Elastic 8.0: opening a new era of speed, scale, relevance and simplicity

Reprint ：Elastic 8.0： Opening speed 、 scale 、 A new era of relevance and simplicity

We are pleased to announce Elastic 8.0 Official release of , With Elasticsearch Enhancement of vector retrieval function of 、 Yes NLP Native support for 、 More and more simplified data access , And a streamlined security experience ,Elastic 8.0 Ushered in a speed 、 scale 、 A new era of relevance and simplicity .

Elastic Stack 8.0 introduction

Elastic Stack 8.0 introduction_ Bili, Bili _bilibili

Whether it's connecting people and teams to important content 、 Keep mission critical applications and infrastructure online , Or to protect the entire digital ecosystem from increasingly complex cyber threats ,Elastic 8.0 All laid the foundation for the next generation of search driven solutions , Help everyone reach a new level of success .

Are you ready to start ？Elastic 8.0 Now you can Elastic Cloud Upper use -- This is the only one that includes all the new features in the latest version Elasticsearch Managed products .

Speed 、 Scale and relevance ： A new beginning , But the first intention is the same

Every end means a new beginning . When we start the new year （2021 See you in ,2022 Hello in ）-- We are also starting to enter a speed 、 Scale and relevance Elastic 8.0 The new age .

Our customers and communities know ,Elastic Always adhere to speed 、 Commitment to scale and relevance . every last Elastic Both versions have enhancements and optimizations , In order to ensure that Elasticsearch Is the fastest 、 Most scalable 、 The most powerful search engine .

in fact , In the past three years , We have made great progress in the following areas ： Reduce memory usage （ Allow each node to manage more data ）, Reduce query costs （ Especially for large deployments ）, And introduce some new features to improve the relevance .

for example , With 7.x Release of version , We have improved the speed of date histogram and search aggregation 、 Enhanced the performance of page caching 、 And created a new “ Pre filtration ” Search phase . Besides , We reduce the memory stack , Fully support ARM framework , Introduce new ways to reduce storage , Enable our customers to easily separate computing from storage through new frozen tiers and searchable snapshots , This reduces the demand for resources .

For us , Yes Elastic Stack The best thing about infinite optimization is , No matter how you choose to use your data for work , These enhancements essentially help you search at speed and scale 、 Resolution and success —— No extra work is required .

Using native vector retrieval to improve the relevance of search

Elastic 8.0 It brings a set of native vector retrieval functions , This enables customers and their employees to search in their own language and words and obtain highly relevant results .

In the past two years , We have been trying to make Elasticsearch Become a good place for vector retrieval . As early as Elasticsearch 7.0 When it was released , We introduce field types for high-dimensional vectors . stay Elasticsearch 7.3 and Elasticsearch 7.4 in , We introduce support for vector similarity functions . These early versions demonstrated the introduction of vector retrieval techniques into Elasticsearch The future of the ecosystem . We are pleased to see that our customers and communities are eager to use them in a wide range of cases .

today , stay Elasticsearch 8.0 in , We treat natural language processing （NLP） The local support of the model is introduced directly Elasticsearch, Make vector search more practical . Besides ,Elasticsearch 8.0 It also includes right ANN Native support for search -- Make vector based query and vector based document library fast 、 Large scale comparisons become Probably .

use NLP The power of opens up a new world of analysis

Elasticsearch Always do NLP A good place , But before that it needed to be in Elasticsearch Do something else , Or write some rather complicated plug-ins . stay 8.0 In the version , Users will be able to go directly to Elasticsearch Named entity recognition in 、 Sentiment analysis 、 Text classification, etc , Without additional components or code . stay Elasticsearch Internally calculate and create vectors , Not only in terms of horizontal scalability " victory "（ By distributing computing in the server cluster ）-- This change is also called Elasticsearch Users save a lot of time and energy .

Elastic Stack 8.0 Introduce 2

Elastic Stack 8.0 Introduce 2_ Bili, Bili _bilibili

stay Elastic 8.0 in , Users can directly Elasticsearch Use in PyTorch Machine learning model （ Such as BERT）, And in Elasticsearch Use these models for reasoning . These models can be your own custom models , It can also be a model published to the community , Such as Hugging Face.

By letting users directly in Elasticsearch To make a reasoning in , It is easier than ever to bring Modernity NLP The power of is integrated into the search application and experience （ think about it ： No need to write code ！）, Essentially more efficient （ thank Elasticsearch Distributed computing power ）, and NLP Itself becomes faster , Because you don't need to move data to a separate process or system .

Quick search , Scale search

Whereas Elastic 8.0 Is based on Lucene 9.0 Of , Those who take advantage of modern NLP Search experience , Due to ANN Of （ new ） Native support , It can be more rapid and large-scale .ANN It makes a quick and effective comparison between vector based queries and vector based document corpora （ Whether it's small 、 Large or huge corpus ） Make it possible .

Elastic Stack 8.0 Introduce 3

Elastic Stack 8.0 Introduce 3_ Bili, Bili _bilibili

think about it , This is just the beginning ......

All in all , Through the modern NLP Local support for models and support for ANN Local support for search , bring Elastic 8.0 It releases the power of vector retrieval for our customers and communities .

Elastic Stack 8.0 Introduce 4

Elastic Stack 8.0 Introduce 4_ Bili, Bili _bilibili

future , Search driven solutions （ Such as Elastic Enterprise Search） It will make it very simple to use the powerful ability of vector retrieval . In this case , The user's search term does not need to match the best results exactly . Vector retrieval can " Connecting these points ", And easily emerge the most relevant results .

Simple things should be simplified

Through two new AWS Integration to simplify cloud native observation capabilities

In the past few versions , We simplify getting any type of data from any source to Elastic Stack The process of . With Elastic 8.0 Release , We have further expanded cloud native integration , There are two new ones AWS Integrate , namely ： A new AWS Lambda Application integration and a new Amazon S3 Storage Lens Integrate .

new AWS Lambda The application is published in AWS Serverless application library （SAR） in , Enables users to simplify their architecture and simplify the data ingestion process , You don't have to pay for configuring virtual machines or installing data transmitters . In short , Users now only need to AWS Click a few times in the console , You can change the log from Amazon S3 Ingest their Elastic Cloud Deploying .

adopt Elastic stay 12 First announced in January Amazon S3 Storage Lens Integrate , Users can easily put Storage Lens Indicators are delivered to their Elastic Cloud Deploy , Use Elasticsearch Powerful search and analysis capabilities and Kibana Powerful dashboard functions to optimize Amazon S3 Use cost , On the basis of ensuring the efficiency of data protection , Monitor user activity trends . in addition , Now you can go through AWS Marketplace register , Get started easily AWS Upper Elastic Cloud Free trial of .

Protect your data from unauthorized access with simplified stack security

Proper security has never been more important than it is today . Although the security features have been free for years and embedded in Elastic Stack in , But we still further simplify the steps required to configure security —— Because we believe , No cluster should run without protection . stay Elastic 8.0 in , For self managed clusters , Security is now enabled by default . This ensures that the data 、 The network and users are Elastic Stack The security , Prevent data leakage and unauthorized access . Through assistance within the product , Such as automatically generated token And certificates , We have streamlined and simplified this process , To save time and energy , Make it safe for everyone .

and , If you are already using （ Or prepare to try ）Elastic Cloud, don't worry , Security and full role-based access control are always enabled .