OpenSearch 3.0 is here and brings more performance and scalability

Almost 3 years after version 2.0, the OpenSearch project has released a new major version of the search and analysis suite at the OpenSearchCon Europe conference. This is accompanied by a series of improvements and some new functions. The starting signal for the new main version comes from Apache Lucene. The well-known search library is an important pillar of OpenSearch and has been available in version 10 for a few months now. The jump to this new version and to Java 24 alone improves the search speed, i.e. reduces the waiting time for results.

The network level does not remain untouched. OpenSearch is a distributed system. This means that data and information must flow back and forth. Thanks to the use of gRPC and protobuf, Opensearch has also become faster here. Thanks to the HTTP/2 basis, gRPC is capable of multiplexing and can process bidirectional data streams, i.e. send and receive simultaneously via the same TCP connection. In general, the OpenSearch project wants to focus more on microservices and cloud-native computing. This also includes the further development of the "in-house" Kubernetes operator.

OpenSearch is now also breaking new ground in the area of data storage. Until now, the company has tended to take a monolithic approach. The user feeds new information into the system as a complete document. OpenSearch breaks this down and stores derived information in the corresponding subsystems. This means that the data is stored redundantly, so to speak, in the form of the overall document and in the subsystems. OpenSearch no longer necessarily stores the former.

Why? The complete document can be restored from the derived information in the subsystems, and updates are also more efficient. Once the source document has been split up, it is always clear which data is new and which is not. This makes the entire data management process much more efficient. A detailed list of the new features is documented on the project website.

From fork to foundation

OpenSearch was launched in 2021 as a fork of Elasticsearch after it dropped its Apache license. In the fall of 2024, the Linux Foundation took the project under its wing. A broad ecosystem has now formed around OpenSearch, which was mainly based at AWS for a long time. Companies such as Eliatra from Ireland, for example, offer an OpenSearch distribution and corresponding support contracts, similar to other companies that sell Linux or Kubernetes as their own distribution.

In cooperation with IONOS, there are also GDPR-compliant Managed OpenSearch offerings. Under the name NeoNephos-Foundation, the Linux Foundation Europe founded a new organization in the spring that is dedicated to promoting open technologies and digital sovereignty in Europe.

Upcoming OpenSearch events include OpenSearchCon India from June 3-4 and OpenSearchCon North America from September 8-10. (mki)