The Magic of Precision Fetching with Tiered Storage in StarTree Cloud

In this lightboard, Neha Pawar and Tim Berglund concoct a compelling metaphor comparing the concept of Tiered Storage in StarTree Cloud to the wizarding world of Harry Potter. The eccentric new Minister of Magic (representing a cost-sensitive cloud solution) has issued a mandate that all wizards must lug around every single item they might need at all times. The paraphernalia is kept in weighty trunks, requiring Hippogriffs and elf buddies for transportation, mirroring the burdensome cost and performance issues related to scaling storage.

Neha and Tim propose a more efficient, loosely coupled system akin to a magical charm, which allows for local, network attached, and cloud-based storage, reducing costs. Apache Pinot fetches specific data “pages” based on a query to minimize latency, and employs advanced strategies like selective columnar fetch and automatic local pinning based on usage patterns. Their approach has led to significant improvements in query response times, effectively demonstrating the efficiency of the system.

LEARN MORE ABOUT STARTREE CLOUD

Contents
Share
MIT Technology Review

Transform customer experience with real-time analytics

This recent report explores how companies like Stripe and Uber are able to embed real-time analytics in their apps to deliver continuous insights, improve customer engagement, and open new revenue opportunities
Download your copy
Subscribe to get notifications of the latest news, events, and releases at StarTree