It has been more than a month since I had a post on this blog. I was trying to do at least some part of the ‘WeStand!’ features and come back, however things were not in favor of my plan. I am still working on it and other parallel things. For today I would like to shortly describe what the presentation I had at the GTUG Addis Event last Saturday. I had a presentation on NoSQL databases. The following is a short summary of the presentation:
1. The motivation of NoSQL databases
The discussion was on what drives the NoSQL field and the naming of the ‘NoSQL’. It is a little confusing that ‘NoSQL’ might mean no query. But it is to mean that No RDBMS (Relation Database Management System). Out of the major characteristics of RDBMS (Convenience, multi-user support, Safety, Persistent files, Reliability, Massive scaling and efficiency), NoSQL trades-off some of the features to gain another boost on other features like massive scaling and efficiency.
Pros and cons of No SQL are flexible schema, quick setup (which is a controversial idea), massive scale in the first group and More programming, fewer guarantee in the second group respectively.
2. The second part of the discussion was examples which really require the use of NoSql databases for efficiency and showing that the RDBMS is really abused.
3. The type of NoSQL data stores
There are four types of classes of NoSql
i)MapReduce Framework (optimized for OLAP)
examples include Hadoop (an open source from Google)
ii)Key value stores (for OLTP)
examples include Googles BigTable, Amazons Dynamo …
iii) Document Store
examples include MongoDB, CouchDB and RavenDB (which was the main discussion point)
iV) Graph database
examples include Neo4j, FlockDB
We had other related interesting discussion. Thank You GTUGAddis for inviting me to present. It was a good learning atmosphere.