Shulou(Shulou.com)11/24 Report--

Zero: preface

The era of cloud native has arrived, and today bloggers share the practical experience of cloud solutions on Huawei's cloud database.

First, background introduction

With the rapid development of business in various industries, the amount of data is increasing, and the data type is becoming more complex. In order to meet different business needs, a variety of database deployment methods are widely used. But at the same time, these industries are also faced with many challenges, such as unable to achieve flexible capacity expansion, high operation and maintenance costs and so on.

Currently, the customer's database is mainly deployed in the following three ways:

Local database: the traditional local database deployment method is used to deploy the database in the IDC server room instead of using public cloud services.

Cloud ECS deployment database: in order to reduce costs and optimize architecture design, choose to deploy the database on the cloud vendor's server instead of using cloud database services.

Cloud database services: to improve data reliability, security, scalability and flexibility, purchase cloud database services directly from cloud vendors.

1.1. Pain points of the problem

When using the above deployment methods, users are mainly faced with the following problems:

The speed of business launch is slow: whether it is a local database or an ECS self-built database, users need to install their own software to solve patch upgrades, high availability and other problems, which is very time-consuming.

Unable to achieve flexible capacity expansion: resources cannot be quickly adjusted according to business needs. For example, after traffic growth, the self-built database cannot quickly upgrade the configuration, which may result in the loss of potential customers. When the peak of traffic falls down, it can not be reduced quickly, resulting in a waste of resources.

High manpower cost of operation and maintenance: self-built database requires users to deal with daily operation and maintenance problems on their own. When the number of instances reaches a certain extent, management will become very difficult, and the manpower cost of DBA will be high.

There are many security protection problems: self-built database requires users to download and install packages by themselves, which may be implanted with Trojans, posing a threat to the database. At the same time, if there is no disaster preparedness mechanism, users need to solve the security problems on their own. In addition, the lack of professional teams to provide security for the database is also a major problem.

1.2. Cloud solution on Huawei Cloud Database

Say the important thing three times: cloud on the database! Cloud on the database! Cloud on the database!

Core advantage

Flexible and flexible deployment

Auto scaling: cloud database is ready for use and can be flexibly scaled. Specifications can be raised or lowered, allowing enterprises to flexibly grasp resource input.

Pay-on-demand: pay for resources on demand, expand rapidly with business needs, and deploy flexibly to help customers deal with problems such as tight resources and slow response caused by business growth.

Intelligent operation and maintenance, reduce cost and increase efficiency

Intelligent operation and maintenance: compared with customers' original self-built database, cloud database is more convenient in backing up, adding read-only, monitoring alarm and so on.

Reduce operation and maintenance costs: let DBA have time to focus on more valuable business such as application optimization, and reduce DBA operation and maintenance costs. At the same time, Huawei Cloud has a professional data center operation and maintenance team to enable enterprises to focus on business development.

Lighten the burden on enterprises: enterprises do not need to spend a long time to build IDC, do not need high one-time investment, and do not have to worry about equipment depreciation.

Disaster preparedness is perfect, safe and reliable.

Perfect disaster preparedness mechanism Huawei cloud database has the mechanism design of active and standby, two places and three centers, recycling bin, security group and so on, which greatly improves the security and reliability of the database.

Strictly data neutrality: Huawei Cloud does not use technical means to illegally obtain customer data and does not commercially cash customer data.

Focus on infrastructure: Huawei Cloud focuses on building cloud infrastructure services to help the cloud on the enterprise database and master the core data independently

Data center stability: T4-level computer room, up to support 3AZ high reliability, meet customer high security needs, all 3 copies of data storage, cross-room cross-regional storage

Rich in security protection: provide 10 + security services such as WAF / DDoS to ensure business security.

Higher availability, stronger performance

Cloud database has stronger performance: compared with self-built stand-alone database, cloud database with active / standby and distributed architecture is more powerful to meet the needs of large amount of data.

Let's take several different database products of Huawei Cloud as an example to share the cloud on the database from zero to one.

2. Cloud database RDS for MySQL

Cloud database RDS for MySQL is a relational database management system. With excellent performance and LAMP, RDS for MySQL has become an efficient solution for WEB development.

2.1, advantages

Ultra-high performance, ultimate experience

High-quality hardware: Huawei after years of research, innovation and development of server hardware, stable, high performance.

SQL optimization: slow SQL detection is provided and corresponding optimization suggestions are given.

High-speed access: with elastic CVM in the same region, it communicates through private network to shorten application response time and save public network traffic fees.

Lower costs, more services

Using Huawei cloud database RDS for MySQL, you only need RDS for MySQL instance fee, no need to purchase and install any software and hardware, and simplify operation and maintenance operations. In contrast, the use of self-built database requires to pay server, system, database and other software and hardware costs + computer room hosting costs + operation and maintenance costs.

Full series of high security certification, database security without worry

Access control: inbound and outbound rules are restricted to control the range of the network that can connect to the database.

Security protection: under the protection of multi-layer firewall, it can effectively resist all kinds of malicious attacks and ensure data security.

Data encryption: transmission encryption is realized through TLS encryption and SSL encryption. Encrypt the data through static encryption and tablespace encryption.

Operation audit: cloud audit service that records operation events related to cloud database RDS instances.

Multiple deployment and disaster recovery solutions to meet a variety of availability requirements

Dual-server hot backup: the cloud database RDS service adopts hot backup architecture, which automatically switches over within seconds of failure.

Remote disaster recovery: you can use backup files to restore new RDS instances in remote areas to restore business.

Disaster recovery in the same city: support the deployment of multiple availability zones, interconnection of private networks between availability zones, and physical isolation of power and networks between different availability zones.

Strong scalability to meet business needs at different stages

Rich product specifications: provide stand-alone, active / standby, read-only instances, one-button stand-alone to active and standby.

Flexible capacity expansion: support adding and deleting read-only instances to cope with the surge / decline of business in a timely manner.

A variety of configurations: general-purpose (maximize CPU utilization through resource reuse, high cost performance), exclusive (completely exclusive CPU and memory, long-term stable performance).

Rich operation and maintenance tools to help focus on core business

Backup recovery: support by backup set and specified point-in-time recovery, 732 days automatic backup recovery, data is not lost.

Intelligent operation and maintenance: intelligent monitoring and diagnosis optimization of the database in all directions.

Monitoring alarm: customize the monitoring target and notification policy to keep abreast of the operation status of the cloud database RDS for MySQL service.

Differentiation advantage

Performance-to-price ratio: excellent product performance, ensuring leading or even industry, and more competitive in price

You can add read-only instances to stand-alone instances to meet the multi-read requirements of non-core businesses.

Support second-level monitoring, including 1-second monitoring and 5-second monitoring, to improve the instantaneous accuracy of monitoring indicators

Support to view the record of the instance to which the parameter template is applied.

2.2, applicable scenarios

For e-commerce customers

Cloud database RDS for MySQL provides reliable, economical and efficient data storage for e-commerce and mobile commerce applications, so that applications can run quickly and safely on the network.

Service value

RDS for MySQL can help customers effectively manage users, recharge, billing and other key data, and quickly change specifications to deal with traffic peaks.

For game customers

With its high reliability and high performance, cloud database RDS for MySQL can help customers easily deploy mobile online game services.

Service value

RDS for MySQL can help customers effectively manage data such as user, recharge, bill, shopping cart, etc.

For IoT customers of the Internet of things

Scale and reliability are critical, and cloud databases are required to provide fast response time to support a large number of connections

Service value

Cloud database RDS for MySQL provides high throughput and high concurrency, and can support a large number of connections through fast response.

For e-government customers

E-government faces the challenge of serving billions of people on a potential scale, requiring high-performance, high-throughput and highly available cloud services

Service value

The cloud database RDS for MySQL with high performance, high throughput and high availability can help customers easily cope with the challenge of the scale of potential users.

2.3. Purchase experience

Sign in to Huawei Cloud account first.

Log in to the administrative console and click purchase Database instance.

Please buy as needed. In order to learn, the blogger bought a stand-alone version with 4 cores of 8g. The recommended number of connections is 2500.

Note: the production database of an enterprise should use the active and standby instance type (the classic high availability architecture of one master and one standby).

Set up the network:

Set the password for the root account

The last step is to buy.

2.4, connection experienc

Bind an elastic public network IP

Select the instance you bought, click the instance name, and go to the basic Information page of the instance.

Select the elastic public network IP, click "Yes", and submit the binding task.

So the binding is successful.

Set security group rules

Select the instance you bought, click the instance name, and go to the basic Information page of the instance.

On the left navigation bar, click connection Management, and at the current Security Group in the Security Group rules module, click the security group name to go to the security group page.

Direction rules and exit direction rules can be added.

Connect RDS for MySQL instances via public network

Use client-side navicat

The password is the password set when you buy it.

Test the connection.

The next operation on the connection is the same as the local operation.

3. Redis goes to the cloud

As an enterprise-level Redis of Huawei Cloud, Cloud Database GaussDB (for Redis) is committed to providing enterprise customers with stable and reliable core data storage services with ultra-high concurrency and rapid elastic expansion. GaussDB (for Redis) adopts cloud native distributed architecture, is fully compatible with Redis protocol, supports rich data types, and provides real-time data persistence, strong consistency protection for multiple copies, as well as real-time monitoring, auto scaling, automatic backup and other one-stop services.

3.1, advantages

High availability

If a node fails, the takeover can be completed within seconds. Even if the NMUI 1 node fails at the same time, as long as one instance is available, the business can still operate normally, minimizing the impact on the business.

High huff and puff

Multi-node can be written at the same time, and multi-threaded architecture is adopted, throughput is easily doubled, QPS can scale horizontally, and effectively withstand traffic peak.

Strong consistency

The redundancy of three copies of data is realized through GaussDB basic component service, which ensures the loss of 0 and meets the strong consistency requirements of multi-point access at the same time, which makes business design easier.

Large capacity

The computing and storage separation architecture is adopted to fully support large data scenarios, and the product positioning breaks through the limitation of "cache", which can help the business to carry more important data for a long time.

Second capacity expansion

Adjust the granularity of 1GB as needed at any time, and take effect in seconds. GaussDB basic component service enables smooth storage scaling and business zero awareness.

Reduce cost

Compared with the pure memory architecture, the hardware cost is reduced by 75% to 90%, and almost all the capacity is available, and it can be purchased according to the fine granularity of 1GB without excessive waste.

Differentiation advantage

High reliability-zero data loss, strong consistency, stable performance

Low cost-75% price savings and larger capacity

Elastic expansion-smooth expansion, second-rate experience

Security operation and maintenance-professional cloud service, one-stop operation and maintenance

A single node supports maximum 1TB, and data can be stored persistently to ensure strong and consistent data storage.

Limit the access of the source IP to the instance through the security group, which can be refined to the port

The permissions to access the account can be set and managed through the console

Customers can customize the retention time of backup files in the console, with a maximum retention period of 35 days to meet customer regulatory needs.

Customers can restart the instance and modify some running parameters in the console.

3.2. Applicable scenarios

For e-commerce customers

During the promotion period, there will be a large number of concurrent visits, and there are both popular and unpopular goods in commodity search, and there are high requirements for data consistency.

Service value

Using GaussDB (for Redis) as the front-end cache (need to configure large memory) can relieve the pressure on the back-end database and smoothly ride through the business peak; at the same time, the hot product information resides in memory as hot data, and the unpopular product information is replaced into the shared storage pool, which can further reduce the database pressure. In addition, strong consistent counter makes flash sale activity no longer "avalanche", inventory deduction is no longer worried about "oversold", and sorted set can be used to help build spam mechanism and geo to achieve logistics tracking.

For game customers

There are some scenarios that are very sensitive to delay, such as cross-server games, and it is necessary to build agile, iterative and fully functional game systems.

Service value

GaussDB (for Redis) provides customers with a wealth of data types to help build efficient game systems. By using the sorted set data type, you can easily build real-time rankings to reflect players' rankings in real time. On the other hand, the pub / sub mechanism can realize the function of real-time chat room, so that players can keep communicating anytime and anywhere. In addition, GaussDB (for Redis) also supports the deployment of dynamic servers to update players'"moments", so that players can share their game experiences and experiences at any time. The account server is used to store basic user information and ensure the security of players' accounts. The friend server is used to persistently store a large and complex network of players, making it easy for players to find and keep in touch.

What is more worth mentioning is that GaussDB (for Redis) can be used as a front-end cache to speed up application access and effectively reduce latency by configuring large memory. This is of great significance to improve the game experience and response speed. In a word, with its rich data types and flexible deployment methods, GaussDB (for Redis) provides a comprehensive solution for game developers and helps to build a stable and efficient game system.

For IoT customers of the Internet of things

A database with high reliability, high availability and high performance is needed to ensure the stable development of business and prevent blocking problems.

Service value

For example, multi-city vehicle charging platforms often need to manage a wide range of endpoint devices. In order to make more terminals easily interconnected and safely managed, we can build a highly concurrent task scheduling system, end device monitoring and more edge computing power with the help of GaussDB (for Redis).

3.3. Purchase experience

Sign in to Huawei Cloud account first.

Log in to the management console and select Cloud Database GaussDB (for Redis).

Buy on demand, the blogger here chooses to buy 2-core 24GB.

Select a network:

Set the password:

The last step is to buy.

3.4, connection experience

Get the Redis client:

Extract the client kit:

Connection instance:

The following message appears indicating that the connection was successful:

4. InfluxDB goes to the cloud

GaussDB (for Influx) is a cloud native time series database that adopts computing storage separation architecture and is compatible with InfluxDB ecology. Provide high-performance read and write, high compression rate, hot and cold layered storage, flexible capacity expansion, monitoring alarm and other service capabilities. It is suitable for monitoring application scenarios such as operation and maintenance, Internet of things, industrial production and so on.

4.1, advantages

Extreme performance

Billion-level timeline data management, easily support trillions of records written every day, aggregate query performance is excellent.

Massive data analysis

Support multi-dimensional conditional combination query, aggregate analysis, timing insight, and provide full scene support.

Low-cost storage

Column storage, adaptive compression algorithm, automatic hot and cold hierarchical storage, low storage cost.

Extreme elasticity

Computing and storage can be expanded independently on demand, without the need to migrate data, and the expansion and scaling of database clusters can be completed quickly.

4.2. Applicable scenarios

For IoT customers of the Internet of things

Time series data analysis of IoT sensors helps the Internet of everything.

Service value

Super write: the performance of the small batch write scenario is 4.5times that of the open source version InfluxDB, and the performance of the mass write scenario is 3.3times that of the open source version InfluxDB.

Multi-dimensional analysis: support multi-dimensional conditional combination query, easy to analyze massive data

Flexible expansion: based on the distributed architecture of computing and storage separation, the capacity of minute-level computing nodes is expanded to cope with business peak periods.

For application operation and maintenance monitoring

Unified application of operation and maintenance real-time monitoring, global control of the running status of the system.

Service value

Massive data support: support trillions of monitoring metrics per day, write performance linear expansion

Anomaly detection: supports detection and alarm of multiple timing anomaly types, visual display of anomalies, convenient for operators to identify anomalies, simple alarm policy setting, and no machine learning background.

Intelligent prediction: the multi-dimensional time series prediction algorithm is adopted to improve the prediction accuracy by using the correlation between multi-indexes, which is higher than the traditional prediction algorithm.

4.3. Purchase experience

Sign in to Huawei Cloud account first.

Log in to the administrative console and select GaussDB (for Influx).

Buy on demand, the blogger chooses to buy 4-core 16:

Network settings and password settings:

The last step is to buy.

4.4, connection experienc

Connect the instance through Java language:

Pom.xml dependency add:

Example of Java code for connecting an instance using SSL:

An example of Java code that connects an instance in a non-SSL way:

Example of Java code for connecting an instance using connection pooling:

5. Cloud on document database

Huawei Cloud document Database Service (Document Database Service, referred to as DDS) is fully compatible with MongoDB protocol, providing secure, highly available, highly reliable, scalable and easy-to-use database services. It supports three deployment architectures: sharded cluster (Sharding), replica set (ReplicaSet) and single node. It also provides one-click deployment, elastic capacity expansion, disaster recovery, backup, recovery, monitoring and alarm functions.

5.1, advantages

Flexible and diverse instance types

The document database service DDS supports cluster instances, replica set instances, single-node instances and other instance types to meet the needs of different business scenarios.

Flexible configuration

The document database service DDS provides the function of changing instance storage space, number of nodes, CPU and memory specifications. Customers can flexibly configure the required resources according to their business needs.

Flexible management visual control

The management console provides comprehensive monitoring information, easy to use, flexible management, visual and controllable

Multi-tier data security

Effectively resist various malicious attacks and ensure data security through multi-layer security protection systems such as virtual private cloud, subnet, security group, storage encryption, DDoS protection and SSL security access; provide audit log function, record customer operations on the database or collection, help customers to carry out security audit, fault root cause analysis and other operations on the database; support fine-grained permissions to achieve fine-grained rights management

Rich tools to simplify operation and maintenance

In terms of backup and recovery, full backup and incremental backup are supported, and backup recovery is supported at any point in time, and backup data can be saved for up to 732 days. In terms of monitoring and alarm, you can monitor the running metrics of the instance in real time. Users can customize monitoring targets and notification policies to keep abreast of the operation of the database service.

Differentiation advantage

1. Support single-node deployment, solve test scenarios without high availability requirements, and save resource costs.

2. Support users to set cross-region backup policies, including on / off, backup area, and backup retention period.

3. The instance supports the migration of availability zones to solve the problem of disaster recovery at the AZ level of users.

4. Provide error log display and download function

5. Support tenants to use their own keys uploaded on KMS services to encrypt DDS data at rest to improve data security. Some foreign companies have encryption requirements, and it is a necessary threshold for cloud hosting.

6. Read-only replication delay: customers can set the replication delay time of read-only nodes in the console and quickly retrieve data through delayed read-only nodes in case of emergency (mistakenly deleting data).

5.2. Applicable scenarios

For large enterprises and Internet customers

Often need to deal with TB-level data, there are strong data storage requirements. And big data scenario, but also need to meet the needs of real-time query and dynamic analysis, such as real-time writing of business online data to database, big data calculation and analysis, analysis result feedback.

Service value

The replica set mode of DDS adopts the high availability architecture of three-node Replica Set, and the three data nodes form an anti-affinity group, which is deployed on different physical servers and synchronizes data automatically. The Primary node and the Secondary node provide services, and the two nodes have independent private network addresses respectively, which cooperate with Driver to achieve read pressure distribution. It can not only meet the needs of large enterprises and Internet industry customers for massive data storage, but also solve the needs of data analysis scenarios.

For game customers

In the game application, the game player is active at the peak and has a high demand for concurrency ability. at the same time, with the development of the business, the table structure may need to be changed when the game method changes, so it is difficult and heavy for operation and maintenance.

Service value

You can store your game equipment and points and other data in the DDS database. By adopting the cluster type of DDS, you can effectively deal with high concurrency scenarios and ensure the stability and availability of the game. In addition, the replica set and cluster architecture of DDS have high availability features, which can meet the requirements of continuous and stable operation of the game in high concurrency scenarios.

DDS is compatible with MongoDB and uses No-Schema, which means that when the game changes, you don't have to change the table structure. This flexibility is well suited to deal with the changing needs of the game business. You can easily manage structured data with fixed schemas by storing it in the cloud database RDS. At the same time, efficient data access can be achieved by storing flexible business data in DDS. To reduce the input cost of storing data, you can store hot data in GaussDB (for Redis). This approach optimizes data storage and management, thereby reducing operating costs. Through the rational use of these technologies and tools, you can build an efficient, stable and flexible game data storage solution to meet the changing needs of the game business.

For IoT customers of the Internet of things

IoT intelligent terminals often need to carry out a variety of data collection, and need to store device log information, and carry on multi-dimensional analysis of these information. At the same time, with the rapid growth of IoT business, large amount of data and rapid growth of visits, data storage needs the ability to scale horizontally.

Service value

DDS is compatible with MongoDB, with high performance and asynchronous data writing capabilities, and can achieve the processing capacity of in-memory database in specific scenarios. At the same time, the cluster instance in DDS can dynamically expand the capacity and increase the performance specification and number of mongos and shard components, and the performance and storage space can be expanded rapidly, which is very suitable for IoT scenarios with high concurrent writes.

5.3. Purchase experience

Sign in to Huawei Cloud account first.

Log in to the administrative console and select DDS.

Buy on demand, what the blogger chooses is the copy set, 4-core 8G.

Network Settings:

The last step is to buy.

5.4. Connectivity experienc

Java example (no certificate connection):

When connecting to a cluster instance through Java, the format of the Java link in the code is as follows:

Parameter description:

Summary

This article shares a lot of content, analyzing the current situation and pain points of the pan-Internet industry, such as games, e-commerce, social networking and so on. The main problems involved are: slow online speed of business, inability to flexibly expand and reduce capacity, high labor costs of operation and maintenance, and many security protection problems. The blogger suggests that the solution is: cloud solution on Huawei Cloud Database!

There are four main core advantages of the cloud on the database:

Flexible and flexible deployment

Intelligent operation and maintenance, reduce cost and increase efficiency

Disaster preparedness is perfect, safe and reliable.

Higher availability, stronger performance

Through the study of this article, we can have some understanding and practical experience of cloud on Huawei cloud database.

In addition to theoretical sharing, this paper also introduces the sharing of cloud practices of different database types. The main databases are cloud database RDS for MySQL, cloud database GaussDB (for Redis), GaussDB (for Influx) and Huawei cloud document database service (Document Database Service, referred to as DDS). From the advantages to the applicable scenarios, to the purchase experience and connection experience in detail.

At present, it is Huawei Yunshuang 11 marketing season. Cloud database RDS for MySQL 1 core 2GB costs only 15.3 CNY / 3 months. There are also products and services such as cloud database GaussDB (for Redis) and GaussDB (for Influx) that can be used first and bought later, so that the database can be easily put on the cloud. If you are interested, go to Huawei's cloud official website and have a look.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.