Storage Facilities

The server can use different storage types for different types of data. Here is a comparison of the available storage types and their feasibility:

Storage Type

Applicable For

Pros

Cons

NoSQL Database

Configuration, Events, Binary Data

High insert/update performance. Storage performance scalability via storage clustering. Failover clustering is supported.

Cannot be used to store topologies.

Relational Database

Configuration, Events, Binary Data, Topologies

Works out-of-the-box. Data is directly available for third-party applications. Native and Iotellect Server-based failover clustering are supported.

Low performance on data inserts and updates.

File Storage

Configuration, Binary Data

Very low system performance overhead.

Failover clustering is not available.

Disabled

Events, Binary Data

No performance overhead.

Data is not stored persistently. All historical values are not preserved.

Round-Robin Database

Statistics (aggregated time series)

Constant data footprint. Highest insert/update/retrieval performance.

Only time series aggregates of numeric values can be stored.

Graph Database

Topologies

High-performance, specialized tools for graph processing.

Large topologies may require a standalone commercial version of a graph database.

Key-Value Database (deprecated)

Configuration, Binary Data

High insert/update performance. Failover clustering is supported.

Not suitable as an event storage.

Default Storage Configuration

By default, Iotellect Server is configured to keep data in a NoSQL Database on Unix systems, and the bundled MySQL database in Windows installations. If the bundled MySQL database was not installed, the server will rely to the embedded relational database engine that’s not designed for production usage.

Database section of the server’s global configuration provides a way to change storage type independently for each type of stored data.

Was this page helpful?