Distributed Database Middleware

Distributed Database Middleware (DDM)

Distributed Database Middleware (DDM) is a database middleware service (to be developed into a distributed relational database service) that is compatible with MySQL licenses. It uses a storage-compute decoupled architecture that makes it easy to scale out compute and storage resources to handle a huge number of concurrent requests.

Reasons for Distributed Database Middleware in Open Telekom Cloud

Icon folder and two arrows pointing up and down

Data import and export

External data can be imported to help you migrate databases to the cloud, and DDM instance data can be exported based on service requirements.

Online monitoring

DDM monitors the read/write ratio and slow SQL logs on the console, helping you detect resource and performance bottlenecks as fast as possible.

DDM console

The DDM console enables you to manage and maintain DDM instances and schematics online.

Key Features of Distributed Database Middleware

Hands on a laptop keyboard displaying digital icons

Scaling

DDM is scalable in concurrency, computing, and data storage, freeing you from worrying about scalability and O&M issues as your services grow.

DDM is highly scalable, easy to use, quick to deploy, and cost-effective and can deliver high performance.

Highly scalable

Provides automatic horizontal sharding
Allows you to change your instance class within minutes
Allows you to add instance nodes without impacting applications
Allows you to add DB instances with minimal downtime.

Easy to use

Compatible with MySQL licenses, syntax, and clients
Makes it easy to import data and migrate databases to the cloud
Splits read and write requests with no need to modify service code

Quick to deploy

DDM instances are easy to deploy online. This shortens project cycles and helps migrate your services to the cloud. You do not need to purchase, deploy, or configure a DDM instance as you do for a self-built database.

Cost-effective

DDM provides stable performance, comprehensive O&M, and excellent technical support. With these advantages, it is more cost-effective than open-source products. Different instance classes are available, so there is always a choice well suited to your specific needs.

Architecture Use Case and Service

Using a DDM instance is as if you are still using a single-node database no matter how many database shards there are. You can perform database O&M on the DDM console and connect to a database using the Java database connectivity (JDBC) driver or a MySQL client to read and write data, just as you would do with a single-node database.

Supported functionalities

Function	Description
Horizontal sharding	When creating a distributed database, select a sharding key and DDM will generate a sharding rule to horizontally shard data by the sharding key.
Scaling	DDM supports both compute and storage scaling. You can scale a DDM instance up or out by adding nodes. Compute scaling is undetectable to your applications, and storage scaling minimizes service interruption to seconds. Storage scaling can be implemented with two methods: rebalancing or resharding. Rebalancing keeps the total database shards unchanged. Resharding doubles the number of database shards.
Distributed transactions	DDM provides three types of transactions, including single-shard, FREE, and XA. Single-shard: means that transactions cannot be committed across shards. FREE: allows data to be written to multiple shards. If a transaction fails to be committed in some of the shards, the transaction is not rolled back which causes data inconsistencies. XA: allows transactions to be committed in two phases. If a transaction fails to be committed in some of participating database shards, all work done will be rolled back to ensure data consistency.
SQL syntax	DDM is highly compatible with the MySQL licenses and syntax.
Read and write splitting	Read and write splitting is transparent to applications. You only need to create read replicas for a DB instance associated with your DDM instance and configure a read policy, without the need to modify application code. This function improves database performance and access efficiency, helping your applications process high-concurrency transactions.
Global sequence	DDM supports globally unique, distributed, and ascending SNs to meet requirements for primary or unique keys, or other specific requirements.

Available flavors

Class	Architecture	vCPUs	Memory (GB)
General-enhanced	x86	8	16
		16	32
		32	64

Restrictions

Currently, only MySQL DB instances of versions 5.7 and 8.0 are supported.
DDM does not support SSL for MySQL DB instances.
When you modify configurations of a MySQL DB instance associated with DDM, an exception may occur. After the modification, you need to click Synchronize DB Instance Data on the DDM instance details page to synchronize the change to DDM.

Syntax restrictions on DDL

Item

Restriction

DDL syntax

Sharding keys cannot be modified.
ALTER DATABASE Syntax is not supported.
ALTER cannot be used to modify sharding keys or global sequence fields.
TEMPORARY sharded and broadcast tables cannot be created.
Sharded or broadcast tables cannot be created from another table.
Only un-sharded tables support foreign keys. Sharded and broadcast tables do not support such keys.
Un-sharded tables can be created only from another un-sharded or broadcast table and not from another sharded table.
CREATE TABLE Syntax and DROP TABLE Syntax do not support annotations and are not restricted by the sql_execute_timeout parameter.

Syntax restrictions on DML

Item	Restriction
DELETE statement	PARTITION clauses are not supported. Subqueries are not supported.
UPDATE statement	PARTITION clauses are not supported. Subqueries are not supported.

Restrictions on database administration statements

Item	Restriction
Database administration statements	Executing SET Syntax to modify global variables is not supported. SHOW TRIGGERS Syntax is not supported.

Restrictions on advanced SQL functions

Item	Restriction
Advanced SQL functions	Customized data types and functions are not supported. Views, stored procedures, triggers, and cursors are not supported. Compound statements such as BEGIN...END, LOOP...END LOOP, REPEAT...UNTIL...END REPEAT, and WHILE...DO...END WHILE are not supported. Process control statements such as IF and WHILE are not supported.

Use Cases

Internet

In e-commerce, finance, Online to Offline (O2O), retail, and social networking applications, core transactional systems slow down when there is too much data generated by large user bases and frequent marketing activities, which impairs business development. DDM can be scaled out to improve database processing capabilities and access efficiency, helping your applications handle high-concurrency transactions.

IoT

A large number of sensors and monitoring devices are used in industrial monitoring, remote control, smart city extension, smart home, and Internet of Vehicles (IoV), where samples are frequently collected, involving huge amounts of data beyond the storage capacity of single-node databases. DDM provides horizontal capability expansion to help you inexpensively store massive data.

Traditional industry

Government agencies, large-sized enterprises, and banks use commercial solutions to support large-scale data storage and high-concurrency database access. These solutions are expensive because they need to rely on mid-range computers and high-end storage devices. DDM, deployed in clusters of general computing ECSs, provides cost-efficient database solutions with the same as or even higher performance than traditional commercial database solutions.

Find out more

Documentation

Distributed Database Middleware