Elasticsearch Not Working Out?

Try StarTree Cloud

Elasticsearch is a great search engine, but it wasn’t designed for real-time, user-facing analytics. If your workload has these characteristics, here are good reasons to consider migrating your data to StarTree Cloud, powered by Apache Pinot™.

Low-Latency Aggregations at Scale

 

StarTree Cloud provides performant aggregations against petabytes of data, with latencies measured in milliseconds; Elasticsearch hits the wall with high-dimensionality aggregations

Fast & Flexible Indexing

 

StarTree Cloud offers multiple indexing options, including the star-tree index, for fast and efficient query results; outperforms Elasticsearch’s inverted index

Real-Time Upserts

 

StarTree Cloud supports real-time upserts — a vital feature if your data is consistently being updated from event streaming or CDC; Elasticsearch does not support real-time upserts

User-Facing Analytics

 

StarTree Cloud supports extremely high concurrency queries (100,000+ QPS), unlocking your data for all internal & external users; Elasticsearch can’t maintain QPS at the same rate

Resource Efficiency to Lower Infrastructure Spend

 

StarTree is more efficient with memory & storage, saving significantly on your infrastructure spend; Elasticsearch doesn’t use efficient column storage and is memory intensive

Designed for Purpose

 

StarTree Cloud was built for fast, real-time user-facing online analytics processing (OLAP). Elasticsearch was designed as a full-text query search engine. Use the right tool for the right job.

StarTree Cloud, Powered by Apache Pinot™

StarTree Cloud is our Database-as-a-Service (DBaaS), powered by Apache Pinot™. StarTree Cloud is faster than Apache Druid and more scalable than ClickHouse. It also supports highly concurrent queries (high QPS) better than other real-time analytics databases — a capability that we refer to as user-facing analytics.

StarTree Cloud also includes StarTree Data Manager, which makes for easy no-code ingestion of data from event streaming systems like Apache Kafka®, live Change Data Capture (CDC) from transactional databases, as well as batch-oriented systems, object stores and a wide variety of data formats.

Why Elasticsearch Users Switched to Apache Pinot

— Yupeng Fu

We use Apache Pinot as a core system to empower mission-critical use cases

— Yupeng Fu, Uber



Uber fully replaced Elasticsearch with Apache Pinot for its time-sensitive real-time analytics. This allowed them to handle increasing traffic growth and numbers of users, support a broader variety of use cases, all while saving on operation costs and optimizing for feature development. With Apache Pinot, Uber can do real-time upserts and get query results from 1.5PB of data with <100ms P99 latencies. Beyond performance improvemets, it also provided improved reliability and scalability, while reducing engineering cost.

— Vaibhav Mittal

Once we saw the raw numbers we decided Elasticsearch should no longer be considered for our further analyses.

— Vaibhav Mittal, Cisco WebEx



Cisco WebEx moved from Elasticsearch to Apache Pinot after seeing how it outperformed their existing infrastructure. Apache Pinot brought query times of 10-30 seconds in Elasticsearch down to sub-second speeds. Apache Pinot was also found to be 4× faster than Clickhouse in most cases in Cisco WebEx’ head-to-head benchmarking comparison.

Get Started Today

With StarTree Cloud you can set up data ingestion in minutes. There’s no time like the present to sign up for a free trial account. Data migrations, however, can    be a bit tricky. So if you want to talk to our team about how to migrate your data from Elasticsearch to StarTree Cloud, sign up now for a free consultation.

SIGN UP FOR FREECONSULT WITH OUR EXPERTS