Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.Original topic: 【TiDB 社区智慧合集】2022 TiDB 社区用户实践案例
|
username: Billmay表妹
This article will bring you various user practice cases in the TiDB technical column from 2022 to the present, including those from: Ctrip, DMALL, NetEase, BaiRong, Getui, McDonald’s, Sina, CCB Fintech, and Cisco Webex.
TiDB Implementation at Getui | Solving Hotspot Issues and Improving Performance by Over a Thousand Times
As a data intelligence company, Getui provides messaging push and other developer services to hundreds of thousands of apps, while also offering professional digital solutions to numerous industry clients. As the business rapidly grows, the company’s data volume is also increasing at a high speed. Over time, the data volume has become so large that MySQL can no longer meet the company’s needs for fast data querying and analysis. A new type of database that supports horizontal elastic scaling, can effectively handle high concurrency and massive data scenarios, and is highly compatible with MySQL has become Getui’s selection requirement.
Exploration and Practice of Distributed Database TiDB at BaiRong Cloud
Like every tech company, when a company’s business and data grow rapidly, it is the most challenging time for the technical department. In mid-2021, the rapid development of BaiRong Cloud’s SaaS-end data analysis and AI business brought new challenges to our technical platform.
Application of TiDB in DMALL’s Digital Retail Scenario
This article is based on the sharing by Feng Guangpu, head of the DMALL database team, at the TUG Enterprise Tour Chengdu Station, introducing the usage of TiDB in DMALL’s digital retail scenario, core business scenario support, value analysis, and experience summary.
Application Exploration of TiDB in Chain Fast Food Enterprises | Massive Transactions and Real-time Analysis
For many modern people, a tasty, energy-packed, quality-assured, and quick and convenient burger, fries, and fried chicken is indeed a good choice during busy work. Moreover, the salt, sugar, fat, and carbohydrates in fast food are more likely to give people a sense of satisfaction. But while enjoying the happiness brought by fast food, have you ever wondered how fast food enterprises manage to perform high-level digital operations such as online coupon purchase, offline pickup, home delivery, and in-store pickup while operating hundreds or thousands of stores?
Best Practices for Deploying TiDB v5.1.2 Production Environment on Bank & Kylin v10
The author recently conducted PoC testing for a bank project. Since the client chose to use the TiDB database, the author selected a relatively stable and bug-free version: TiDB v5.1.2. Although there are fewer bugs, some issues were inevitably discovered during testing and were resolved through parameter adjustments. After PoC testing and solution formulation, the production environment deployment began. Deploying a production environment is not as simple as deploying a test environment. The test environment can be simplified by reducing some non-essential configuration items, such as whether to use the performance mode for the CPU frequency cpufreq module or whether to set the I/O scheduler for the storage medium to noop. However, in the production environment, all optimization items must be set up, even if it only optimizes a little bit. This article will organize the deployment architecture and the considerations for the production environment in the project to provide some references.
NetEase Games & TiDB Cold and Hot Storage Separation Solution
TiDB 6.0 officially provides the data placement framework (Placement Rules in SQL) function. Users can configure the placement of data in the TiKV cluster through SQL to directly manage the data. Let’s see how NetEase uses this function to archive 330TB of cluster data.
DMALL & TiDB Practice of Splitting Hundreds of Terabytes of Data
To improve TiDB availability, it is necessary to split the existing hundreds of terabytes of TiDB clusters into two sets.
TiDB at Ctrip | Optimization Practice of Real-time Tag Processing Platform
Ctrip is a leading global one-stop travel platform, owning brands such as Ctrip.com, Qunar.com, and Skyscanner. Ctrip.com provides hotel booking, hotel reviews, special hotel searches, flight booking, flight ticket searches, timetables, fare searches, and flight searches to over 90 million members. With billions of data, Ctrip leverages TiDB HTAP capabilities to improve business operation efficiency. The application scenarios mainly include international business CDP platform, hotel settlement, and risk control.
Have You Learned NetEase’s Impressive Migration Solution? [DDB Migration to TiDB Solution Design]
Currently, several businesses in the company have gone live with TiDB services, including NetEase Payment Reconciliation Center and NetEase Cloud Music Heartbeat Chart System. However, these are all new businesses directly launched on TiDB. To explore the migration of existing businesses to TiDB, this article summarizes some migration solutions.
New Energy Industry (Gas) & Mixed Deployment Practice of TiDB Cluster in Production Environment
Due to various external factors, we cannot freely choose the ideal hardware environment. Moreover, the hardware configuration of a single physical machine is often higher than the demand. To plan resources more reasonably, a single server cannot “luxuriously” deploy only one instance. Instead, we consider deploying multiple instances of TiDB or TiKV on a single machine. This requires building a TiDB cluster that meets high availability and high performance in the existing environment as much as possible. This article mainly shares the process of mixed deployment of TiDB clusters in an actual production environment for reference.
Logistics System & A Small Case of Performance Improvement by 666 Times Illustrates the Importance of Correctly Using Indexes in TiDB
Recently, I conducted a TiDB PoC test for a logistics system. This system is developed based on MySQL. The business data involved in this test includes about 10 databases with approximately 900 tables, with the largest single table having more than 60 million rows. This scale is not large. The test data and database table structure were exported from MySQL using Dumpling and imported into TiDB using Lightning. The entire process went very smoothly. After the system was running on TiDB, a SQL query regularly appeared on the slow query page through the Dashboard. Upon opening the SQL, it was found to be a simple single-table query, which seemed suspicious.
Sina & A Small Operation, SQL Query Speed Increased by 1000 Times
Application Practice of Chaos Engineering in CCB Fintech
Currently, the financial industry is developing rapidly, with increasing demands and rapid product iterations. As the system scale expands, traditional monolithic architectures can no longer meet the system’s development needs. Distributed microservice architectures are increasingly being applied in the industry. The financial industry involves a large number of fund transactions, covering complex infrastructures such as multi-data centers, multi-active, disaster recovery, containers, and virtual machines. The interaction between systems is very complex. The application of distributed technology makes the infrastructure more complex compared to traditional architectures, and the system’s runtime state is uncertain. The characteristics of financial business place extremely high requirements on the system’s stability, availability, and reliability.
Deployment and Application of TiDB in Cisco Webex Architecture
Since its inception, Cisco Webex has been committed to bringing people using the internet from all over the world together, collaborating creatively, and conducting business. It enables users to expand learning, special events, and multimedia presentations based on integrated real-time video and audio technology, saving the time and cost of traditional meetings. Today, Webex has become the world’s number one web conferencing provider.
Making Flash Sales More Manageable: The Database Behind Big Promotions (Part 2)
The annual 6.18 shopping festival is here again. The best gift for tech people is a technical guide! Over the years, shopping festivals have not been limited to the e-commerce industry. Nowadays, various industries adopt similar methods for promotional activities. The automotive industry has 818, Xiaomi has Mi Fan Festival, etc. This poses many new challenges to basic software, including databases, and has accumulated many best practices. PingCAP has conducted in-depth discussions with users such as JD.com, ZTO Express, Autohome, and Bitauto to reveal the technical challenges behind the soaring sales year by year. What technical architecture can withstand the traffic peaks smoothly? This article is the second part of the “Database Behind Big Promotions” series, introducing the application practice of TiDB in the Super Auto Carnival.
Dry Goods | Practice of Distributed Database TiDB at Ctrip
Since around 2014, Ctrip has fully adopted MySQL databases. With business growth and data volume surge, single-instance bottlenecks gradually appeared, such as increased query time for historical data due to large single-table rows and insufficient disk space due to large single-database capacity. To address these issues, we have taken various measures such as horizontal splitting of databases and tables, read-write separation with one master and multiple slaves, hardware SSD upgrades, and adding front-end Redis caches. However, these measures also made the entire business layer architecture more complex and could not achieve transparent elasticity. Therefore, we began to shift our focus to distributed databases to solve these pain points.