There has been considerable interest in leveraging GPUs' computational power and high memory bandwidth for analytical database workloads. However, their limited memory capacity remains a fundamental limitation for databases whose sizes far exceed the GPU memory size. This challenge is exacerbated by the slow PCIe data transfer speed, that creates a bottleneck in overall system performance. In this work, we introduce a hybrid CPU-GPU query processing strategy that leverages the distinct strengths of CPU and GPU to alleviate the data transfer bottleneck. Our approach performs highly efficient data filtering on the CPU, which substantially reduces the volume of data transferred to the GPU via PCIe, and offloads compute-intensive operators such as joins to the GPU for further processing. Our evaluation on the TPC-H benchmark at scale factors up to 1000 (1TB), using a single A100 GPU with 80GB memory, demonstrates that our approach can effectively handle datasets significantly larger than the GPU memory size. Moreover, it substantially outperforms a state-of-the-art CPU-only database system in both performance and cost-effectiveness.

Abstract This paper evaluates the suitability of Apache Arrow, Parquet, and ORC as formats for subsumption in an analytical DBMS. We systematically identify and explore the high-level features that are important to support efficient querying in modern OLAP DBMSs and evaluate the ability of each format to support these features. We find that each format has trade-offs that make it more or less suitable for use as a format in a DBMS and identify opportunities to more holistically co-design a unified in-memory and on-disk data representation. Notably, for certain popular machine learning tasks, none of these formats perform optimally, highlighting significant opportunities for advancing format design. Our hope is that this study can be used as a guide for system developers designing and using these formats, as well as provide the community with directions to pursue for improving these common open formats.

We explore the use of Relational Database Manage-ment Systems to reduce technical debt in Machine Learning de-ployments, specifically focusing on in-DBMS prediction serving. We evaluate the performance of ML pipelines in Sklearn and ML.NET against their SQL counterparts executed on MySQL and SQL Server. Results suggest comparable performance when data resides in the database, indicating the potential feasibility of executing prediction processes directly in SQL on DBMSs.

research-article Share on Query Processing on Gaming Consoles Authors: Wei Cui Microsoft Research Asia, CN Microsoft Research Asia, CN 0009-0005-9362-3585View Profile , Qianxi Zhang Microsoft Research Asia, CN Microsoft Research Asia, CN 0000-0002-0646-5365View Profile , Spyros Blanas The Ohio State University, US The Ohio State University, US 0009-0004-2703-7177View Profile , Jesús Camacho-Rodríguez Microsoft, US Microsoft, US 0009-0008-9151-6024View Profile , Brandon Haynes Microsoft Gray Systems Lab, US Microsoft Gray Systems Lab, US 0000-0002-1501-9586View Profile , Yinan Li Microsoft Research, US Microsoft Research, US 0009-0004-5483-2862View Profile , Ravi Ramamurthy Microsoft, USA Microsoft, USA 0000-0002-3484-0038View Profile , Peng Cheng Microsoft Research, CN Microsoft Research, CN 0000-0003-4014-4757View Profile , Rathijit Sen Microsoft, US Microsoft, US 0000-0003-4736-2837View Profile , Matteo Interlandi Microsoft, US Microsoft, US 0000-0002-5756-8321View Profile Authors Info & Claims DaMoN '23: Proceedings of the 19th International Workshop on Data Management on New HardwareJune 2023Pages 86–88https://doi.org/10.1145/3592980.3595313Published:18 June 2023Publication History 0citation191DownloadsMetricsTotal Citations0Total Downloads191Last 12 Months191Last 6 weeks191 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

Introducing new (learned) features into a DBMS requires considerable experimentation and benchmarking to avoid regressions in production (customer) workloads. Using standard benchmarks such as TPC-H and TCH-DS is common practice, but, unfortunately, these do not represent the complexity of real production workloads. To solve this problem, in this demo, we propose a technique that generates a synthetic dataset from query logs and metadata—without touching the original data. The keystone of our approach is to map the data generation as a SAT problem where constraints, such as runtime cardinalities, are extracted from query logs and metadata. We show that our approach can generate representative benchmarks mirroring the performance of the original data without trading off privacy. The demo will guide the attendees through the various steps involved in the data generation and testing process.

GPUs offer massive parallelism and high-bandwidth memory access, making them an attractive option for accelerating data analytics in database systems. However, while modern GPUs possess more resources than ever before (e.g., higher DRAM bandwidth), efficient system implementations and judicious resource allocations for query processing are still necessary for optimal performance. Database systems can save GPU runtime costs through just-enough resource allocation or improve query throughput with concurrent query processing by leveraging new GPU resource-allocation capabilities, such as Multi-Instance GPU (MIG). In this paper, we do a cross-stack performance and resource-utilization analysis of four GPU database systems, including Crystal (the state-of-the-art GPU database, performance-wise) and TQP (the latest entry in the GPU database space). We evaluate the bottlenecks of each system through an in-depth microarchitectural study and identify resource underutilization by leveraging the classic roofline model. Based on the insights gained from our investigation, we propose optimizations for both system implementation and resource allocation, using which we are able to achieve 1.9x lower latency for single-query execution and up to 6.5x throughput improvement for concurrent query execution.

Data pipelines (i.e., converting raw data to features) are critical for machine learning (ML) models, yet their development and management is time-consuming. Feature stores have recently emerged as a new "DBMS-for-ML" with the premise of enabling data scientists and engineers to define and manage their data pipelines. While current feature stores fulfill their promise from a functionality perspective, they are resource-hungry---with ample opportunities for implementing database-style optimizations to enhance their performance. In this paper, we propose a novel set of optimizations specifically targeted for point-in-time join, which is a critical operation in data pipelines. We implement these optimizations on top of Feathr: a widely-used feature store, and evaluate them on use cases from both the TPCx-AI benchmark and real-world online retail scenarios. Our thorough experimental analysis shows that our optimizations can accelerate data pipelines by up to 3× over state-of-the-art baselines.

This paper evaluates the suitability of Apache Arrow, Parquet, and ORC as formats for subsumption in an analytical DBMS. We systematically identify and explore the high-level features that are important to support efficient querying in modern OLAP DBMSs and evaluate the ability of each format to support these features. We find that each format has trade-offs that make it more or less suitable for use as a format in a DBMS and identify opportunities to more holistically co-design a unified in-memory and on-disk data representation. Our hope is that this study can be used as a guide for system developers designing and using these formats, as well as provide the community with directions to pursue for improving these common open formats.

In the past decade, many approaches have been suggested to execute ML workloads on a DBMS. However, most of them have looked at in-DBMS ML from a training perspective, whereas ML inference has been largely overlooked. We think that this is an important gap to fill for two main reasons: (1) in the near future, every application will be infused with some sort of ML capability; (2) behind every web page, application, and enterprise there is a DBMS, whereby in-DBMS inference is an appealing solution both for efficiency (e.g., less data movement), performance (e.g., cross-optimizations between relational operators and ML) and governance. In this article, we study whether DBMSs are a good fit for prediction serving. We introduce a technique for translating trained ML pipelines containing both featurizers (e.g., one-hot encoding) and models (e.g., linear and tree-based models) into SQL queries, and we compare in-DBMS performance against popular ML frameworks such as Sklearn and ml.net. Our experiments show that, when pushed inside a DBMS, trained ML pipelines can have performance comparable to ML frameworks in several scenarios, while they perform quite poorly on text featurization and over (even simple) neural networks.

Prediction queries are widely used across industries to perform advanced analytics and draw insights from data. They include a data processing part (e.g., for joining, filtering, cleaning, featurizing the datasets) and a machine learning (ML) part invoking one or more trained models to perform predictions. These parts have so far been optimized in isolation, leaving significant opportunities for optimization unexplored. We present Raven, a production-ready system for optimizing prediction queries. Raven follows the enterprise architectural trend of collocating data and ML runtimes. It relies on a unified intermediate representation that captures both data and ML operators in a single graph structure to unlock two families of optimizations. First, it employs logical optimizations that pass information between the data part (and the properties of the underlying data) and the ML part to optimize each other. Second, it introduces logical-to-physical transformations that allow operators to be executed on different run-times (relational, ML, and DNN) and hardware (CPU, GPU). Novel data-driven optimizations determine the runtime to be used for each part of the query to achieve optimal performance. Our evaluation shows that Raven is able to improve performance of prediction queries on Apache Spark and SQL Server by up to 13.1x and 330x, respectively. Finally, for complex models where GPU acceleration is beneficial, Raven provides up to 8× speedup compared to state-of-the-art systems.

Modern analytical workloads are highly heterogeneous and massively complex, making generic out of the box query optimizers untenable for many customers and scenarios. As a result, it is important to specialize these optimizers to instances of the workloads. In this paper, we continue a recent line of work in steering a query optimizer towards better plans for a given workload, and make major strides in pushing previous research ideas to production deployment. Along the way we solve several operational challenges including, making steering actions more manageable, keeping the costs of steering within budget, and avoiding unexpected performance regressions in production. Our resulting system, QO-Advisor, essentially externalizes the query planner to a massive offline pipeline for better exploration and specialization. We discuss various aspects of our design and show detailed results over production SCOPE workloads at Microsoft, where the system is currently enabled by default.

We demonstrate Tensor Query Processor (TQP): a query processor that automatically compiles relational operators into tensor programs. By leveraging tensor runtimes such as PyTorch, TQP is able to: (1) integrate with ML tools (e.g., Pandas for data ingestion, Tensorboard for visualization); (2) target different hardware (e.g., CPU, GPU) and software (e.g., browser) backends; and (3) end-to-end accelerate queries containing both relational and ML operators. TQP is generic enough to supports the TPC-H benchmark, and it provides performance that are comparable to, and often better than, that of specialized CPU and GPU query processors.

The huge demand for computation in artificial intelligence (AI) is driving unparalleled investments in hardware and software systems for AI. This leads to an explosion in the number of specialized hardware devices, which are now offered by major cloud vendors. By hiding the low-level complexity through a tensor-based interface, tensor computation runtimes (TCRs) such as PyTorch allow data scientists to efficiently exploit the exciting capabilities offered by the new hardware. In this paper, we explore how database management systems can ride the wave of innovation happening in the AI space. We design, build, and evaluate Tensor Query Processor (TQP): TQP transforms SQL queries into tensor programs and executes them on TCRs. TQP is able to run the full TPC-H benchmark by implementing novel algorithms for relational operators on the tensor routines. At the same time, TQP can support various hardware while only requiring a fraction of the usual development effort. Experiments show that TQP can improve query execution time by up to 10X over specialized CPU- and GPU-only systems. Finally, TQP can accelerate queries mixing ML predictions and SQL end-to-end, and deliver up to 9X speedup over CPU baselines.

Inference of Machine Learning (ML) models, i.e. the process of obtaining predictions from trained models, is often an overlooked problem. Model inference is however one of the main contributors of both technical debt in ML applications and infrastructure complexity. MASQ is a framework able to run inference of ML models directly on DBMSs. MASQ not only averts expensive data movements for those predictive scenarios where data resides on a database, but it also naturally exploits all the "Enterprise-grade" features such as governance, security and auditability which make DBMSs the cornerstone of many businesses. MASQ compiles trained models and ML pipelines implemented in scikit-learn directly into standard SQL: no UDFs nor vendor-specific syntax are used, and therefore queries can be readily executed on any DBMS. In this demo, we will showcase MASQ's capabilities through a GUI allowing attendees to: (1) train ML pipelines composed of data featurizers and ML models; (2) compile the trained pipelines into SQL, and deploy them on different DBMSs (MySQL and SQLServer in the demo); and (3) compare the related performance under different configurations (e.g., the original pipeline on the ML framework against the SQL implementations).

In recent years, there has been tremendous interest in research that applies machine learning to database systems. Being one of the most complex components of a DBMS, query optimizers could benefit from adaptive policies that are learned systematically from the data and the query workload. Recent research has brought up novel ideas towards a learned query optimizer, however these ideas have not been evaluated on a commercial query processor or on large scale, real-world workloads. In this paper, we take the approach used by Marcus et al. in Bao and adapt it to SCOPE, a big data system used internally at Microsoft. Along the way, we solve multiple new challenges: we define how optimizer rules affect final query plans by introducing the concept of a rule signature, we devise a pipeline computing interesting rule configurations for recurring jobs, and we define a new learning problem allowing us to apply such interesting rule configurations to previously unseen jobs. We evaluate the efficacy of the approach on production workloads that include 150K daily jobs. Our results show that alternative rule configurations can generate plans with lower costs, and this can translate to runtime latency savings of 7-30% on average and up to 90% for a non trivial subset of the workload.

The goal of this tutorial is to educate the audience about the state of the art in ML for cloud data systems, both in research and in practice. The tutorial is divided in two parts: the progress, and the path forward. Part I covers the recent successes in deploying machine learning solutions for cloud data systems. We will discuss the practical considerations taken into account and the progress made at various levels. The goal is to compare and contrast the promise of ML for systems with the ground actually covered in industry. Finally, Part II discusses practical issues of machine learning in the enterprise covering the generation of explanations, model debugging, model deployment, model management, constraints on eyes-on data usage and anonymization, and a discussion of the technical debt that can accrue through machine learning and models in the enterprise.

Deep Learning (DL) has created a growing demand for simpler ways to develop complex models and efficient ways to execute them. Thus, a significant effort has gone into frameworks like PyTorch or TensorFlow to support a variety of DL models and run efficiently and seamlessly over heterogeneous and distributed hardware. Since these frameworks will continue improving given the predominance of DL workloads, it is natural to ask what else can be done with them. This is not a trivial question since these frameworks are based on the efficient implementation of tensors, which are well adapted to DL but, in principle, to nothing else. In this paper we explore to what extent Tensor Computation Runtimes (TCRs) can support non-ML data processing applications, so that other use cases can take advantage of the investments made on TCRs. In particular, we are interested in graph processing and relational operators, two use cases very different from ML, in high demand, and complement quite well what TCRs can do today. Building on HUMMINGBIRD, a recent platform converting traditional machine learning algorithms to tensor computations, we explore how to map selected graph processing and relational operator algorithms into tensor computations. Our vision is supported by the results: our code often outperforms custom-built C++ and CUDA kernels, while massively reducing the development effort, taking advantage of the cross-platform compilation capabilities of TCRs.

While deep neural networks (DNNs) have shown to be successful in several domains like computer vision, non-DNN models such as linear models and gradient boosting trees are still considered state-of-the-art over tabular data. When using these models, data scientists often author machine learning (ML) pipelines: DAG of ML operators comprising data transforms and ML models, whereby each operator is sequentially trained one-at-a-time. Conversely, when training DNNs, layers composing the neural networks are simultaneously trained using backpropagation. In this paper, we argue that the training scheme of ML pipelines is sub-optimal because it tries to optimize a single operator at a time thus losing the chance of global optimization. We therefore propose WindTunnel: a system that translates a trained ML pipeline into a pipeline of neural network modules and jointly optimizes the modules using backpropagation. We also suggest translation methodologies for several non-differentiable operators such as gradient boosting trees and categorical feature encoders. Our experiments show that fine-tuning of the translated WindTunnel pipelines is a promising technique able to increase the final accuracy.

Easy-to-use programming interfaces paired with cloud-scale processing engines have enabled big data system users to author arbitrarily complex analytical jobs over massive volumes of data. However, as the complexity and scale of analytical jobs increase, they encounter a number of unforeseen problems, hotspots with large intermediate data on temporary storage, longer job recovery time after failures, and worse query optimizer estimates being examples of issues that we are facing at Microsoft. To address these issues, we propose Phoebe, an efficient learning-based checkpoint optimizer. Given a set of constraints and an objective function at compile-time, Phoebe is able to determine the decomposition of job plans, and the optimal set of checkpoints to preserve their outputs to durable global storage. Phoebe consists of three machine learning predictors and one optimization module. For each stage of a job, Phoebe makes accurate predictions for: (1) the execution time, (2) the output size, and (3) the start/end time taking into account the inter-stage dependencies. Using these predictions, we formulate checkpoint optimization as an integer programming problem and propose a scalable heuristic algorithm that meets the latency requirement of the production environment. We demonstrate the effectiveness of Phoebe in production workloads, and show that we can free the temporary storage on hotspots by more than 70% and restart failed jobs 68% faster on average with minimum performance impact. Phoebe also illustrates that adding multiple sets of checkpoints is not cost-efficient, which dramatically reduces the complexity of the optimization.

Continuous integration (CI) has been a de facto standard for building industrial-strength software. Yet, there is little attention towards applying CI to the development of machine learning (ML) applications until the very recent effort on the theoretical side. In this paper, we take a step forward to bring the theory into practice.

Many data analysis and knowledge mining tasks require a basic understanding of the content of a dataset prior to any data access. In this demo, we showcase how data descriptions---a set of compact, readable and insightful formulas of boolean predicates---can be used to guide users in understanding datasets. Finding the best description for a dataset is, unfortunately, both computationally hard and task-specific. This demo shows that not only we can generate descriptions at interactive speed, but also that diverse user needs---from anomaly detection to data exploration---can be accommodated through a user-driven process exploiting dynamic programming in concert with a set of heuristics.

Machine Learning is transitioning from an art and science into a technology available to every developer. In the near future, every application on every platform will incorporate trained models to encode data-based decisions that would be impossible for developers to author. This presents a significant engineering challenge, since currently data science and modeling are largely decoupled from standard software development processes. This separation makes incorporating machine learning capabilities inside applications unnecessarily costly and difficult, and furthermore discourage developers from embracing ML in first place. In this paper we present ML.NET, a framework developed at Microsoft over the last decade in response to the challenge of making it easy to ship machine learning models in large software applications. We present its architecture, and illuminate the application demands that shaped it. Specifically, we introduce DataView, the core data abstraction of ML.NET which allows it to capture full predictive pipelines efficiently and consistently across training and inference lifecycles. We close the paper with a surprisingly favorable performance study of ML.NET compared to more recent entrants, and a discussion of some lessons learned.

To process massive quantities of data, developers leverage Data-Intensive Scalable Computing (DISC) systems such as Apache Spark. In terms of debugging, DISC systems support only post-mortem log analysis and do not provide any debugging functionality. This demonstration paper showcases BigDebug: a tool enhancing Apache Spark with a set of interactive debugging features that can help users in debug their Big Data Applications.

There is great interest in exploiting the opportunity provided by cloud computing platforms for large-scale analytics. Among these platforms, Apache Spark is growing in popularity for machine learning and graph analytics. Developing efficient complex analytics in Spark requires deep understanding of both the algorithm at hand and the Spark API or subsystem APIs (e.g., Spark SQL, GraphX). Our BigDatalog system addresses the problem by providing concise declarative specification of complex queries amenable to efficient evaluation. Towards this goal, we propose compilation and optimization techniques that tackle the important problem of efficiently supporting recursion in Spark. We perform an experimental comparison with other state-of-the-art large-scale Datalog systems and verify the efficacy of our techniques and effectiveness of Spark in supporting Datalog-based analytics.

One notoriously hard data cleaning problem is, given a database, how to precisely capture which value is correct (i.e., proof positive) or wrong (i.e., proof negative). Although integrity constraints have been widely studied to capture data errors as violations, the accuracy of data cleaning using integrity constraints has long been controversial. Overall they deem one fundamental problem: Given a set of data values that together forms a violation, there is no evidence of which value is proof positive or negative. Hence, it is known that integrity constraints themselves cannot guide dependable data cleaning. In this work, we introduce an automated method for proof positive and negative in data cleaning, based on Sherlock rules and reference tables. Given a tuple and reference tables, Sherlock rules tell us what attributes are proof positive, what attributes are proof negative and (possibly) how to update them. We study several fundamental problems associated with Sherlock rules. We also present efficient algorithms for cleaning data using Sherlock rules. We experimentally demonstrate that our techniques can not only annotate data with proof positive and negative, but also repair data when enough information is available.

Debugging data processing logic in Data-Intensive Scalable Computing (DISC) systems is a difficult and time consuming effort. Today's DISC systems offer very little tooling for debugging programs, and as a result programmers spend countless hours collecting evidence ( e.g. , from log files) and performing trial and error debugging. To aid this effort, we built Titian , a library that enables data provenance ---tracking data through transformations---in Apache Spark. Data scientists using the Titian Spark extension will be able to quickly identify the input data at the root cause of a potential bug or outlier result. Titian is built directly into the Spark platform and offers data provenance support at interactive speeds---orders-of-magnitude faster than alternative solutions---while minimally impacting Spark job performance; observed overheads for capturing data lineage rarely exceed 30% above the baseline job execution time.

We showcase QUEST (QUEry generator for STructured sources), a search engine for relational databases that combines semantic and machine learning techniques for transforming keyword queries into meaningful SQL queries. The search engine relies on two approaches: the forward, providing mappings of keywords into database terms (names of tables and attributes, and domains of attributes), and the backward, computing the paths joining the data structures identified in the forward step. The results provided by the two approaches are combined within a probabilistic framework based on the Dempster-Shafer Theory. We demonstrate QUEST capabilities, and we show how, thanks to the flexibility obtained by the probabilistic combination of different techniques, QUEST is able to compute high quality results even with few training data and/or with hidden data sources such as those found in the Deep Web.

The adoption of business intelligence technology in industries is growing rapidly. Business managers are not satisfied with ad hoc and static reports and they ask for more flexible and easy to use data analysis tools. Recently, application interfaces that expand the range of operations available to the user, hiding the underlying complexity, have been developed. The paper presents eLog, a business intelligence solution designed and developed in collaboration between the database group of the University of Modena and Reggio Emilia and eBilling, an Italian SME supplier of solutions for the design, production and automation of documentary processes for top Italian companies. eLog enables business managers to define OLAP reports by means of a web interface and to customize analysis indicators adopting a simple meta-language. The framework translates the user's reports into MDX queries and is able to automatically select the data cube suitable for each query. Over 140 medium and large companies have exploited the technological services of eBilling S.p.A. to manage their documents flows. In particular, eLog services have been used by the major media and telecommunications Italian companies and their foreign annex, such as Sky, Media set, H3G, Tim Brazil etc. The largest customer can provide up to 30 millions mail pieces within 6 months (about 200 GB of data in the relational DBMS). In a period of 18 months, eLog could reach 150 millions mail pieces (1 TB of data) to handle.