arXiv subjects

Carlo Curino

Publications and source records attributed to Carlo Curino.

At least 19 recordsLinked to original sources

arXiv2025

GPU Acceleration of SQL Analytics on Compressed Data

GPUs are uniquely suited to accelerate (SQL) analytics workloads thanks to their massive compute parallelism and High Bandwidth Memory (HBM) -- when datasets fit in the GPU HBM, performance is unparalleled. Unfortunately, GPU HBMs remain typically small when compared with lower-bandwidth CPU main memory. Besides brute-force scaling across many GPUs, current solutions to accelerate queries on large datasets include leveraging data partitioning and loading smaller data batches in GPU HBM, and hybrid execution with a connected device (e.g., CPUs). Unfortunately, these approaches are exposed to the limitations of lower main memory and host-to-device interconnect bandwidths, introduce additional I/O overheads, or incur higher costs. This is a substantial problem when trying to scale adoption of GPUs on larger datasets. Data compression can alleviate this bottleneck, but to avoid paying for costly decompression/decoding, an ideal solution must include computation primitives to operate directly on data in compressed form. This is the focus of our paper: a set of new methods for running queries directly on light-weight compressed data using schemes such as Run-Length Encoding (RLE), index encoding, bit-width reductions, and dictionary encoding. Our novelty includes operating on multiple RLE columns without decompression, handling heterogeneous column encodings, and leveraging PyTorch tensor operations for portability across devices. Experimental evaluations show speedups of an order of magnitude compared to state-of-the-art commercial CPU-only analytics systems, for real-world queries on a production dataset that would not fit into GPU memory uncompressed. This work paves the road for GPU adoption in a much broader set of use cases, and it is complementary to most other scale-out or fallback mechanisms.

cs.DB↗

arXiv2025

Terabyte-Scale Analytics in the Blink of an Eye

For the past two decades, the DB community has devoted substantial research to take advantage of cheap clusters of machines for distributed data analytics -- we believe that we are at the beginning of a paradigm shift. The scaling laws and popularity of AI models lead to the deployment of incredibly powerful GPU clusters in commercial data centers. Compared to CPU-only solutions, these clusters deliver impressive improvements in per-node compute, memory bandwidth, and inter-node interconnect performance. In this paper, we study the problem of scaling analytical SQL queries on distributed clusters of GPUs, with the stated goal of establishing an upper bound on the likely performance gains. To do so, we build a prototype designed to maximize performance by leveraging ML/HPC best practices, such as group communication primitives for cross-device data movements. This allows us to conduct thorough performance experimentation to point our community towards a massive performance opportunity of at least 60$\times$. To make these gains more relatable, before you can blink twice, our system can run all 22 queries of TPC-H at a 1TB scale factor!

cs.DB↗

arXiv2025

MotherNet: Fast Training and Inference via Hyper-Network Transformers

Foundation models are transforming machine learning across many modalities, with in-context learning replacing classical model training. Recent work on tabular data hints at a similar opportunity to build foundation models for classification for numerical data. However, existing meta-learning approaches can not compete with tree-based methods in terms of inference time. In this paper, we propose MotherNet, a hypernetwork architecture trained on synthetic classification tasks that, once prompted with a never-seen-before training set generates the weights of a trained ``child'' neural-network by in-context learning using a single forward pass. In contrast to most existing hypernetworks that are usually trained for relatively constrained multi-task settings, MotherNet can create models for multiclass classification on arbitrary tabular datasets without any dataset specific gradient descent. The child network generated by MotherNet outperforms neural networks trained using gradient descent on small datasets, and is comparable to predictions by TabPFN and standard ML methods like Gradient Boosting. Unlike a direct application of TabPFN, MotherNet generated networks are highly efficient at inference time. We also demonstrate that HyperFast is unable to perform effective in-context learning on small datasets, and heavily relies on dataset specific fine-tuning and hyper-parameter tuning, while MotherNet requires no fine-tuning or per-dataset hyper-parameters.

cs.LG↗

arXiv2025

AutoComp: Automated Data Compaction for Log-Structured Tables in Data Lakes

The proliferation of small files in data lakes poses significant challenges, including degraded query performance, increased storage costs, and scalability bottlenecks in distributed storage systems. Log-structured table formats (LSTs) such as Delta Lake, Apache Iceberg, and Apache Hudi exacerbate this issue due to their append-only write patterns and metadata-intensive operations. While compaction--the process of consolidating small files into fewer, larger files--is a common solution, existing automation mechanisms often lack the flexibility and scalability to adapt to diverse workloads and system requirements while balancing the trade-offs between compaction benefits and costs. In this paper, we present AutoComp, a scalable framework for automatic data compaction tailored to the needs of modern data lakes. Drawing on deployment experience at LinkedIn, we analyze the operational impact of small file proliferation, establish key requirements for effective automatic compaction, and demonstrate how AutoComp addresses these challenges. Our evaluation, conducted using synthetic benchmarks and production environments via integration with OpenHouse--a control plane for catalog management, schema governance, and data services--shows significant improvements in file count reduction and query performance. We believe AutoComp's built-in extensibility provides a robust foundation for evolving compaction systems, facilitating future integration of refined multi-objective optimization approaches, workload-aware compaction strategies, and expanded support for broader data layout optimizations.

cs.DB↗

arXiv2025

Optimizing open-domain question answering with graph-based retrieval augmented generation

In this work, we benchmark various graph-based retrieval-augmented generation (RAG) systems across a broad spectrum of query types, including OLTP-style (fact-based) and OLAP-style (thematic) queries, to address the complex demands of open-domain question answering (QA). Traditional RAG methods often fall short in handling nuanced, multi-document synthesis tasks. By structuring knowledge as graphs, we can facilitate the retrieval of context that captures greater semantic depth and enhances language model operations. We explore graph-based RAG methodologies and introduce TREX, a novel, cost-effective alternative that combines graph-based and vector-based retrieval techniques. Our benchmarking across four diverse datasets highlights the strengths of different RAG methodologies, demonstrates TREX's ability to handle multiple open-domain QA types, and reveals the limitations of current evaluation methods. In a real-world technical support case study, we demonstrate how TREX solutions can surpass conventional vector-based RAG in efficiently synthesizing data from heterogeneous sources. Our findings underscore the potential of augmenting large language models with advanced retrieval and orchestration capabilities, advancing scalable, graph-based AI solutions.

cs.IR↗

arXiv2024

Towards Query Optimizer as a Service (QOaaS) in a Unified LakeHouse Ecosystem: Can One QO Rule Them All?

Customer demand, regulatory pressure, and engineering efficiency are the driving forces behind the industry-wide trend of moving from siloed engines and services that are optimized in isolation to highly integrated solutions. This is confirmed by the wide adoption of open formats, shared component libraries, and the meteoric success of integrated data lake experiences such as Microsoft Fabric. In this paper, we study the implications of this trend to Query Optimizer (QO) and discuss our experience of building Calcite and extending Cascades into QO components of Microsoft SQL Server, Fabric Data Warehouse (DW), and SCOPE. We weigh the pros and cons of a drastic change in direction: moving from bespoke QOs or library-sharing (à la Calcite) to rewriting the QO stack and fully embracing Query Optimizer as a Service (QOaaS). We report on some early successes and stumbles as we explore these ideas with prototypes compatible with Fabric DW and Spark. The benefits include centralized workload-level optimizations, multi-engine federation, and accelerated feature creation, but the challenges are equally daunting. We plan to engage CIDR audience in a debate on this exciting topic.

cs.DB↗

arXiv2024

Towards Building Autonomous Data Services on Azure

Modern cloud has turned data services into easily accessible commodities. With just a few clicks, users are now able to access a catalog of data processing systems for a wide range of tasks. However, the cloud brings in both complexity and opportunity. While cloud users can quickly start an application by using various data services, it can be difficult to configure and optimize these services to gain the most value from them. For cloud providers, managing every aspect of an ever-increasing set of data services, while meeting customer SLAs and minimizing operational cost is becoming more challenging. Cloud technology enables the collection of significant amounts of workload traces and system telemetry. With the progress in data science (DS) and machine learning (ML), it is feasible and desirable to utilize a data-driven, ML-based approach to automate various aspects of data services, resulting in the creation of autonomous data services. This paper presents our perspectives and insights on creating autonomous data services on Azure. It also covers the future endeavors we plan to undertake and unresolved issues that still need attention.

cs.DC↗

arXiv2024

LST-Bench: Benchmarking Log-Structured Tables in the Cloud

Data processing engines increasingly leverage distributed file systems for scalable, cost-effective storage. While the Apache Parquet columnar format has become a popular choice for data storage and retrieval, the immutability of Parquet files renders it impractical to meet the demands of frequent updates in contemporary analytical workloads. Log-Structured Tables (LSTs), such as Delta Lake, Apache Iceberg, and Apache Hudi, offer an alternative for scenarios requiring data mutability, providing a balance between efficient updates and the benefits of columnar storage. They provide features like transactions, time-travel, and schema evolution, enhancing usability and enabling access from multiple engines. Moreover, engines like Apache Spark and Trino can be configured to leverage the optimizations and controls offered by LSTs to meet specific business needs. Conventional benchmarks and tools are inadequate for evaluating the transformative changes in the storage layer resulting from these advancements, as they do not allow us to measure the impact of design and optimization choices in this new setting. In this paper, we propose a novel benchmarking approach and metrics that build upon existing benchmarks, aiming to systematically assess LSTs. We develop a framework, LST-Bench, which facilitates effective exploration and evaluation of the collaborative functioning of LSTs and data processing engines through tailored benchmark packages. A package is a mix of use patterns reflecting a target workload; LST-Bench makes it easy to define a wide range of use patterns and combine them into a package, and we include a baseline package for completeness. Our assessment demonstrates the effectiveness of our framework and benchmark packages in extracting valuable insights across diverse environments. The code for LST-Bench is open-sourced and is available at https://github.com/microsoft/lst-bench/ .

cs.DB↗

arXiv2024

XTable in Action: Seamless Interoperability in Data Lakes

Contemporary approaches to data management are increasingly relying on unified analytics and AI platforms to foster collaboration, interoperability, seamless access to reliable data, and high performance. Data Lakes featuring open standard table formats such as Delta Lake, Apache Hudi, and Apache Iceberg are central components of these data architectures. Choosing the right format for managing a table is crucial for achieving the objectives mentioned above. The challenge lies in selecting the best format, a task that is onerous and can yield temporary results, as the ideal choice may shift over time with data growth, evolving workloads, and the competitive development of table formats and processing engines. Moreover, restricting data access to a single format can hinder data sharing resulting in diminished business value over the long term. The ability to seamlessly interoperate between formats and with negligible overhead can effectively address these challenges. Our solution in this direction is an innovative omni-directional translator, XTable, that facilitates writing data in one format and reading it in any format, thus achieving the desired format interoperability. In this work, we demonstrate the effectiveness of XTable through application scenarios inspired by real-world use cases.

cs.DB↗

arXiv2024

Sibyl: Forecasting Time-Evolving Query Workloads

Database systems often rely on historical query traces to perform workload-based performance tuning. However, real production workloads are time-evolving, making historical queries ineffective for optimizing future workloads. To address this challenge, we propose SIBYL, an end-to-end machine learning-based framework that accurately forecasts a sequence of future queries, with the entire query statements, in various prediction windows. Drawing insights from real-workloads, we propose template-based featurization techniques and develop a stacked-LSTM with an encoder-decoder architecture for accurate forecasting of query workloads. We also develop techniques to improve forecasting accuracy over large prediction windows and achieve high scalability over large workloads with high variability in arrival rates of queries. Finally, we propose techniques to handle workload drifts. Our evaluation on four real workloads demonstrates that SIBYL can forecast workloads with an $87.3\%$ median F1 score, and can result in $1.7\times$ and $1.3\times$ performance improvement when applied to materialized view selection and index selection applications, respectively.

cs.DB↗

arXiv2023

OneProvenance: Efficient Extraction of Dynamic Coarse-Grained Provenance from Database Logs [Technical Report]

Provenance encodes information that connects datasets, their generation workflows, and associated metadata (e.g., who or when executed a query). As such, it is instrumental for a wide range of critical governance applications (e.g., observability and auditing). Unfortunately, in the context of database systems, extracting coarse-grained provenance is a long-standing problem due to the complexity and sheer volume of database workflows. Provenance extraction from query event logs has been recently proposed as favorable because, in principle, can result in meaningful provenance graphs for provenance applications. Current approaches, however, (a) add substantial overhead to the database and provenance extraction workflows and (b)~extract provenance that is noisy, omits query execution dependencies, and is not rich enough for upstream applications. To address these problems, we introduce OneProvenance: an efficient provenance extraction system from query event logs. OneProvenance addresses the unique challenges of log-based extraction by (a)~identifying query execution dependencies through efficient log analysis, (b) extracting provenance through novel event transformations that account for query dependencies, and (c)~introducing effective filtering optimizations. Our thorough experimental analysis shows that OneProvenance can improve extraction by up to ~18X compared to state-of-the-art baselines; our optimizations reduce the extraction noise and optimize performance even further. OneProvenance is deployed at scale by Microsoft Purview and actively supports customer provenance extraction needs (https://bit.ly/3N2JVGF).

cs.DB↗

arXiv2023

Query Processing on Tensor Computation Runtimes

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 10$\times$ over specialized CPU- and GPU-only systems. Finally, TQP can accelerate queries mixing ML predictions and SQL end-to-end, and deliver up to 9$\times$ speedup over CPU baselines.

cs.DB↗

arXiv2022

The Tensor Data Platform: Towards an AI-centric Database System

Database engines have historically absorbed many of the innovations in data processing, adding features to process graph data, XML, object oriented, and text among many others. In this paper, we make the case that it is time to do the same for AI -- but with a twist! While existing approaches have tried to achieve this by integrating databases with external ML tools, in this paper we claim that achieving a truly AI-centric database requires moving the DBMS engine, at its core, from a relational to a tensor abstraction. This allows us to: (1) support multi-modal data processing such as images, videos, audio, text as well as relational; (2) leverage the wellspring of innovation in HW and runtimes for tensor computation; and (3) exploit automatic differentiation to enable a novel class of "trainable" queries that can learn to perform a task. To support the above scenarios, we introduce TDP: a system that builds upon our prior work mapping relational queries to tensors. Thanks to a tighter integration with the tensor runtime, TDP is able to provide a broader coverage of new emerging scenarios requiring access to multi-modal data and automatic differentiation.

cs.DB↗

arXiv2022

LlamaTune: Sample-Efficient DBMS Configuration Tuning

Tuning a database system to achieve optimal performance on a given workload is a long-standing problem in the database community. A number of recent works have leveraged ML-based approaches to guide the sampling of large parameter spaces (hundreds of tuning knobs) in search for high performance configurations. Looking at Microsoft production services operating millions of databases, sample efficiency emerged as a crucial requirement to use tuners on diverse workloads. This motivates our investigation in LlamaTune, a tuner design that leverages domain knowledge to improve the sample efficiency of existing optimizers. LlamaTune employs an automated dimensionality reduction technique based on randomized projections, a biased-sampling approach to handle special values for certain knobs, and knob values bucketization, to reduce the size of the search space. LlamaTune compares favorably with the state-of-the-art optimizers across a diverse set of workloads. It identifies the best performing configurations with up to $11\times$ fewer workload runs, and reaching up to $21\%$ higher throughput. We also show that benefits from LlamaTune generalize across both BO-based and RL-based optimizers, as well as different DBMS versions. While the journey to perform database tuning at cloud-scale remains long, LlamaTune goes a long way in making automatic DBMS tuning practical at scale.

cs.DB↗

arXiv2021

KEA: Tuning an Exabyte-Scale Data Infrastructure

Microsoft's internal big-data infrastructure is one of the largest in the world -- with over 300k machines running billions of tasks from over 0.6M daily jobs. Operating this infrastructure is a costly and complex endeavor, and efficiency is paramount. In fact, for over 15 years, a dedicated engineering team has tuned almost every aspect of this infrastructure, achieving state-of-the-art efficiency (>60% average CPU utilization across all clusters). Despite rich telemetry and strong expertise, faced with evolving hardware/software/workloads this manual tuning approach had reached its limit -- we had plateaued. In this paper, we present KEA, a multi-year effort to automate our tuning processes to be fully data/model-driven. KEA leverages a mix of domain knowledge and principled data science to capture the essence of our cluster dynamic behavior in a set of machine learning (ML) models based on collected system data. These models power automated optimization procedures for parameter tuning, and inform our leadership in critical decisions around engineering and capacity management (such as hardware and data center design, software investments, etc.). We combine "observational" tuning (i.e., using models to predict system behavior without direct experimentation) with judicious use of "flighting" (i.e., conservative testing in production). This allows us to support a broad range of applications that we discuss in this paper. KEA continuously tunes our cluster configurations and is on track to save Microsoft tens of millions of dollars per year. At the best of our knowledge, this paper is the first to discuss research challenges and practical learnings that emerge when tuning an exabyte-scale data infrastructure.

cs.DB↗

arXiv2020

A Tensor Compiler for Unified Machine Learning Prediction Serving

Machine Learning (ML) adoption in the enterprise requires simpler and more efficient software infrastructure---the bespoke solutions typical in large web companies are simply untenable. Model scoring, the process of obtaining predictions from a trained model over new data, is a primary contributor to infrastructure complexity and cost as models are trained once but used many times. In this paper we propose HUMMINGBIRD, a novel approach to model scoring, which compiles featurization operators and traditional ML models (e.g., decision trees) into a small set of tensor operations. This approach inherently reduces infrastructure complexity and directly leverages existing investments in Neural Network compilers and runtimes to generate efficient computations for both CPU and hardware accelerators. Our performance results are intriguing: despite replacing imperative computations (e.g., tree traversals) with tensor computation abstractions, HUMMINGBIRD is competitive and often outperforms hand-crafted kernels on micro-benchmarks on both CPU and GPU, while enabling seamless end-to-end acceleration of ML pipelines. We have released HUMMINGBIRD as open source.

cs.LG↗

arXiv2020

Seagull: An Infrastructure for Load Prediction and Optimized Resource Allocation

Microsoft Azure is dedicated to guarantee high quality of service to its customers, in particular, during periods of high customer activity, while controlling cost. We employ a Data Science (DS) driven solution to predict user load and leverage these predictions to optimize resource allocation. To this end, we built the Seagull infrastructure that processes per-server telemetry, validates the data, trains and deploys ML models. The models are used to predict customer load per server (24h into the future), and optimize service operations. Seagull continually re-evaluates accuracy of predictions, fallback to previously known good models and triggers alerts as appropriate. We deployed this infrastructure in production for PostgreSQL and MySQL servers across all Azure regions, and applied it to the problem of scheduling server backups during low-load time. This minimizes interference with user-induced load and improves customer experience.

cs.DC↗

arXiv2020

MLOS: An Infrastructure for Automated Software Performance Engineering

Developing modern systems software is a complex task that combines business logic programming and Software Performance Engineering (SPE). The later is an experimental and labor-intensive activity focused on optimizing the system for a given hardware, software, and workload (hw/sw/wl) context. Today's SPE is performed during build/release phases by specialized teams, and cursed by: 1) lack of standardized and automated tools, 2) significant repeated work as hw/sw/wl context changes, 3) fragility induced by a "one-size-fit-all" tuning (where improvements on one workload or component may impact others). The net result: despite costly investments, system software is often outside its optimal operating point - anecdotally leaving 30% to 40% of performance on the table. The recent developments in Data Science (DS) hints at an opportunity: combining DS tooling and methodologies with a new developer experience to transform the practice of SPE. In this paper we present: MLOS, an ML-powered infrastructure and methodology to democratize and automate Software Performance Engineering. MLOS enables continuous, instance-level, robust, and trackable systems optimization. MLOS is being developed and employed within Microsoft to optimize SQL Server performance. Early results indicated that component-level optimizations can lead to 20%-90% improvements when custom-tuning for a specific hw/sw/wl, hinting at a significant opportunity. However, several research challenges remain that will require community involvement. To this end, we are in the process of open-sourcing the MLOS core infrastructure, and we are engaging with academic institutions to create an educational program around Software 2.0 and MLOS ideas.

cs.DC↗