A NoSql database designed for maximum plugablitily and configurability.
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
src
.gitignore Add .gitignore Jan 7, 2015
.travis.yml Rename .travis to .travis.yml Dec 23, 2015
LICENSE
README.md Update README.md Dec 24, 2015
hexagons.ppt hexagons.ppt Jan 16, 2016
pom.xml RPC yea you know me Jan 4, 2016

README.md

nibiru Build status

Nibiru is a NoSql database designed for maximum plugablitily and configurability.

Motivation

There has been a recent proliferation in NoSql datastores. Many systems evolved from specific use cases or a specific industry. Often these systems are pitted against each other in apples to orange comparisons. Because apples and oranges can not be compared easily the debates typically degrade into a debate over what is the "best" implementation for the "real" users.

Nibiru takes a different approach by building API's throughout the codebase. There is no "locked in" implementation that is the "best" rather there is a system in place to provide alternative implementations. For example, Nibiru supports a gossiper for cluster discovery and a static provider. However this implemenentation is plugable so that an interested user can build a Zookeeper based implementation, an etcd based implementation, or whatever.

Plugability and is the key focus throughout the codebase. Nibiru is not even built around a specific data model. We already have support for a Column Family data model and a Key Value data model! Again Nibiru does not aim to tell you "apples are the best" or "we only support oranges". A user should be able to plug together the pieces they desire into the system they want! For example if a user wants a fast, in memory, non durable, key value store for one table, and a strongly consistent ColumnFamily store in another table they should have be able to have that.

Quickstart

If you want to launch an instance of nibiru we have configuration files that will create a two node cluster. Run the following commands in two separate shells. You will find the server on port 7070 on 127.0.0.1 and 127.0.0.2.

mvn exec:java -Dexec.mainClass="io.teknek.nibiru.Server" -Dexec.args="src/test/resources/nibiru_1.xml"
mvn exec:java -Dexec.mainClass="io.teknek.nibiru.Server" -Dexec.args="src/test/resources/nibiru_2.xml"

There is a simple CLI that wraps a facade around many of the client functions in the database.

mvn exec:java -Dexec.mainClass="io.teknek.nibiru.cli.Cli"  

Welcome to a very minimal Cli type 'connect <host> <port>' to get started 
ok> 
connect localhost 7070
ok>