AccelerateWithOptane / lab

Request access to Optane powered bare metal infrastructure for performance-testing and analysis purposes
15 stars 3 forks source link

Proposal: TiDB #5

Open ceohockey60 opened 6 years ago

ceohockey60 commented 6 years ago

Name, email, company, job title

Ed Huang, huang at pingcap.com, PingCAP, co-founder and CTO Kevin Xu, kevin at pingcap.com, PingCAP, General Manager of U.S.

Project Title and description

TiDB is an open-source distributed scalable Hybrid Transactional and Analytical Processing (HTAP) database, designed to provide infinite horizontal scalability, strong consistency, and high availability. It is MySQL compatible and serves as a one-stop data warehouse for both OLTP and OLAP workloads. It has three core components:

TiDB: a stateless SQL layer that’s MySQL compatible TiKV: a distributed transactional key-value storage layer TiSpark: a Spark plug-in that works on top of TiDB/TiKV

Which members of the community would benefit from your work?

We’re going to bench TiDB/TiKV with Optane in a large scale cloud cluster deployment, so the wider database community can better evaluate TiDB’s performance on Optane SSDs based on the results.

Is the code that you’re going to run 100% open source? If so, what is the URL or URLs where it is located?

Yes. TiDB: https://github.com/pingcap/tidb TiKV:https://github.com/pingcap/tikv TiSpark: https://github.com/pingcap/tispark

What infrastructure (computing resources and network access) do you need? (see: https://www.packet.net/bare-metal/)?

We would like to request access for 10 to 20 Optane SSDs that are either c1.xlarge.x86 or m2.xlarge.x86.

What performance-focused articles has your project published?

We’ve published a performance tuning article on InfoWorld in February 2018: https://www.infoworld.com/article/3258810/sql/tidb-performance-tuning-a-distributed-newsql-database.html

We also include performance results in relevant product announcement like our 2.0 blog post: https://www.pingcap.com/blog/tidb-2-0-announcement/

Please state your contributions to the open source community and any other relevant initiatives

All core components of TiDB (TiDB, TiKV, TiSpark) have been open-sourced from day one and are currently deployed in production in more than 200 companies. Our team actively contributes to RocksDB (which we use as storage engine). We are also maintainers of open-source projects like rust-prometheus, grpc-rs, and raft-rs, which are used in other open-source distributed systems.

Would you be willing to share your analysis and results publicly?

Yes.