I use the term Enterprise NoSQL in NoSQL For Dummies to distinguish two types of systems:-
- (Open Source) NoSQL systems
- Science Projects – ‘Hey I’ve built an internal wiki with this NoSQL database!’
- Secondary stores – Where a NoSQL database is a read only cache for a different (normally relational) database
- Enterprise NoSQL systems
- Mission critical systems where the NoSQL database is the PRIMARY store for data
- Where a down day is a death day for the business – healthcare, defence, financial services trade stores
- Requires High Availability data replication between nodes in the same cluster
- Requires scaling up / down the size of the cluster based on current workload, including auto data rebalancing (aka re-sharding)
- Requires Disaster Recovery replication to 1 or more alternate data centres
- Requires backup, restore, systems management and live systems monitoring
- Requires secure storage of information, with security at or below the record (row, document, graph) level managed inside the database
I personally only care about Enterprise NoSQL Systems. I have government clients where all of the requirements of Enterprise software are true.
Enterprise requirements mean that the old adage of ‘Hey we can use an open source NoSQL database for this!’ IS A MYTH.
There is not a single NoSQL database whose Enterprise version is free AND provides all of the Enterprise features listed above.
This means Total Cost of Ownership cannot be decided on NoSQL database license cost alone. In particular, many open source vendors have high day rates on consultancy. If you think about it they have to – because they don’t make any money from the open source variant of their software!
If you have a mission-critical system that needs a NoSQL database, evaluate the Commercial only vendors AND the Enterprise version only of the open source vendors.
Blog posts on Enterprise NoSQL
You can read all my blog posts on Enterprise NoSQL. Try starting with these:-
- Why you need Enterprise NoSQL
- A NoSQL database for the Enterprise
- An independent review of NoSQL and MarkLogic