Swedish startup Logical Clocks takes a crack at scaling MySQL backend for live recommendations

Takes a 'different approach' to YouTube's Vitess to munch complex transactions in microseconds

Swedish startup Logical Clocks is launching a new key-value database as a managed service, based on the MySQL derivative MySQL NDB Cluster.

The vendor told us its RonDB can be used to provide live data to machine learning models for real-time decision-making – as commonly used in online recommendations and fraud detection.

Although it has a history going back to the late 1990s, the new open-source distribution is currently in closed beta, with interested users encouraged to apply to participate. General availability is expected in the second quarter.

Logical Clocks said the database can respond in 100-200 microseconds on individual requests, in less than a millisecond on batched read requests and perform complex transactions in a highly loaded cluster within 10 milliseconds. It can perform hundreds of millions of read or write operations per second, the company added, and apparently offers 99.9999 per cent availability – no more than 30 seconds of downtime per year.

Individual RonDB data nodes can store up to 16TB of in-memory data and many tens of terabytes of disk data. Clusters can scale up to 144 data nodes offering petabytes of data, the vendor said.

RonDB is an open-source distribution of MySQL NDB Cluster, the MySQL back-end used as a high-performance alternative to the InnoDB storage engine.

Roots in the '90s

Its creator, Logical Clocks head of data Mikael Ronström, told The Register he began work on the database in the 1990s as part of his PhD thesis, which Swedish telco Ericsson helped to fund at Linköping University.

As the resulting database was bought by MySQL, the product was named MySQL NDB Cluster, which Ronström said had seen success in the telecoms industry, as well as in gaming and financial applications.

Oracle bought MySQL, the company supporting the open-source relational database, when it acquired Sun Microsystems for $7.4bn in 2009 (Sun having bought MySQL a year earlier).

Although MySQL NDB Cluster was proving popular, Logical Clocks said it wanted to make it easier to deploy and maintain to accelerate its adoption.

"It's been very successful but mostly with customers that look for the really high performance, the ones that have enough time to spend on actually integrating the database, configuring it and trying it out. We're trying to make it a lot easier to use and get rid of the barriers to adoption," Ronström said.

Logical Clocks said a global European online media company was beta-testing the database as an alternative to the wide-column store database Cassandra to provide the data for machine learning recommendations, involving 4 million lookups per second.

Andy Pavlo, associate professor of databaseology in the computer science department at Carnegie Mellon University, said RonDB was trying to tackle the scaling problem in MySQL known about since the mid-2000s.

Google, Facebook, and others were building their own middleware until database startup Clustrix came out of AOL as a commercial product and Vitess (and the hosted PlanetScale) came out of YouTube, where the former served all of the streaming platform's database traffic from 2011 to 2019.

"RonDB predates all of these approaches. It takes a different approach from that [of] Vitess," Pavlo added.

While Vitess, which was donated to the Cloud Native Computing Foundation in February 2018, intercepts the queries using MySQLProxy and then rewrites them to standalone MySQL instances running InnoDB to execute the query, RonDB and MySQL NDB Cluster use a different storage engine than InnoDB. It is responsible for sending low-level physical commands to the cluster, the associate professor said.

"These are just different ways to make a single-node DBMS like MySQL become a distributed system. There are pros and cons to each of them that are probably not relevant to most users," he added. ®