Flink Forward SF 2017: Timo Walther - Table & SQL API – unified APIs for batch and stream processing

Current 2022 talk. Speaker: Yingjun Wu Title: Rethinking State Management in Cloud-Native Streaming Systems. Abstract: Stream processing is becoming increasingly essential for extracting business value from data in real-time. To achieve strict user-defined SLAs under constantly changing workloads, modern streaming systems have started taking advantage of the cloud for scalable and resilient resources. New demand opens new opportunities and challenges for state management, which is at the core of streaming systems. Existing approaches typically use embedded key-value storage so that each worker can access it locally to achieve high performance. However, it requires an external durable file system for checkpointing, is complicated and time-consuming to redistribute state during scaling and migration, and is prone to performance throttling. Therefore, we propose shared storage based on LSM-tree. State gets stored at cloud object storage and seamlessly makes itself durable, and the high bandwidth of cloud storage enables fast recovery. The location of a partition of the state decouples with compute nodes thus making scaling straightforward and more efficient. Compaction in this shared LSM-tree is now globally coordinated with opportunistic serverless boosting instead of relying on individual compute nodes. We design a streaming-aware compaction and caching strategy to achieve smoother and better end-to-end performance.

Flink Forward San Francisco 2022. Resource Elasticity is a frequently requested feature in Apache Flink: Users want to be able to easily adjust their clusters to changing workloads for resource efficiency and cost saving reasons. In Flink 1.13, the initial implementation of Reactive Mode was introduced, later releases added more improvements to make the feature production ready. In this talk, we’ll explain scenarios to deploy Reactive Mode to various environments to achieve autoscaling and resource elasticity. We’ll discuss the constraints to consider when planning to use this feature, and also potential improvements from the Flink roadmap. For those interested in the internals of Flink, we’ll also briefly explain how the feature is implemented, and if time permits, conclude with a short demo. by Robert Metzger

During last two major versions (1.9 & 1.10), Apache Flink community spent lots of effort to improve the architecture for further unified batch & streaming processing. One example for that is Flink SQL added the ability to support multiple SQL planners under the same API. This talk will first discuss the motivation behind these movements, but more importantly will have a deep dive into Flink SQL. The presentation shows the unified architecture to handle streaming and batch queries and explain how Flink translates queries into the relational expressions, leverages Apache Calcite to optimize them, and generates efficient runtime code for execution. Besides, this talk will also describe the lifetime of a query in detail, how optimizer improve the plan based on relational node patterns, how Flink leverages binary data format for its basic data structure, and how does certain operator works. This would give audience better understanding of Flink SQL internals.

Flink Forward San Francisco 2022. The Apache Flink Kubernetes Operator provides a consistent approach to manage Flink applications automatically, without any human interaction, by extending the Kubernetes API. Given the increasing adoption of Kubernetes based Flink deployments the community has been working on a Kubernetes native solution as part of Flink that can benefit from the rich experience of community members and ultimately make Flink easier to adopt. In this talk we give a technical introduction to the Flink Kubernetes Operator and demonstrate the core features and use-cases through in-depth examples." by Thomas Weise

Spring Batch is a framework for writing batch jobs that can run as scheduled or on-demand processes without user interaction. It provides reusable components for connecting to databases or other systems, processing and transforming data in chunks, and writing output. The basic architecture includes a job launcher, job made of steps, and components for reading input, processing it, and writing output in chunks. Spring Batch Admin provides a web-based interface for monitoring and managing batch jobs.

Arrow Flight is a proposed RPC layer for Apache Arrow that allows for efficient transfer of Arrow record batches between systems. It uses GRPC as the foundation to define streams of Arrow data that can be consumed in parallel across locations. Arrow Flight supports custom actions that can be used to build services on top of the generic API. By extending GRPC, Arrow Flight aims to simplify the creation of data applications while enabling high performance data transfer and locality awareness.

This document discusses Aspect Oriented Programming (AOP) using the Spring Framework. It defines AOP as a programming paradigm that extends OOP by enabling modularization of crosscutting concerns. It then discusses how AOP addresses common crosscutting concerns like logging, validation, caching, and transactions through aspects, pointcuts, and advice. It also compares Spring AOP and AspectJ, and shows how to implement AOP in Spring using annotations or XML.

The document discusses code optimization techniques in Spark SQL's Catalyst optimizer. It describes how function outlining can improve performance of generated Java code by splitting large routines into smaller ones. The document outlines a Spark SQL query optimization case study where outlining a 300+ line routine from Catalyst code generation improved query performance by up to 19% on a Power8 cluster. Overall, the document examines how function outlining and other code generation optimizations in Catalyst can help the Java JIT compiler better optimize Spark SQL queries.

There are a growing set of optimization mechanisms that allow you to achieve competitive SQL performance. Spark has extension points that help third parties to add customizations and optimizations without needing these optimizations to be merged into Apache Spark. This is very powerful and helps extensibility. We have added some enhancements to the existing extension points framework to enable some fine grained control. This talk will be a deep dive at the extension points that is available in Spark today. We will also talk about the enhancements to this API that we developed to help make this API more powerful. This talk will be of benefit to developers who are looking to customize Spark in their deployments.

The document discusses Struts framework and internationalization (I18N) applications. Some key points: 1. Struts is a MVC framework that simplifies development of web applications. It provides components like ActionServlet and tag libraries. 2. I18N applications display output based on the user's locale/language. This is achieved using properties files with language-specific translations and the ResourceBundle class. 3. In Struts, properties files are configured in struts-config.xml and accessed in JSPs using the <message> tag. Keys not found will result in errors unless null=false is specified.

This document provides an overview and comparison of the algorithms, phases, and commonalities of modern concurrent garbage collectors in HotSpot, including G1, Shenandoah, and Z GC. It begins with laying the groundwork on stop-the-world vs concurrent collection and heap layout. It then introduces the key differences between the three collectors in their marking, barrier, and compaction approaches. The goal of the document is to provide a technical introduction and high-level differences between these concurrent garbage collectors.

Flink Forward San Francisco 2022. Being in the payments space, Stripe requires strict correctness and freshness guarantees. We rely on Flink as the natural solution for delivering on this in support of our Change Data Capture (CDC) infrastructure. We heavily rely on CDC as a tool for capturing data change streams from our databases without critically impacting database reliability, scalability, and maintainability. Data derived from these streams is used broadly across the business and powers many of our critical financial reporting systems totalling over $640 Billion in payment volume annually. We use many components of Flink’s flexible DataStream API to perform aggregations and abstract away the complexities of stream processing from our downstreams. In this talk, we’ll walk through our experience from the very beginning to what we have in production today. We’ll share stories around the technical details and trade-offs we encountered along the way. by Jeff Chao

Tutorial Videos: https://www.youtube.com/playlist?list=PLD8nQCAhR3tQ7KXnvIk_v_SLK-Fb2y_k_ Day 1 : Introduction to React, Babel and Webpack Prerequisites of starting the workshop ( Basic understanding of Node & Express ) What is Virtual DOM? What is React and why should we use it? Install and set up React: a-Using create-react-app b-From scratch using Babel and Webpack. We will use Webpack Dev Server. Day 2 : React Basic Concepts Types of Components: Class-based and Functional based Components Use of JSX Parent, Child, and Nested Components Difference between State and Props Create and Handle Routes Component Lifecycle Methods Create a form and handling form inputs Use of arrow functions and Spread Operator Day 3: Advanced Concepts in React Use of Refs What are Higher Order Components( HOC )? How to use HOC Understanding Context in React

The presentation for the Shift Split 2017 conf workshop. Example project repo: https://github.com/infinum/shift-2017

Speaker: Timo Walther (https://www.linkedin.com/in/twalthr/) Video: https://www.youtube.com/watch?v=Ych5bbmDIoA Talk presented during Bangalore Kafka group's stream processing meetup at Hotstar https://www.meetup.com/Bangalore-Apache-Kafka-Group/events/265285812/

Flink Forward San Francisco 2022. Pinterest is a visual discovery engine that serves over 433MM users. Stream processing allows us to unlock value from realtime data for pinners. At Pinterest, we adopt Flink as the unified streaming processing engine. In this talk, we will share our journey in building a stream processing platform with Flink and how we onboarding critical use cases to the platform. Pinterest has supported 90+near realtime streaming applications. We will cover the problem statement, how we evaluate potential solutions and our decision to build the framework. by Rainie Li & Kanchi Masalia

Although NVMe has been more and more popular these years, a large amount of HDD are still widely used in super-large scale big data clusters. In a EB-level data platform, IO(including decompression and decode) cost contributes a large proportion of Spark jobs’ cost. In another word, IO operation is worth optimizing. In ByteDancen, we do a series of IO optimization to improve performance, including parallel read and asynchronized shuffle. Firstly we implement file level parallel read to improve performance when there are a lot of small files. Secondly, we design row group level parallel read to accelerate queries for big-file scenario. Thirdly, implement asynchronized spill to improve job peformance. Besides, we design parquet column family, which will split a table into a few column families and different column family will be in different Parquets files. Different column family can be read in parallel, so the read performance is much higher than the existing approach. In our practice, the end to end performance is improved by 5% to 30% In this talk, I will illustrate how we implement these features and how they accelerate Apache Spark jobs.

SQL is undoubtedly the most widely used language for data analytics. It is declarative and can be optimized and efficiently executed by most query processors. Therefore the community has made effort to add relational APIs to Apache Flink, a standard SQL API and a language-integrated Table API. Both APIs are semantically compatible and share the same optimization and execution path based on Apache Calcite. Since Flink supports both stream and batch processing and many use cases require both kinds of processing, we aim for a unified relational layer. In this talk we will look at the current API capabilities, find out what's under the hood of Flink’s relational APIs, and give an outlook for future features such as dynamic tables, Flink's way how streams are converted into tables and vice versa leveraging the stream-table duality.

SQL is undoubtedly the most widely used language for data analytics. It is declarative and can be optimized and efficiently executed by most query processors. Therefore the community has made effort to add relational APIs to Apache Flink, a standard SQL API and a language-integrated Table API. Both APIs are semantically compatible and share the same optimization and execution path based on Apache Calcite. Since Flink supports both stream and batch processing and many use cases require both kinds of processing, we aim for a unified relational layer. In this talk we will look at the current API capabilities, find out what's under the hood of Flink’s relational APIs, and give an outlook for future features such as dynamic tables, Flink's way how streams are converted into tables and vice versa leveraging the stream-table duality.

SQL can be used to query both streaming and batch data. Apache Flink and Apache Calcite enable SQL queries on streaming data. Flink uses its Table API and integrates with Calcite to translate SQL queries into dataflow programs. This allows standard SQL to be used for both traditional batch analytics on finite datasets and stream analytics producing continuous results from infinite data streams. Queries are executed continuously by applying operators within windows to subsets of streaming data.

Fabian Hueske presented on stream analytics using SQL on Apache Flink. Flink provides a scalable platform for stream processing that is fast, accurate, and reliable. Its relational APIs allow querying both batch and streaming data using standard SQL or a LINQ-style Table API. Queries on streaming data produce continuously updating results. Windows can be used to compute aggregates over tumbling time intervals. The dynamic tables representing streaming data can be converted to output streams encoding updates as insertions and deletions. While not all queries can be supported, techniques like limiting state size allow bounding computational resources. Use cases like continuous ETL, dashboards, and event-driven architectures were discussed.

Apache Flink's DataStream API is very expressive and gives users precise control over time and state. However, many applications do not require this level of expressiveness and can be implemented more concisely and easily with a domain-specific API. SQL is undoubtedly the most widely used language for data processing but usually applied in the domain of batch processing. Apache Flink features two relational APIs for unified stream and batch processing, the Table API, a language-integrated relational query API for Scala and Java, and SQL. A Table API or SQL query computes the same result regardless whether it is evaluated on a static file or on a Kafka topic. While Flink evaluates queries on batch input like a conventional query engine, queries on streaming input are continuously processed and their results constantly updated and refined. In this talk we present Flink’s unified relational APIs, show how streaming SQL queries are processed, and discuss exciting new use-cases.

Apache Flink's DataStream API is very expressive and gives users precise control over time and state. However, many applications do not require this level of expressiveness and can be implemented more concisely and easily with a domain-specific API. SQL is undoubtedly the most widely used language for data processing but usually applied in the domain of batch processing. Apache Flink features two relational APIs for unified stream and batch processing, the Table API, a language-integrated relational query API for Scala and Java, and SQL. A Table API or SQL query computes the same result regardless whether it is evaluated on a static file or on a Kafka topic. While Flink evaluates queries on batch input like a conventional query engine, queries on streaming input are continuously processed and their results constantly updated and refined. In this talk we present Flink’s unified relational APIs, show how streaming SQL queries are processed, and discuss exciting new use-cases.

Slides of my Apache Flink talk "Big Things" meetup at Proofpoint and "Apache Flink" meetup at Clicktale in January 2018