts have suggested the adoption of hardware-assisted malware detection. In this research, we present a fresh framework for hardware-assisted malware detection that utilizes machine learning to monitor and classify patterns of memory access. This framework offers enhanced automation and coverage by reducing the reliance on specific malware signatures from the user. Our work is based on the fundamental understanding that malware must modify control flow and/or data structures, thereby leaving identifiable traces in program memory accesses. Expanding on this insight, we propose an online framework for malware detection that employs machine learning to classify malicious behaviour based on patterns of virtual memory access. Key elements of this framework include techniques for gathering and summarizing memory access patterns at the function and system call levels, as well as a two-level classification architecture

Improving Racial and Ethnic Disparities in Outpatient Anterior Cervical Discectomy and Fusion Driven by Increasing Utilization of Ambulatory Surgical Centers in New York State.

Retrospective cohort study. The purpose of this study was to assess trends in disparities in utilization of hospital outpatient departments (HOPDs) and ambulatory surgical centers (ASCs) for outpatient ACDF (OP-ACDF) between White, Black, Hispanic, and Asian/Pacific Islander patients from 2015 to 2018 in New York State. Racial and ethnic disparities within the field of spine surgery have been thoroughly documented. To date, it remains unknown how these disparities have evolved in the outpatient setting alongside the rapid emergence of ASCs and whether restrictive patterns of access to these outpatient centers exist by race and ethnicity. We conducted a retrospective review from 2015 to 2018 using the Healthcare Cost and Utilization Project (HCUP) New York State Ambulatory Database. Differences in utilization rates for OP-ACDF were assessed and trended over time by race and ethnicity for both HOPDs and freestanding ASCs. Poisson regression was used to evaluate the association between utilization rates for OP-ACDF and race/ethnicity. Between 2015 and 2018, Black, Hispanic, and Asian patients were less likely to undergo OP-ACDF compared with White patients in New York State. However, the magnitude of these disparities lessened over time, as Black, Hispanic, and Asian patients had greater relative increases in utilization of HOPDs and ASCs for ACDF when compared with White patients ( Ptrend <0.001). The magnitude of the increase in freestanding ASC utilization was such that minority patients had higher ACDF utilization rates in freestanding ASCs by 2018 ( P <0.001). We found evidence of improving racial disparities in the relative utilization of outpatient ACDF in New York State. The increase in access to outpatient ACDF appeared to be driven by an increasing number of patients undergoing ACDF in freestanding ASCs in large metropolitan areas. These improving disparities are encouraging and contrast previously documented inequalities in inpatient spine surgery. III.

H2M: Exploiting Heterogeneous Shared Memory Architectures

Over the past decades, the performance gap between the memory subsystem and compute capabilities continued to spread. However, scientific applications and simulations show increasing demand for both memory speed and capacity. To tackle these demands, new technologies such as high-bandwidth memory (HBM) or non-volatile memory (NVM) emerged, which are usually combined with classical DRAM. The resulting architecture is a heterogeneous memory system in which no single memory is “best”. HBM is smaller but offers higher bandwidth than DRAM, whereas NVM provides larger capacity than DRAM at a reasonable cost and less energy consumption. Despite that, in several cases, DRAM still offers the best latency out of all three technologies.In order to use different kinds of memory, applications typically have to be modified to a great extent. Consequently, vendor-agnostic solutions are desirable. First, they should offer the functionality to identify kinds of memory, and second, to allocate data on it. In addition, because memory capacities may be limited, decisions about data placement regarding the different memory kinds have to be made. Finally, in making these decisions, changes over time in data that is accessed, and the actual access pattern, should be considered for initial data placement and be respected in data migration at run-time.In this paper, we introduce a new methodology that aims to provide portable tools and methods for managing data placement in systems with heterogeneous memory. Our approach allows programmers to provide traits (hints) for allocations that describe how data is used and accessed. Combined with characteristics of the platforms’ memory subsystem, these traits are exploited by heuristics to decide where to place data items. We also discuss methodologies for analyzing and identifying memory access characteristics of existing applications, and for recommending allocation traits.In our evaluation, we conduct experiments with several kernels and two proxy applications on Intel Knights Landing (HBM + DRAM) and Intel Ice Lake with Intel Optane DC Persistent Memory (DRAM + NVM) systems. We demonstrate that our methodology can bridge the performance gap between slow and fast memory by applying heuristics for initial data placement.

Accessibility of Primary Schools in Rural Areas and the Impact of Topography: A Case Study in Nanjiang County, China

In recent years, many developing countries have consolidated rural primary schools, closed small community schools, and enlarged centralized schools, which can reduce the accessibility of education to many communities. Meanwhile, expanding road networks may enable people in far-flung communities to access schools more easily. To evaluate the impacts of both trends on spatial justice in access to education, it is important to examine spatial patterns of primary school accessibility and their predictors. How do the topographic features of villages and surrounding landscapes correlate with primary school accessibility in rural upland areas? Using a digital map route planning application, this study evaluates the primary school accessibility of each village in Nanjiang County, a mountainous county in southwest China. By evaluating relationships between primary school accessibility and village characteristics, this study provides evidence corroborating frequent claims that rural remote mountainous areas have poor primary school accessibility. Additionally, by analyzing the effects of elevation and ruggedness of villages and of the zone between villages and schools as well as the mechanisms driving these effects, we find that, contrary to expectations, with increasing village elevation, a village’s primary school accessibility first decreases and then increases. The ruggedness of the terrain upon which a village is built has no significant effect. The ruggedness of the zone between a village and its nearest school exerts significant effects. These findings demonstrate that the two policies have created a pattern of spatial injustice that disadvantages peripheral villages, illustrating the need to attend to topography in efforts to provide equitable school access in rural mountainous areas.

COVID-19 Vaccine Rollout Plans Should Consider Spatial Distribution of Age-Specific Population

In the early phases of COVID-19 vaccine distribution, most countries adopted age-based rollout plans due to vaccine shortages. The plans, however, might overlook different spatial distributions of each age group. In this study, we first examine separate spatial accessibility to vaccine sites varying by different age groups, transportation modes, and times to find differences in spatial patterns of accessibility scores based on the enhanced two-step floating catchment area method. We also investigate spatial inequality in measured accessibility through the Gini coefficient. Finally, we scrutinize to what extent spatial accessibility to vaccine sites is overestimated or underestimated by different age groups. Results showed that there is spatial disparity in accessibility scores between the total population and multiple age-stratified groups. We also found that transportation modes play a significant role in determining spatial patterns of accessibility to vaccine locations, whereas time was not a major factor making differences in spatial accessibility patterns. Our findings suggest that vaccine rollout plans should incorporate age-specific population distribution to maximize accessibility and minimize spatial disparity.

Grep: A Graph Learning Based Database Partitioning System

Database partitioning is a fundamental but challenging task in distributed databases, which selects specific columns as a partitioning key for each table and uses the partitioning key to allocate the table data into different compute nodes in order to maximize the performance. However, this problem is NP-hard and existing distributed databases require users to manually specify the partitioning keys, which may cause potential performance degradation. Although reinforcement learning based methods have been proposed, they have several limitations. First, they do not capture the complex data distributions and query access patterns, and thus involve high computation cost across different compute nodes to answer a query. Second, they involve an expensive step to repetitively partition the data into different compute nodes in order to train a learned key-selection model, which is a waste of time and resources. To address these limitations, we propose a practical learned database partitioning system Grep. We first adopt a graph model to encode data and query features, where vertices are columns, edges are query relations, and the weights of columns are computed based on the localized graph structures (e.g., data diversity, joined columns). We then utilize graph neural networks to embed the partitioning factors into embedding vectors in order to capture the data and query correlations. Next we propose a key-selection model to select appropriate partitioning keys based on the graph model. Finally, we propose an evaluation model to estimate the partitioning performance without actually partitioning the database. We have implemented Grep in a commercial distributed database, and experiments show the effectiveness of our system (e.g., 68% higher throughput for 30K queries in a real banking scenario).

A searchable encryption scheme with hidden search pattern and access pattern on distributed cloud system

A searchable encryption scheme with hidden search pattern and access pattern on distributed cloud system

Analyzing accessibility of car campgrounds through road network structure software in Beijing

The development of self-driving tourism and automobile campgrounds depends on the construction of road infrastructure. In this study, we utilized aeronautical reconnaissance coverage geographic information system (ArcGIS), spatial analysis technology and accessibility model to quantitatively examine the accessibility and spatial pattern of automobile campgrounds in Beijing. Our findings reveal that Beijing's automobile campgrounds exhibit a circular distribution pattern, with decreasing accessibility from the center to the periphery areas. The overall accessibility maintains high-level road directivity, with an average travel time to automobile campgrounds ranging from 0.028 to 3.37 h. Access to automobile campgrounds is higher in the central and southern-central regions compared to the northwest and southwest regions, highlighting a highly unbalanced spatial pattern of accessibility. Finally, we provide suggestions for optimizing the accessibility of automobile campgrounds in Beijing based on our research.