Introduction to In-memory data grid IMDG

1Introduction

The memory as the primary storage medium is nothing new, and we have many examples of main memory databases (IMDB or MMDB) around us. In memory data grid (IMDG) is similar to IMDB in the use of main storage, but they are completely different in architecture. The IMDG feature can be summarized as follows:

? The data is distributed across multiple servers.

? Each server is active mode.

? The data model is usually object-oriented and non-relational.

? The server is often added or subtracted as needed.

In addition, theIMDG is different from the normal cache system. Similarly, the cache system is similar to IMDG in main memory usage and horizontal scaling. However, the use of the two methods and purposes is completely different. Caching systems are just buffering read pressure, and persistent storage like an RDBMS is a must. For example, the Arcus cache system in. Please refer to the second section for the IMDG architecture.

In other words,IMDG stores the object itself in memory and guarantees extensibility. Common business and open source products are as follows:

? hazelcast

? Terracotta Enterprise Suite

? VMware Gemfire

? Oracle Coherence

? gigaspaces XAP Elastic Caching Edition

? IBM eXtreme Scale

? JBoss Infinispan

2Architecture

IMDG the two core issues that need to be overcome are: capacity constraints and reliability. Typically,IMDG is designed to overcome limits on memory capacity through horizontal scaling, while ensuring reliability through a replication system. A typical IMDG architecture is shown.

Therefore, the difference between the cache system and the IMDG described earlier is obvious.

3features

In addition to providing a distributed implementation of various data structures,IMDG generally uses out-of-heap memory (off-heap, or elastic memory ) to reduce the pressure of garbage collection.

References

1 Introduction to in-memory Data grid:main Features

Introduction to In-memory data grid IMDG