---
title: "Looking inside Uber’s GPU-powered open source, real-time analytics engine"
publication: "DB Weekly"
issue: 239
published_at: "2019-02-01"
canonical_url: "https://dbweekly.com/issues/239"
---

This week's database news       

#239 — February 1, 2019 

[Read on the Web](https://dbweekly.com/issues/239)   

Database Weekly 

[![\](https://res.cloudinary.com/cpress/image/upload/w_1280,e_sharpen:60/vhdp5hvij9fid43z4pjh.jpg)](https://eng.uber.com/aresdb/)  

[Introducing AresDB: Uber’s GPU-Powered Open Source, Real-time Analytics Engine](https://eng.uber.com/aresdb/) — Uber, the ride-sharing service, is _hugely_ data driven and they’ve built a significant system for doing large-scale analysis in real-time using GPUs for massive parallelization. This is a great tour of how it works, and [here’s the code](https://github.com/uber/aresdb). 

Uber Engineering   

[NoSQL for the Serverless Age: Google's Cloud Firestore Now Generally Available](https://cloud.google.com/blog/products/databases/announcing-cloud-firestore-general-availability-and-updates) — The Cloud Firestore database is a NoSQL, serverless solution for mobile and web application development, and is now generally available globally. 

Google Cloud   [![\](https://copm.s3.amazonaws.com/66182737.jpg)](https://www.strongdm.com/pgw-130/?utm_source=&utm_medium=email&utm_campaign=2019-1-30--%5BD%5D-%5BSchD%5D-SDM-PGW%3DTIP3-JAN-FORM-FILL) 

[Tired of Manually Revoking DB Creds & SSH Keys?](https://www.strongdm.com/pgw-130/?utm_source=&utm_medium=email&utm_campaign=2019-1-30--%5BD%5D-%5BSchD%5D-SDM-PGW%3DTIP3-JAN-FORM-FILL) — "strongDM simplifies access management for our teams. It delivers the layer of security we've been looking for." - VP Engineering, Hearst MediaOS | Click here to see strongDM in action and schedule a demo. 

strongDM sponsor   

[TimescaleDB 1.2: A Time Series Data Extension for PostgreSQL](https://blog.timescale.com/timescaledb-1-2-analytical-functions-advanced-data-lifecycle-management-improved-performance/) — A big week for TimescaleDB as not only have their open source and community releases got a few key enhancements (such as Postgres 11 support and [new time-series analytics functions](https://blog.timescale.com/sql-functions-for-time-series-analysis/)) but they’ve [raised $15m](https://blog.timescale.com/raises-another-15m-for-leading-time-series-sql-database-introduces-enterprise-offering/) and unveiled an enterprise offering too. 

Timescale   

[Microsoft Accidentally Deletes (Some) Databases](https://www.theregister.co.uk/2019/01/30/azure_sql_delete/) — The Register reports that an Azure outage earlier this week resulted in some customer databases being deleted but, thankfully, they were restored within 5 minutes. 

The Register   

[Operational Factors to Consider During MongoDB Data Modeling](https://severalnines.com/blog/operational-factors-consider-during-mongodb-data-modeling) — Describes the major factors to consider when performing MongoDB data modeling including scalability, atomicity, performance / data usage, sharding, indexing, storage optimization, document structure, growth, and the data lifecycle. 

Onyancha Brian Henry   

[The Future of Data: A Decentralized Graph Database?](https://hackernoon.com/the-future-of-data-a-decentralized-graph-database-bbb668715bd1) — A paradigm shift is happening that will change the way companies store, compute and transmit data. 

Matthieu Gavaudan   

[Lessons Learned Scaling Postgres to 1.2 Billion Records per Month](https://medium.com/@gajus/lessons-learned-scaling-postgresql-database-to-1-2bn-records-month-edc5449b3067) — The experiences of engineers building a cinema tracking service covering 3200 cinemas and 47,000 showtimes a day. They cover how they reviewed providers, experiences with materialized views, and using the database as a job queue. 

Gajus Kuizinas   

[The Joy and Pain of Using Google BigTable](https://syslog.ravelin.com/the-joy-and-pain-of-using-google-bigtable-4210604c75be) 

Jono MacDougall   

[Data Management Falls Short for Unstructured Data](https://www.datanami.com/2019/01/31/data-management-falls-short-for-unstructured-data/) — While most data-centric companies have solid strategies in place for structured data, current tools are proving insufficient for a wave of _unstructured_ data growth. This topic is coming up a _lot_ lately and I suspect there’s a lot of opportunity available. 

Allison Armstrong   

[InfluxDays: The Only Event Focused on Time Series Data](https://influxdays.com/new-york-2019/) — Attend InfluxDays NYC on 3/13 and learn why TSDB’s are the fastest growing database segment. 20% off code: INFLUXDATANEWS. 

InfluxData sponsor   

[TrailDB: Efficiently Store and Query Series of Events](http://traildb.io/) — A library (implemented in C but with bindings for R, Go, Python and others) for querying series of events at ‘blazing’ speed. 

AdRoll Inc.   

[The Life of a SQL Query](https://numeracy.co/blog/life-of-a-sql-query) — What happens when you run a SQL statement? We follow a Postgres query transformation by transformation as a query is processed and results are returned. 

Eric Feng  

🐦 Spotted on Twitter  

[![\](https://res.cloudinary.com/cpress/image/upload/w_1280,e_sharpen:60/hkahffambllbnuuf71nh.jpg)](https://twitter.com/shl/status/1090742936625528832)        

A **[Cooperpress](https://cooperpress.com/)** publication. 

[Change your email address](https://dbweekly.com/edit_subscription/*|UID|*) or[Stop getting this newsletter](https://dbweekly.com/leave/*|UID|*) 

© Cooper Press Ltd · Fairfield Enterprise Centre, Louth, LN11 0LS, United Kingdom