#225 - Accelerating Modern Databases with Speedb

In this episode, Chris and Martin discuss storage engine efficiency and performance with Speedb CEO, Adi Gelvan. Storage engines form the core of all databases, providing the translation between the structured form of data the application sees, and the way in which data is written onto physical media like disks or SSDs. The performance and resiliency of a database is directly affected by the performance of the storage engine. Speedb has developed technology that optimises the underlying engine for many applications – RocksDB. Optimisation is more than pure performance. Database scaling involves sharding – splitting a database using some characteristic of the data. The result of sharding is increased costs and complexity, especially in the public cloud. Adi explains how Speedb helps avoid sharding and reduces resource consumption.

You can find out more about speedb at https://www.speedb.io/

Elapsed Time: 00:38:41

Timeline

00:00:00 – Intros
00:01:00 – Why Speedb?
00:03:15 – What is a database storage engine?
00:05:00 – MySQL, MongoDB and other databases have pluggable storage engines
00:06:30 – What’s driving a need for better storage engines?
00:08:00 – Data is now trillions of files and metadata
00:11:00 – Existing storage engines are complex – like real engines
00:12:45 – Database performance has lots of “solutions” thrown at it
00:13:30 – RocksDB derived from LevelDB
00:15:15 – Sharding is a solution for scaling databases
00:19:30 – Speedb provides both fast read and fast write performance
00:21:15 – Cloud has made sharding easy – with an associated cost!
00:23:45 – Speedb maintains API compliance with RocksDB
00:26:40 – API compatibility is essential for plug and play components
00:32:15 – Is storage “done”?
00:35:00 – Speedb will initially work with OEMs and look at Open Source
00:38:00 – Wrap Up

Related Podcasts & Blogs

Podcast (storageunpacked): Play in new window | Download