Most developers tend to use a general purpose SQL databases for all kinds of data storage, which is okay most of the times but can add great overhead if you have a massive amount of data and when you want to minimize the deployment overhead. Different data storage databases exist to solve different programming problems, as developer, you should focus on the most efficient and scalable solution that fits your programming problem without taking the SQL database as a granted/generic solution.
I’ve started using Tokyo cabinet 2 years ago in one of the most sophisticated dispersed storage solutions I’ve ever seen/worked-on. I was astonished by the capabilities and performance of such lightweight database, developed by a Japanese engineer called Mikio Hirabayashi who also wrote QDBM previously. A couple of years later he and his brother (apparently) – Hatsuki Hirabayashi – Launched Kyoto cabinet, which is better than tokyo cabinet in almost every aspect. Those two brothers formed a startup called FAL Labs.
Koyoto cabinet is a lightweight concurrent key-value database that has some amazing performance characterstics, it can store one million records in 900ms (0.9s) in a hash database (HDB) and 1100ms (1.1s) in a B+ tree (BDB) database with minimal storage overhead of 4 bytes for BDB and 16 bytes for HDB! Continue reading →












Recent Comments