Introduction to Tachyon and a deep dive into Baidu’s production use case

I pleased to announce a webcast that I’ll be hosting a webcast featuring the co-creator of Tachyon (full disclosure: I’m an advisor to Tachyon Nexus) alongside one of the architects behind Baidu’s big data platform. I hope to see you online on Sept 14th!

Tachyon is a memory-centric fault-tolerant distributed storage system, which enables reliable data sharing at memory-speed. It was born in the UC Berkeley AMPLab and is completely open source. Multiple companies deploy Tachyon, for example, Baidu is running a production Tachyon cluster with 150 nodes managing over 2 PB of storage space. Tachyon has more than 100 contributors from over 30 institutions, including Baidu, IBM, Intel, and Yahoo.

In this webcast, Haoyuan Li from Tachyon Nexus will present an overview of Tachyon, as well as some recent development and use cases. After that, Shaoshan Liu from Baidu will present their experience with Tachyon. He will describe how they achieved 30x end-to-end performance improvement using Tachyon, how they addressed problems encountered when they started using Tachyon, what new features they want to see, and future plans to scale further.

Leave a Reply

%d bloggers like this: