No SQL Required
With a new age of application and infrastructure scaling comes new options for dealing with the ways that we handle data. One of our favorites in this arena is the Riak data store from Basho.
- Load Balancer
- Traffic Director
- Riak Node
- Memory: 2GB Storage: 5GB vCPUs: 2
- Management Node
- Memory: 246MB Storage: 200GB vCPUs: 1
Do you need a massively-scalable database? Are you drinking the NoSQL punch? Take a look at our starter Riak cluster!
- Private Riak cluster
- Initial capacity of ~7500000 objects
- Distributed and "eventually consistent"
- Built-in Map-Reduce engine
- Data storage for non-relational data
- Use as a distributed file system of sorts (similar to Amazon S3)
- Use as a document database for quasi-relational data
What is Riak?
Riak is a different sort of database system. Rather than emulating the relational model like the usual suspects (MySQL, PostgreSQL), it is primarily a distributed key-value store that was inspired by Amazon's Dynamo, but it also shares the qualities of document databases like MongoDB and CouchDB. Additionally, while the initial layout of our clusters support upwards of 7500000** objects, Riak's architecture makes it terribly easy to scale the cluster for either storage capacity or processing power.
(Based on an average bucket size of 15 bytes, average key size of 15 bytes, average value size of 2000 bytes, and three copies of the object written to the cluster.)