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
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.
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.
Get Started
Developers
Company
