HistorianAbout Setting up a Horizontally Scalable System
In a horizontally scalable Historian system, there are multiple Historian servers, all of which are connected to one another. This type of system is used to scale out the system horizontally. For example, if you have 5,00,000 tags in your Historian system, you can distribute them among the various servers to improve performance.
To set up a horizontally scalable system, you must first set up Configuration Hub.
- The Historian servers: You must install the following types of Historian
servers:
- Primary: A primary server is the only server in a system where
the Configuration Manager service runs. For the entire system,
Configuration Manager manages the system configuration licensed by the
user (that is, the number of tags, options, and so on). Each system can
have only one primary server.
You must apply the Enterprise license to the primary server.
- Distributed/Mirror: These servers collect and store data. If
added to a mirror group/location, you can achieve high
availability.
You must apply the Distributed license to the distributed/mirror servers.
- Primary: A primary server is the only server in a system where
the Configuration Manager service runs. For the entire system,
Configuration Manager manages the system configuration licensed by the
user (that is, the number of tags, options, and so on). Each system can
have only one primary server.
- A Historian system: A Historian system is a network of Historian servers
that collect, store, and retrieve data related to tags, alarms, and events.
By default, a system is created when you set up Configuration Hub.
- Data stores: A data store is a logical collection of tags used to store,
organize, and manage tags according to your requirements. The primary use of
data stores is segregating tags by data collection intervals. For example, you
can put name plate or static tags (where the value rarely changes) in one data
store, and put process tags in another data store. This can improve the query
performance.
By default, a user data store is created when you set up Configuration Hub. You can add more as needed.
- Locations: These are virtual entities in which data stores are created.
They are used for storage. The following types of locations are used in a
horizontally scalable system:
- Distributed location: This location is created automatically when you install a Historian mirror primary server, or when you install a Historian distributed/mirror node and add it to the primary server. You cannot modify or delete this location, and you cannot create another one.
- Mirror location: This location is used to replicate data collected in a data store. For more information, refer to About Data Mirroring.
- A collector instance: Collectors are the applications that collect data
from a data source, and send it to an on-premises Historian server or a cloud
destination such as Predix Time Series and Azure IoT hub.
You must add a collector instance to begin collecting data. You can choose the type of the collector depending on your need. You can use any existing instances (created during collector installation or ported during an upgrade).
- Tags: Tags are the parameters for which you want to store data (for
example, temperature, pressure, torque).
You must specify the tags for which you want to collect data.
- Data archiver: This is a service that indexes all the data by tag name
and timestamp, and stores the result in an .iha file.
By default, this is installed when you install the Historian server.
- Clients: These are applications that retrieve data from the archive files
using the Historian API.
By default, these are installed when you set up Configuration Hub.