Nanolat Message Queue - A message in 3 us.

Nanolat Database

About Nanolat Database
  • An in-memory key/value store supporting put/get/del, and cursors for range search.
  • Supports transaction and master-slave replication.
  • Supports Table Plugins to implement your own data structure for efficient data processing for your specific domain.
Target Systems
  • Financial systems such as HFT(High Frequency Trading), FEP(Front End Processor), etc.
  • Monitoring systems that require processing huge amount of data in a second.
Why use it?
  • Low latency data processing, 1 transaction takes < 5 us.
  • Supports 1.5M transactions / second .
  • Supports built in Table Plug-ins using Google LevelDB and TokyoCabinet providing replication feature on top of them.
  • (Under Development in another project) Supports basic SQL that is directly converted to Nanolat Database API at compile time to achieve low latency data processing.
How it works?
  • Uses 40G Infiniband and 10G Ethernet instead of TCP/IP to replicate data from a server to another server within 2.5 us.
  • Instead of flushing transactional log on local disk, transfer it to slave machine.
Supported Platforms
  • RedHat EL5 up6, RedHat EL5 up7, RedHat EL6, EL 6.1 EL 6.2 EL 6.3
  • SLES10 SP4, 11, 11 SP1
  • Scientific Linux 6.1 
  • Other Linux distributions with kernel version 2.6 - 3.1
Source Code

Nanolat Database is an open source software. The source code is available at .

How to get user manual?

Send an email to to get the user manual of Nanolat Database. In case you do not want to wait for the email, you can reach us by calling 1-415-272-0796.