Ontotext Platform 3.5 Enables High Availability

The new version of Ontotext Platform simplifies deployment processes and infrastructure, and enables engineering teams to achieve high availability.

Sofia/Bulgaria Wednesday, June 23, 2021

Ontotext Platform 3.5 provides high availability by eliminating the single point of failure, lowers the complexity by removing rarely used components, updates key components to newer versions, introduces many improvements in deployment and operations, provides additional mutation capabilities to write in named graphs, and enables the Semantic Search Service authentication and the skipping of malformed data indexing.

High Availability

Business continuity is the capability of an organization to withstand disruption and to operate critical business applications and data normally and without interruption in accordance with predefined service-level agreements. To achieve that, a solution should provide protection from unplanned outages caused by hardware failures, human error, software problems and environmental issues.

In order to achieve High Availability with Ontotext Platform version 3.5, we ensure service-critical data integrity by providing multi-master GraphDB cluster support, FusionAuth HA support and multi-node deployment of Elasticsearch. This will create redundancy within the clusters with multiple data storage and they can be redundantly linked across multiple geographical locations.

The changes eliminate the single point of failure, reduce downtime, increase SLA (to nearly 100%) and ensure continuous operations, high-level performance and secure data. The high availability protection will keep the systems up and running, and will minimize the risk of lost revenue, unproductive employees and unhappy customers.

Additionally, in order to lower the Platform complexity and improve the high availability with the new version, we removed the Elasticsearch component from FusionAuth and enabled the use of GraphDB as default schema store. The previously available schema store (MongoDB) is now optional and will be deprecated with some of the next releases. This change also allows us to reduce the number of Platform components and lower the time for passing security and vulnerability checks in big enterprises.

Deployment and Operations Improvements

The other focus in the new release is to improve the deployment and operation of Ontotext Platform. Together with the replacement of MongoDB as schema store, we also made significant improvements in the Helm charts, health checks, logging, etc.

Some of the most important changes are:

  • Upgrades to FusionAuth 1.27 and GraphDB 9.8;
  • Helm charts changes and improvements:
    • Switched the chart components to a supported K8S version (v1.19);
    • Switched to Bitnami PostgreSQL HA Helm chart;
    • Switched to the official Ontotext GraphDB Helm chart;
    • Added options to provide existing PV claims or to enable dynamic provisioning;
    • And many more.

To upgrade to the newer version, just follow the Migration Guide.

Semantic Objects Service Improvements

In this new version, we’ve included a capability to declare a named graph during mutation. The mutated data will be stored in this named graph, which allows the information to be classified depending on the business needs. For example, information from different datasets can be saved in different named graphs and can later be easily accessed/manipulated based on where it originated.

In addition to the supported mutation modes, we’ve introduced a new one, “CHANGES”. Using this mode, all changes during mutation will be written in per-entity graphs.

Semantic Search Service Improvements

In the previous version, we introduced the new Semantic Search Service. Now we have extended the Platform authentication to cover the new service. When the Platform is started with security enabled, the user will have to provide proper credentials in order to access the Search Service. This will ensure that no one can access the data without valid authentication.

In our work with clients, we’ve often had to handle malformed data. This led us to introduce a configuration that ignores malformed data during indexing in Elasticsearch. If set to false, invalid data will prevent the schema from being successfully activated.

Get ready to take advantage of all new Ontotext Platform 3.5 features!

New call-to-action

 

For more information, contact Doug Kimball, Chief Marketing Officer at Ontotext