Open Distro for Elasticsearch 1.8.0 is released, supports Cosine Similarity in k-NNby: Viraj Phanse, Pavani Baddepudi, Alolita Sharma · on:
We are pleased to announce the release for Open Distro for Elasticsearch 1.8.0. that now supports cosine similarity distance metric with k-NN. We released k-NN similarity search feature in Open Distro for Elasticsearch 1.4.0 with support for Euclidean distance to calculate similarity between the vectors. With cosine similarity, you can now measure the orientation between two vectors while Euclidean distance is used to measure the magnitude. We would also like to thank our community for contributing snapshot support in Index Management. This feature enables users to define snapshot action in Index State Management for retention and to migrate indices from one cluster to another. Open Distro for Elasticsearch 1.8.0 can be downloaded here.
The release consists of Apache 2.0 licensed Elasticsearch version 7.7.0, and Kibana version 7.7.0. This distribution also includes alerting, anomaly detection, index management, performance analyzer, security, SQL and k-NN plugins. Other components including SQL Workbench, SQL ODBC and JDBC drivers, SQL CLI client, and PerfTop, a client for Performance Analyzer are also available for download.
Download the latest packages
If you’re using RPMs or DEBs, see our documentation on how to install Open Distro for Elasticsearch with RPMs and install Open Distro for Elasticsearch with Debian packages. A tarball is also available for testing and other applications.
A Windows ready package supporting version 1.8.0 enables users to easily install Elasticsearch and Kibana on Windows. If you’re using Kubernetes, check out the Helm chart to install Open Distro for Elasticsearch.
You can find Open Distro for Elasticsearch security, alerting notification and job scheduler artifacts on Maven Central.
You can download the latest versions of Open Distro for Elasticsearch’s PerfTop client on npm.org, Open Distro for Elasticsearch’s latest SQL CLI client on PyPi. SQL drivers supporting ODBC and JDBC are also available.
- New feature Cosine Similarity is available for use in k-NN plugin.
- The snapshot feature is now available in the Index Management plugin.
- Anomaly Detection plugin releases the new count aggregation feature to detect anomalies.
- Support for connecting PerfTop CLI, a client to interact with Performance Analyzer, to clusters with basic authentication.
Come join our community!
There are many easy ways to participate in the Open Distro community!
Ask questions and share your knowledge with other community members on the Open Distro discussion forums.
Attend our bi-weekly online community meetup to learn more about Elasticsearch, security, performance, machine learning and more.
File an issue, request an enhancement you need or suggest a plugin you need at github.com/opendistro-for-elasticsearch.
Contribute code, tests, documentation and even release packages at github.com/opendistro-for-elasticsearch. If you want to showcase how you’re using Open Distro, write a blog post for opendistro.github.io/blog. If you’re interested, please reach out to us on Twitter. You can find us at Viraj at @vrphanse or Alolita at @alolita or Jon at @jon_handler.
We also invite you to get involved in ongoing development of new Open Distro for Elasticsearch plugins, clients, drivers. Contribute code, documentation or tests for Performance Analyzer RCA Engine.
You can also track upcoming features in Open Distro for Elasticsearch by watching the code repositories or checking the project website.
Thanks again for using and contributing to Open Distro for Elasticsearch and being part of the project’s growing community!
About the Authors
Viraj Phanse is a Senior Product Manager at Amazon Web Services for Search Services. You can find him on GitHub or Twitter @vrphanse.
Pavani Baddepudi is a Senior Product Manager working in Search Services at Amazon Web Services.
Alolita Sharma is a Principal Technologist at AWS focused on all things open source including Open Distro for Elasticsearch. You can find her on GitHub or Twitter @alolita.