Can Go be a good language to implement NoSql database?

I would think so. It has many of the properties you would like in implementing a database system. First, the language is statically typed and executes surprisingly quickly. Second, it can handle high concurrency which is needed in a database since you will have many clients connection. Third, the libraries letting you interface with protobufs and HTTP provides two pretty simple ways to connect your Database interface to the rest of the world.

The perhaps most glaring problem, compared to the world of Erlang where I come from, is the lack of built-in seamless distribution. Many NoSQL databases trades consistency for availability and more machines. You are on your own implementing this.

Also, the lack of a ZooKeeper-like library may be something you will end up implementing. That, or perhaps the basis of the Dynamo paper.

> I would think so. It has many of the properties you would like in implementing a database system. First, the language is statically typed and executes surprisingly quickly. Second, it can handle high concurrency which is needed in a database since you will have many clients connection. Third, the libraries letting you interface with protobufs and HTTP provides two pretty simple ways to connect your Database interface to the rest of the world.

But that is not unique to Go, it may as well apply to Java, Scala, etc...

ZooKeeper in Go: https://github.com/ha/doozerd

LevelDB in Go: http://code.google.com/p/leveldb-go/

It provides in-memory and file-backed implementations.

Thank you, lots of useful information! I am currently in the process of switching to programming in dynamic languages (node.js, ruby) so I am also looking for good screencasts how programmers do work in those languages (using text editors, navigating through code, refactoring).