Hello, everyone!
The FusionInsight LibrA is an MPP system with a shared-nothing architecture. It is composed of multiple independent and shared-nothing logical nodes built based on system resources, such as the CPU, memory, and storage units. Service data is stored on more than one physical node in the shared-nothing architecture system and data analysis tasks are assigned to be executed in locations near the data storage locations. A considerable amount of data is processed in coordination with a control module, which enables rapid data processing responses.
The shared-nothing architecture is also called no-sharing architecture.
Figure 1-1 Comparison between the shared-nothing architecture and other architectures
The shared-nothing architecture has the following advantages:
Strongest scalability
Provides on-demand scaling for high-concurrent computing when processing huge amounts of data during BI and data analysis
Provides an automatic parallel processing mechanism
Automatic internal parallel processing without manual partitioning or optimization
Allows you to load and access data as you would in a common database.
Data is distributed on all parallel nodes.
Every node processes only partial data.
Optimal I/O processing
All nodes process data in parallel.
Nodes share nothing with each other, thus have no I/O conflicts.
Scale-out improves linear storage, queries, and loading performance.
That's all, thanks!