Introduction:In depth interpretation of the new features of Flink version 1.13 +flink’s typical practical application in the mutual entertainment industry.
For the vast number of Flink developers,
What is the most anticipated content?
What information is most useful?
The most anticipated content is naturally the release of the new version of Flink 1.13 and the launch of new functions! The most useful information, of course, is the case sharing of combining theory with enterprise practice! What’s more exciting! What kind of sparks can be produced by combining the two?
This meetup is divided into two sessions, with guests from Alibaba, byte beating, Kwai, iqiyi and little red book.
In the first half, four technical experts will bring in-depth interpretation of the new features of Flink 1.13. For example, windows tvf, datastream & Table API interaction, etc;
In the second half, four other senior industry technical experts will share Flink’s practical application in the mutual entertainment industry. Comprehensively analyze the typical problems faced by the industry, including accurate recommendation, real-time data warehouse, data analysis, etc.
- More practical dry goodsOn the one hand, get the new features and function improvements of version 1.13 at the first time; On the other hand, you can also learn how to explore the practical application of Flink in mutual entertainment scenes, including accurate recommendation, real-time data warehouse, etc., from theory to practice;
- Diversified forms of activitiesThe offline and online services are started synchronously. The city can participate in the offline meetup face-to-face communication, and the live broadcast can also be watched online in different places. The wonderful content is not missed;
- Enrich the surroundings and wait for you, you will have the opportunity to get more beautiful surroundings customized by Flink community!
▼ register now ▼
Introduction to guests and topics
Wang Feng ｜ Alibaba researcher
Wang Feng, a researcher of Alibaba cloud, named “Mo Wen”, is the head of the real-time computing and open platform Department of Alibaba cloud computing platform business unit. At present, he leads the team to build a big data real-time computing platform based on Flink, Hadoop and kubernetes open source technology system, which not only serves all real-time data businesses of Alibaba Group (Taobao, tmall, juhuasuan, Gaode, Youku, Feizhu and rookie, etc.), At the same time, Alibaba cloud also provides the world’s leading real-time computing products and services for the majority of small and medium-sized enterprises.
Morning session – in depth interpretation of the release of the new version of Flink 1.13
In depth interpretation of Flink SQL 1.13
Xubangjiang ｜ Apache Flink contributor, Alibaba Senior Development Engineer
In version 1.13 just released, Flink SQL has brought many new features and function improvements. Here, we will focus on the five aspects of windows tvf, time zone support, datastream & Table API interaction, hive compatibility improvement, and SQL client improvement to deeply understand these core functions.
Xubangjiang (snowed out), Alibaba Senior Development Engineer, focuses on the development of Flink SQL Engine.
《Flink 1.13: Towards Scalable Cloud Native Application》
Songxintong ｜ Apache Flink Committee, Alibaba technical expert
Flink 1.13 has added a passive resource management mode and an adaptive scheduling mode. It has flexible scalability. Combined with the cloud native automatic scaling technology, it can better give play to the advantages of elastic computing resources in the cloud environment. It is another important milestone for Flink to fully embrace the cloud native technology ecology. This topic will introduce the new features of Flink 1.13, such as passive resource management, adaptive scheduling, custom container templates, and share the development history and future planning of Flink’s embrace of cloud native technology ecology.
Songxintong (Wuzang), Apache Flink Committee, Alibaba technical expert, doctor of Peking University. Currently, he works in Alibaba real-time computing team, mainly responsible for R & D related to Flink resource management and deployment.
Flink runtime and datastream API optimization for stream batch integration
Gao Yun ｜ Apache Flink contributor, Alibaba technical expert
In 1.13, aiming at the goal of integrating flow and batch, Flink optimized the performance of large-scale job scheduling and network shuffle in batch execution mode, thus further improving the execution performance of flow jobs and batch jobs; At the same time, in terms of datastream API, Flink is also improving the exit semantics of finite flow jobs, so as to further improve the consistency between semantics and results under different execution modes.
Gao Yun (Yun Qian), from the real-time computing team of Alibaba computing platform business unit, Flink contributor, is mainly engaged in the design and development of Flink datastream API, operation and other parts.
State backend flink-1.13 optimization and production practice sharing
Tang Yun ｜ Apache Flink committer, Alibaba Senior Development Engineer
In the newly released flink-1.13, the state backend module brings relevant optimization and new features in terms of memory control and access delay investigation. We will share them in combination with relevant information in production practice.
Tang Yun, famous for dried tea, has been engaged in the development of Flink kernel state backend/checkpoint module since he joined Alibaba. At the same time, he has also been involved in the development of Flink metrics components and the promotion of Alibaba cloud’s open source big data. Currently, relevant modules such as state backend are maintained in the Apache Flink community.
Afternoon session – Application Practice of mutual entertainment industry
Flink’s practice in Little Red Book
Luan Yanming xiaohongshu data flow team Engineer
The application of Flink in the recommended scenarios of little red book and the evolution of the real-time platform in the last half year, including batch processing, multi cloud support, etc.
Luanyanming, engineer of xiaohongshu data flow team, joined xiaohongshu and began to iterate the real-time data platform from 0.
Flink’s landing practice in byte runout recommendation feature system
Guowenfei ｜ person in charge of basic service of byte bounce recommendation system
From headlines to Tiktok, recommendation is the core business scenario of byte beating, and features are the basic fuel of the recommendation system. Building an efficient real-time feature system is very important for the business iteration of the recommendation system. This sharing mainly introduces the basic practice and future planning of the real-time feature scenario of byte beating recommendation based on Flink SQL and Flink state.
Guowenfei, head of basic service of byte bounce recommendation system. Byte was added at the beginning of 2015, mainly responsible for recommending the basic service direction of the system, such as weight elimination, counting, features, etc.
Scenario based practice of Kwai building real-time data warehouse based on Flink
Litianshuo ｜ technical director of Kwai real time computing data team
As an important application output scenario of Kwai data, real-time computing plays an important role in large-scale activities such as Spring Festival Gala, operation system construction and feature construction. It provides a large data screen, various real-time data Kanban and real-time data push, and supports various application scenarios such as management decision-making, real-time monitoring and strategic services. This sharing will introduce some of our practices and Thoughts on real-time data research and development and real-time data warehouse construction.
He joined Kwai in 2019 and once worked in meituan. At present, he is the technical director of Kwai real-time computing data team, mainly responsible for the construction of real-time data warehouse, the improvement of real-time link SLA and quality assurance. He has been responsible for the construction and guarantee of real-time large screen and real-time application products for large-scale operation activities for two consecutive years.
Flink’s practice in iqiyi advertising business
Hanhonggen ｜ iqiyi Technical Manager
With the rapid development of effect advertising, information flow and other forms of advertising, it is very important for the platform and advertisers to better understand users and achieve fine delivery. In many business scenarios of iqiyi advertising, including large screen display, reports, feature engineering, system monitoring, etc., the requirements for data timeliness and data quality are becoming higher and higher. This sharing introduces the development history, technology selection, problems encountered in actual production and solutions of real-time computing around specific businesses.
Joined iqiyi in 2016, with more than 9 years of big data processing experience and rich practical experience in big data processing, real-time computing, etc. At present, he is mainly responsible for the real-time calculation, task scheduling management system, data access, etc. of iqiyi advertising business.
Agenda and registration
■ event agenda
■ event details
Time: 9:30-18:00, may 22
Venue: Fangheng fashion center, Haidian District, Beijing (byte beat)
Apache Flink community contributor exclusive benefits!
On May 22, Apache Flink meetup · Beijing station community contributor exclusive benefit! Apache Flink contributors, committers and PMC certified by the community can receive a limited number of newly customized mugs from Flink community on site for free.
Limited quantity, first come, first served!
We look forward to more small partners participating in the community contribution!
There are many good gifts at the event site
Flink community customized T-shirt (subject to the actual object)
Welcome to like Flink and send star~
For more Flink related technical issues, you can scan the code to join the community nail exchange group ~
Copyright notice:The content of this article is spontaneously contributed by Alibaba cloud real name registered users. The copyright belongs to the original author. The Alibaba cloud developer community does not own the copyright, nor does it assume corresponding legal responsibilities. Please refer to Alibaba cloud developer community user service agreement and Alibaba cloud developer community intellectual property protection guidelines for specific rules. If you find any content suspected of plagiarism in the community, fill in the infringement complaint form to report. Once verified, the community will immediately delete the content suspected of infringement.