Vectorized execution engines process large datasets by decomposing computations into concise (tight) loops, which can be more efficiently executed by modern hardware. Providing loops that are optimal for execution usually adds burden to the software development process, as developers are required to understand details of vectorized execution, columnar data layout, data encodings, and the code compilation process itself, presenting a steep learning curve and challenges to organizations building and scaling large engineering teams. Due to their large quantity, scalar function authoring accentuates this problem. In our experience building the Velox open source execution engine, we have observed that exposing a large number of developers to the complexity inherent to vectorization resulted in a disproportionate amount of bugs and performance inefficiencies. In this paper, we describe the simple function interface (SFI) created to address this issue. SFI highly simplifies scalar function authoring by encapsulating the vectorization complexity required to generate tight loops, and presenting developers with a simpler, conciser, and more natural row-based interface - without sacrificing performance. SFI also hides columnar layout details, while providing developers the flexibility to efficiently implement advanced features such as functions with nested and recursive parameter types, type variables, variadic parameters, and generic types. Today, more than a thousand functions have been added to Velox using the SFI, implementing popular open source SQL dialects and internal domain-specific use cases at Meta, and are in active production use. While this paper presents implementation details, performance pitfalls, experimental results, and our overall experience developing the state-of-the-art Velox vectorized execution engine, we believe the concepts and trade-offs to be fundamentally equivalent and generally applicable to other vectorized engines.

Datacenter-scale AI training clusters consisting of thousands of domain-specific accelerators (DSA) are used to train increasingly-complex deep learning models. These clusters rely on a data storage and ingestion (DSI) pipeline, responsible for storing exabytes of training data and serving it at tens of terabytes per second. As DSAs continue to push training efficiency and throughput, the DSI pipeline is becoming the dominating factor that constrains the overall training performance and capacity. Innovations that improve the efficiency and performance of DSI systems and hardware are urgent, demanding a deep understanding of DSI characteristics and infrastructure at scale.

The ad-hoc development of new specialized computation engines targeted to very specific data workloads has created a siloed data landscape. Commonly, these engines share little to nothing with each other and are hard to maintain, evolve, and optimize, and ultimately provide an inconsistent experience to data users. In order to address these issues, Meta has created Velox, a novel open source C++ database acceleration library. Velox provides reusable, extensible, high-performance, and dialect-agnostic data processing components for building execution engines, and enhancing data management systems. The library heavily relies on vectorization and adaptivity, and is designed from the ground up to support efficient computation over complex data types due to their ubiquity in modern workloads. Velox is currently integrated or being integrated with more than a dozen data systems at Meta, including analytical query engines such as Presto and Spark, stream processing platforms, message buses and data warehouse ingestion infrastructure, machine learning systems for feature engineering and data preprocessing (PyTorch), and more. It provides benefits in terms of (a) efficiency wins by democratizing optimizations previously only found in individual engines, (b) increased consistency for data users, and (c) engineering efficiency by promoting reusability.

Stream processing applications reduce the latency of batch data pipelines and enable engineers to quickly identify production issues. Many times, a service can log data to distinct streams, even if they relate to the same real-world event (e.g., a search on Facebook's search bar). Furthermore, the logging of related events can appear on the server side with different delay, causing one stream to be significantly behind the other in terms of logged event times for a given log entry. To be able to stitch this information together with low latency , we need to be able to join two different streams where each stream may have its own characteristics regarding the degree in which its data is out-of-order . Doing so in a streaming fashion is challenging as a join operator consumes lots of memory, especially with significant data volumes. This paper describes an end-to-end streaming join service that addresses the challenges above through a streaming join operator that uses an adaptive stream synchronization algorithm that is able to handle the different distributions we observe in real-world streams regarding their event times. This synchronization scheme paces the parsing of new data and reduces overall operator memory footprint while still providing high accuracy. We have integrated this into a streaming SQL system and have successfully reduced the latency of several batch pipelines using this approach.

Realtime data processing powers many use cases at Facebook, including realtime reporting of the aggregated, anonymized voice of Facebook users, analytics for mobile applications, and insights for Facebook page administrators. Many companies have developed their own systems; we have a realtime data processing ecosystem at Facebook that handles hundreds of Gigabytes per second across hundreds of data pipelines.