How to Use Real Time Computing Flink Version Correctly

1. What is the real-time computing Flink version?

The real-time computing Flink version (Alibaba Cloud Realtime Compute for Apache Flink, Powered by Ververica) is an enterprise-level, high-performance real-time big data processing system built by Alibaba Cloud based on Apache Flink. It is officially produced by the founding team of Apache Flink and has a global unified commercial brand , is fully compatible with the open source Flink API, and provides a wealth of enterprise-level value-added functions.

Based on the core functions of Apache Flink, the real-time computing Flink version also enhances cluster stability, performance optimization, security control, system monitoring, and job management that enterprise users are concerned about.

At the same time, it can support sub-second-level processing delays on PB-level data sets, empowering users with standard real-time data processing processes and industry solutions; while supporting Datastream API job development, it provides complete SQL semantics, enabling BI scenarios The development under the environment becomes easier; the rich upstream and downstream connectors ensure the seamless connection with the big data components already used by users; the intelligent job tuning and diagnosis functions further simplify the development and use of users.

Alibaba Cloud's "real-time computing team" is currently the world's largest Flink team with the largest number of committers and the most professional expertise, providing enterprise-level management and consulting services for real-time computing users. In June 2019, the results of the big data product capability evaluation initiated by the Data Center Alliance were authoritatively released. Alibaba Cloud's real-time computing Flink version passed the newly formulated basic capability evaluation of the distributed stream processing platform and became the first batch of domestic stream computing product capabilities to pass the evaluation. products, and was delineated by the Data Center Alliance as the first echelon of domestic big data flow computing basic platforms. In 2019, in the evaluation of Forrester, a well-known international consulting and research company, the real-time computing Flink version became the only real-time stream computing product in China that entered the Forrester quadrant.

2. Comparison between real-time computing Flink version and Apache Flink

3. Real-time computing Flink version product architecture

4. Why use the real-time computing Flink version?

Product advantages:

• Superior performance: Single-core CPU can process hundreds of thousands of records per second, with sub-second data processing delay between ports, supporting tens of thousands of concurrent ultra-large-scale real-time task calculations.

• Powerful functions: a one-stop SQL development and maintenance platform, intelligent diagnosis and automatic configuration and optimization, and seamless connection with Alibaba Cloud's mainstream data products.

• High quality and low price: Single-core CPU computing is as low as 0.25 yuan per hour, and it can be elastically scaled and paid according to the load. The TCO is much lower than that of IDC self-built.

• Stable and reliable: 99.9% SLA stability guarantee, monitoring and alarming of all link indicators, after years of double eleven inspections by Alibaba.

• Compatible with open source: 100% compatible with Apache Flink, supports smooth migration of open source Flink to the cloud, and seamlessly connects with mainstream open source big data ecology.

• Brand certification: Officially produced by the founding team of Flink, authoritatively certified by the China Academy of Information and Communications Technology, and the only real-time stream computing product in China that has entered the Forrester quadrant.

5. What are the application scenarios of the real-time computing Flink version?

Real-time Computing Flink Edition - Product Solutions

1. Internet trading industry solutions

With the rapid development of the domestic Internet trading industry, the original "large-scale" and "spreading the net" operation methods can no longer meet business needs. Quickly mine user characteristics through real-time computing, analyze needs and preferences, and help enterprises complete the transformation in the direction of digital, refined, and personalized operations.

Real-time user portraits: make real-time and efficient personalized recommendations for different types of customers, and improve the conversion rate of orders

Real-time association recommendation: establish real-time association analysis after customer purchase information to increase the probability of second purchase

2. Online education industry solutions

In the future, the K12 online education market will continue to expand, and the combination of big data and education will become closer. By using the Flink version of Alibaba Cloud's real-time computing, it can effectively solve the real-time/accurate computing problems faced in key online education scenarios such as renewing reports, resuming classes, and monitoring the video quality of live classes. At the same time, the automatic elastic expansion and contraction capabilities of the product can be used to effectively cope with traffic peaks and better meet the needs of complex and changing scenarios.

Efficient data processing: meet the needs of real-time analysis and processing of data from different sources (logs, databases, business cluster data, etc.)

Multi-role permission setting: Set real-time report viewing permissions that meet their needs for different roles in the education industry to improve work efficiency.

Business scenario support: According to the business logic of online education, for real-time BI scenarios, with perfect SQL support, to escort the business.

3. Game industry solutions

As an emerging entertainment industry, games have attracted more companies and talents to invest in them due to their ability to attract money and their fun. With the rise of app stores and social networks, the game market has expanded unprecedentedly, and the positive effects of big data and artificial intelligence on game operations, especially in prolonging the life of games, have become more and more obvious.

Forum data tracking: According to the communication and control of player forums, obtain the current players' core attention information and game BUG detection to ensure the fairness of the game.

Operational data control: Real-time analysis of game operation activities is carried out through real-time computing technology to obtain the most effective user conversion channel and improve the effect of operation promotion.

Behavior data analysis: According to the real-time behavior data analysis of players, establish player user portraits, customize different game welfare modes, and improve player retention and conversion rates.

4. Financial industry solutions

In recent years, with the rapid development of the Internet financial industry, various businesses of traditional financial institutions (such as exchanges, securities companies, banks, etc.) are more and more closely integrated with the Internet. Provide overall solutions for the demand link, risk control, credit evaluation, fraud risk, data quality and other needs of banking business. Taking financial institutions as an example, building a real-time data warehouse and anti-fraud system through real-time computing Flink version helps financial institutions quickly build a real-time risk control system.

Fraud detection: Real-time monitoring and differentiation of user accounts and transaction behaviors, quickly identify transaction information, and reduce user losses;

Risk control requirements: optimize the traditional risk control system through real-time calculation of the flunk version, transform intelligent risk control, improve the overall recognition efficiency and apply multiple scenarios;

Credit evaluation: quickly identify malicious fraudulent behaviors such as false information and information counterfeiting, identify high-risk groups, and shorten the review process through intelligent methods;