Carsten Binnig

Johannes Wehrstein, Roman Heinrich, Mihail Stoian et al.

Workload traces from cloud data warehouse providers reveal that standard benchmarks such as TPC-H and TPC-DS fail to capture key characteristics of real-world workloads, including query repetition and string-heavy queries. In this paper, we introduce Redbench, a novel benchmark featuring a workload generator that reproduces real-world workload characteristics derived from traces released by cloud providers. Redbench integrates multiple workload generation techniques to tailor workloads to specific objectives, transforming existing benchmarks into realistic query streams that preserve intrinsic workload characteristics. By focusing on inherent workload signals rather than execution-specific metrics, Redbench bridges the gap between synthetic and real workloads. Our evaluation shows that (1) Redbench produces more realistic and reproducible workloads for cloud data warehouse benchmarking, and (2) Redbench reveals the impact of system optimizations across four commercial data warehouse platforms. We believe that Redbench provides a crucial foundation for advancing research on optimization techniques for modern cloud data warehouses.

Matthias Urban, Vu Huy Nguyen, Gabriele Sanmartino et al.

Semantic data systems integrate Large Language Models (LLMs) and Vision-Language Models (VLMs) directly into database query execution, enabling expressive queries on multi-modal data. However, optimizing these queries requires accurate selectivity estimates to determine the most efficient operator execution order. Contemporary systems rely on online sample-based profiling, a process that incurs severe latency overheads and struggles with low-selectivity queries. In this paper, we introduce Semantic Histograms, a novel selectivity estimator for semantic filters on image data that leverages shared embedding spaces to bypass traditional profiling. We realize that all semantic filters are implicit range queries, as they match a range of different images. Some filter predicates are more general, yielding a wide range, while others are more specific, yielding a smaller range. To address the challenge of implicit ranges, we propose two approaches to estimate the queries' specificity, with an ensemble of the two performing best. The evaluation shows that Semantic Histograms can reduce the end-to-end runtime overhead of query optimization and execution by up to 86%.

Marius Gassen, Benjamin Hättasch, Benjamin Hilprecht et al.

Databases for OLTP are often the backbone for applications such as hotel room or cinema ticket booking applications. However, developing a conversational agent (i.e., a chatbot-like interface) to allow end-users to interact with an application using natural language requires both immense amounts of training data and NLP expertise. This motivates CAT, which can be used to easily create conversational agents for transactional databases. The main idea is that, for a given OLTP database, CAT uses weak supervision to synthesize the required training data to train a state-of-the-art conversational agent, allowing users to interact with the OLTP database. Furthermore, CAT provides an out-of-the-box integration of the resulting agent with the database. As a major difference to existing conversational agents, agents synthesized by CAT are data-aware. This means that the agent decides which information should be requested from the user based on the current data distributions in the database, which typically results in markedly more efficient dialogues compared with non-data-aware agents. We publish the code for CAT as open source.

Jigao Luo, Qi Chen, Carsten Binnig

Parquet is the de facto columnar file format in modern analytical systems, yet its configuration guidelines have largely been shaped by CPU-centric execution models. As GPU-accelerated data processing becomes increasingly prevalent, Parquet files generated with CPU-oriented defaults can severely underutilize GPU parallelism, turning GPU scans into a performance bottleneck. In this work, we systematically study how Parquet configurations affect GPU scan performance. We show that Parquet's poor GPU performance is not inherent to the format itself but rather a consequence of suboptimal configuration choices. By applying GPU-aware configurations, we increase effective read bandwidth up to 125 GB/s without modifying the Parquet specification.

Jigao Luo, Nils Boeschen, Muhammad El-Hindi et al.

The AI hardware boom has led modern data centers to adopt HPC-style architectures centered on distributed, GPU-centric computation. Large GPU clusters interconnected by fast RDMA networks and backed by high-bandwidth NVMe storage enable scalable computation and rapid access to storage-resident data. Tensor computation runtimes (TCRs), such as PyTorch, originally designed for AI workloads, have recently been shown to accelerate analytical workloads. However, prior work has primarily considered settings where the data fits in aggregated GPU memory. In this paper, we systematically study how TCRs can support scalable, distributed query processing for large-scale, storage-resident OLAP workloads. Although TCRs provide abstractions for network and storage I/O, naive use often underutilizes GPU and I/O bandwidth due to insufficient overlap between computation and data movement. As a core contribution, we present PystachIO, a prototype of a PyTorch-based distributed OLAP engine that combines fast network and storage I/O with key optimizations to maximize GPU, network, and storage utilization. Our evaluation shows up to 3x end-to-end speedups over existing distributed GPU-based query processing approaches.

Yannis Chronis, Yawen Wang, Yu Gan et al.

Cardinality estimation is crucial for enabling high query performance in relational databases. Recently learned cardinality estimation models have been proposed to improve accuracy but there is no systematic benchmark or datasets which allows researchers to evaluate the progress made by new learned approaches and even systematically develop new learned approaches. In this paper, we are releasing a benchmark, containing thousands of queries over 20 distinct real-world databases for learned cardinality estimation. In contrast to other initial benchmarks, our benchmark is much more diverse and can be used for training and testing learned models systematically. Using this benchmark, we explored whether learned cardinality estimation can be transferred to an unseen dataset in a zero-shot manner. We trained GNN-based and transformer-based models to study the problem in three setups: 1-) instance-based, 2-) zero-shot, and 3-) fine-tuned. Our results show that while we get promising results for zero-shot cardinality estimation on simple single table queries; as soon as we add joins, the accuracy drops. However, we show that with fine-tuning, we can still utilize pre-trained models for cardinality estimation, significantly reducing training overheads compared to instance specific models. We are open sourcing our scripts to collect statistics, generate queries and training datasets to foster more extensive research, also from the ML community on the important problem of cardinality estimation and in particular improve on recent directions such as pre-trained cardinality estimation.

Benjamin Hilprecht, Christian Hammacher, Eduardo Reis et al.

In this paper, we present our vision of differentiable ML pipelines called DiffML to automate the construction of ML pipelines in an end-to-end fashion. The idea is that DiffML allows to jointly train not just the ML model itself but also the entire pipeline including data preprocessing steps, e.g., data cleaning, feature selection, etc. Our core idea is to formulate all pipeline steps in a differentiable way such that the entire pipeline can be trained using backpropagation. However, this is a non-trivial problem and opens up many new research questions. To show the feasibility of this direction, we demonstrate initial ideas and a general principle of how typical preprocessing steps such as data cleaning, feature selection and dataset selection can be formulated as differentiable programs and jointly learned with the ML model. Moreover, we discuss a research roadmap and core challenges that have to be systematically tackled to enable fully differentiable ML pipelines.

Benjamin Hättasch, Michael Truong-Ngoc, Andreas Schmidt et al.

Since data is often stored in different sources, it needs to be integrated to gather a global view that is required in order to create value and derive knowledge from it. A critical step in data integration is schema matching which aims to find semantic correspondences between elements of two schemata. In order to reduce the manual effort involved in schema matching, many solutions for the automatic determination of schema correspondences have already been developed. In this paper, we propose a novel end-to-end approach for schema matching based on neural embeddings. The main idea is to use a two-step approach consisting of a table matching step followed by an attribute matching step. In both steps we use embeddings on different levels either representing the whole table or single attributes. Our results show that our approach is able to determine correspondences in a robust and reliable way and compared to traditional schema matching approaches can find non-trivial correspondences.

Benjamin Hilprecht, Kristian Kersting, Carsten Binnig

While there has been extensive work on deep neural networks for images and text, deep learning for relational databases (RDBs) is still a rather unexplored field. One direction that recently gained traction is to apply Graph Neural Networks (GNNs) to RBDs. However, training GNNs on large relational databases (i.e., data stored in multiple database tables) is rather inefficient due to multiple rounds of training and potentially large and inefficient representations. Hence, in this paper we propose SPARE (Single-Pass Relational models), a new class of neural models that can be trained efficiently on RDBs while providing similar accuracies as GNNs. For enabling efficient training, different from GNNs, SPARE makes use of the fact that data in RDBs has a regular structure, which allows one to train these models in a single pass while exploiting symmetries at the same time. Our extensive empirical evaluation demonstrates that SPARE can significantly speedup both training and inference while offering competitive predictive performance over numerous baselines.

Matthias Urban, Carsten Binnig

In this paper, we propose Multi-Modal Databases (MMDBs), which is a new class of database systems that can seamlessly query text and tables using SQL. To enable seamless querying of textual data using SQL in an MMDB, we propose to extend relational databases with so-called multi-modal operators (MMOps) which are based on the advances of recent large language models such as GPT-3. The main idea of MMOps is that they allow text collections to be treated as tables without the need to manually transform the data. As we show in our evaluation, our MMDB prototype can not only outperform state-of-the-art approaches such as text-to-table in terms of accuracy and performance but it also requires significantly less training data to fine-tune the model for an unseen text collection.

Benjamin Hättasch, Jan-Micha Bodensohn, Carsten Binnig

In this paper, we propose a new system called ASET that allows users to perform structured explorations of text collections in an ad-hoc manner. The main idea of ASET is to use a new two-phase approach that first extracts a superset of information nuggets from the texts using existing extractors such as named entity recognizers and then matches the extractions to a structured table definition as requested by the user based on embeddings. In our evaluation, we show that ASET is thus able to extract structured data from real-world text collections in high quality without the need to design extraction pipelines upfront.

Roman Heinrich, Carsten Binnig, Harald Kornmayer et al.

In this work, we present COSTREAM, a novel learned cost model for Distributed Stream Processing Systems that provides accurate predictions of the execution costs of a streaming query in an edge-cloud environment. The cost model can be used to find an initial placement of operators across heterogeneous hardware, which is particularly important in these environments. In our evaluation, we demonstrate that COSTREAM can produce highly accurate cost estimates for the initial operator placement and even generalize to unseen placements, queries, and hardware. When using COSTREAM to optimize the placements of streaming operators, a median speed-up of around 21x can be achieved compared to baselines.

Zixuan Yi, Sami Abu-el-Haija, Yawen Wang et al.

DB engines produce efficient query execution plans by relying on cost models. Practical implementations estimate cardinality of queries using heuristics, with magic numbers tuned to improve average performance on benchmarks. Empirically, estimation error significantly grows with query complexity. Alternatively, learning-based estimators offer improved accuracy, but add operational complexity preventing their adoption in-practice. Recognizing that query workloads contain highly repetitive subquery patterns, we learn many simple regressors online, each localized to a pattern. The regressor corresponding to a pattern can be randomly-accessed using hash of graph structure of the subquery. Our method has negligible overhead and competes with SoTA learning-based approaches on error metrics. Further, amending PostgreSQL with our method achieves notable accuracy and runtime improvements over traditional methods and drastically reduces operational costs compared to other learned cardinality estimators, thereby offering the most practical and efficient solution on the Pareto frontier. Concretely, simulating JOB-lite workload on IMDb speeds-up execution by 7.5 minutes (>30%) while incurring only 37 seconds overhead for online learning.

Liane Vogel, Benjamin Hilprecht, Carsten Binnig

Tabular representation learning has recently gained a lot of attention. However, existing approaches only learn a representation from a single table, and thus ignore the potential to learn from the full structure of relational databases, including neighboring tables that can contain important information for a contextualized representation. Moreover, current models are significantly limited in scale, which prevents that they learn from large databases. In this paper, we thus introduce our vision of relational representation learning, that can not only learn from the full relational structure, but also can scale to larger database sizes that are commonly found in real-world. Moreover, we also discuss opportunities and challenges we see along the way to enable this vision and present initial very promising results. Overall, we argue that this direction can lead to foundation models for relational databases that are today only available for text and images.

Benjamin Hilprecht, Carsten Binnig

In this paper, we introduce zero-shot cost models which enable learned cost estimation that generalizes to unseen databases. In contrast to state-of-the-art workload-driven approaches which require to execute a large set of training queries on every new database, zero-shot cost models thus allow to instantiate a learned cost model out-of-the-box without expensive training data collection. To enable such zero-shot cost models, we suggest a new learning paradigm based on pre-trained cost models. As core contributions to support the transfer of such a pre-trained cost model to unseen databases, we introduce a new model architecture and representation technique for encoding query workloads as input to those models. As we will show in our evaluation, zero-shot cost estimation can provide more accurate cost estimates than state-of-the-art models for a wide range of (real-world) databases without requiring any query executions on unseen databases. Furthermore, we show that zero-shot cost models can be used in a few-shot mode that further improves their quality by retraining them just with a small number of additional training queries on the unseen database.

Benjamin Hilprecht, Carsten Binnig

In this paper, we present our vision of so called zero-shot learning for databases which is a new learning approach for database components. Zero-shot learning for databases is inspired by recent advances in transfer learning of models such as GPT-3 and can support a new database out-of-the box without the need to train a new model. Furthermore, it can easily be extended to few-shot learning by further retraining the model on the unseen database. As a first concrete contribution in this paper, we show the feasibility of zero-shot learning for the task of physical cost estimation and present very promising initial results. Moreover, as a second contribution we discuss the core challenges related to zero-shot learning for databases and present a roadmap to extend zero-shot learning towards many other tasks beyond cost estimation or even beyond classical database systems and workloads.

Tiemo Bang, Norman May, Ilia Petrov et al.

In this paper, we propose a radical new approach for scale-out distributed DBMSs. Instead of hard-baking an architectural model, such as a shared-nothing architecture, into the distributed DBMS design, we aim for a new class of so-called architecture-less DBMSs. The main idea is that an architecture-less DBMS can mimic any architecture on a per-query basis on-the-fly without any additional overhead for reconfiguration. Our initial results show that our architecture-less DBMS AnyDB can provide significant speed-ups across varying workloads compared to a traditional DBMS implementing a static architecture.

Cagatay Turkay, Nicola Pezzotti, Carsten Binnig et al.

Data science requires time-consuming iterative manual activities. In particular, activities such as data selection, preprocessing, transformation, and mining, highly depend on iterative trial-and-error processes that could be sped-up significantly by providing quick feedback on the impact of changes. The idea of progressive data science is to compute the results of changes in a progressive manner, returning a first approximation of results quickly and allow iterative refinements until converging to a final result. Enabling the user to interact with the intermediate results allows an early detection of erroneous or suboptimal choices, the guided definition of modifications to the pipeline and their quick assessment. In this paper, we discuss the progressiveness challenges arising in different steps of the data science pipeline. We describe how changes in each step of the pipeline impact the subsequent steps and outline why progressive data science will help to make the process more effective. Computing progressive approximations of outcomes resulting from changes creates numerous research challenges, especially if the changes are made in the early steps of the pipeline. We discuss these challenges and outline first steps towards progressiveness, which, we argue, will ultimately help to significantly speed-up the overall data science process.

Moritz Kulessa, Alejandro Molina, Carsten Binnig et al.

Interactive visualizations are arguably the most important tool to explore, understand and convey facts about data. In the past years, the database community has been working on different techniques for Approximate Query Processing (AQP) that aim to deliver an approximate query result given a fixed time bound to support interactive visualizations better. However, classical AQP approaches suffer from various problems that limit the applicability to support the ad-hoc exploration of a new data set: (1) Classical AQP approaches that perform online sampling can support ad-hoc exploration queries but yield low quality if executed over rare subpopulations. (2) Classical AQP approaches that rely on offline sampling can use some form of biased sampling to mitigate these problems but require a priori knowledge of the workload, which is often not realistic if users want to explore a new database. In this paper, we present a new approach to AQP called Model-based AQP that leverages generative models learned over the complete database to answer SQL queries at interactive speeds. Different from classical AQP approaches, generative models allow us to compute responses to ad-hoc queries and deliver high-quality estimates also over rare subpopulations at the same time. In our experiments with real and synthetic data sets, we show that Model-based AQP can in many scenarios return more accurate results in a shorter runtime. Furthermore, we think that our techniques of using generative models presented in this paper can not only be used for AQP in databases but also has applications for other database problems including Query Optimization as well as Data Cleaning.

Prasetya Utama, Nathaniel Weir, Fuat Basik et al.

The ability to extract insights from new data sets is critical for decision making. Visual interactive tools play an important role in data exploration since they provide non-technical users with an effective way to visually compose queries and comprehend the results. Natural language has recently gained traction as an alternative query interface to databases with the potential to enable non-expert users to formulate complex questions and information needs efficiently and effectively. However, understanding natural language questions and translating them accurately to SQL is a challenging task, and thus Natural Language Interfaces for Databases (NLIDBs) have not yet made their way into practical tools and commercial products. In this paper, we present DBPal, a novel data exploration tool with a natural language interface. DBPal leverages recent advances in deep models to make query understanding more robust in the following ways: First, DBPal uses a deep model to translate natural language statements to SQL, making the translation process more robust to paraphrasing and other linguistic variations. Second, to support the users in phrasing questions without knowing the database schema and the query features, DBPal provides a learned auto-completion model that suggests partial query extensions to users during query formulation and thus helps to write complex queries.