Introduction to MeayunDB embedded high-speed Memory Database

Main features of MeayunDB

Cloud Platform Architecture

The cloud platform consists of N> = 1 MeayunDB sub-cluster. The applications on each sub-cluster are identical. The only difference is that the data stored in each MeayunDB sub-cluster is different. All your data is distributed to each sub-cluster on the cloud platform, and each sub-cluster only stores part of your data.

The number of MeayunDB instances in the MeayunDB sub-cluster must be greater than or equal to 1. The data of the MeayunDB instances in the same sub-cluster is identical and the same business applications are provided to the outside world, the MeayunDB instances in the same subcluster are mutually dependent data backups. You can increase the number of data backups by increasing the number of MeayunDB instances.

The cloud platform does not adopt a master-slave architecture, and there is no single point of failure (spof). As the business expands, the number of sub-clusters can be linearly increased to improve throughput, so as to easily handle storage and real-time analysis of hundreds of millions of rows of data.

MapReduce Process

After a user submits a task to the cloud platform, the cloud platform breaks down the user task, schedules the MeayunDB instance of the cloud platform, processes the user task in parallel, and finally merges the task result, the merged result can be used as the input for the next round of parallel computing.

MeayunDB does not move data, which reduces the communication overhead between client and server processes and computes data in the memory, improving the system performance as much as possible.

MeayunDB Performance

The software and hardware environment used in this test:

Hardware configuration: Intel (R) Xeon (R) CPU E5-2609 @ 2.40 GHz, 8-core 8-thread, 32 GB memory

Operating System: Windows Server 2008 R2 Enterprise

Data Table Structure:

1. query test:

MeayunDB instance	Number of Records)	Milliseconds)
Query instance 1 in a single thread	10000000	1641
Single-thread query instance 2	10000000	1590
Single-thread query instance 3	10000000	1246
Single-thread query instance 4	10000000	1593
Single-thread query instance 5	10000000	1484
Single-thread query instance 6	10000000	1694
Querying instances in a single thread 7	10000000	1376
Single-thread query instance 8	10000000	1581
80 million time consumed by the data to sum F5 of the double data column		2307
Time used to query each record		0.0288375 microseconds
Throughput per second (rows/s)		34677070 rows

2. Insert test:

MeayunDB instance	Number of Records)	Milliseconds)
Insert instance 1 in a single thread	10000000	59814
Time used to insert each record		5.9814 microseconds
Throughput per second (rows/s)		167184.93 rows

MeayunDB Value Analysis

1. Simple development, fast development, low technical requirements, and friendly to developers

2. High scalability and elastic scalability on demand

3. Seamless integration with relational databases

4. Reduce the impact of human factors and reduce project risks

5. Low latency, high concurrency, and microsecond-level data access efficiency.

6. Big Data Storage and Real-Time Parallel Computing

7. Management, development, and maintenance costs reduced by 50-80%

8. Improved work efficiency by 2-4 times

9. Performance Improvement 10-times