History[06][07]

Originally released commercially in 2002 under GemStone Systems as Gemfire, it was first used in Wall Street trading platforms. In 2010, GemStone Systems was acquired by VMWare and by 2013, Pivotal had spun out of VMWare and took Gemfire with it. Pivotal then proceeded to submit Gemfire to the Apache Incubator in 2015 under the name Geode, and was graduated to a top-level project in 2016. Now, Geode is used in many businesses with high-scale applications.

Concurrency Control[05]

Any modifications to caches are replicated and checked for consistency before acknowledging success. If there are any concurrent modifications, then changes are made according to time stamp.

Data Model[08]

Foreign Keys[09]

Indexes[05]

Keys are hashed using fixed number of buckets, with the default number of buckets being 113.

Isolation Levels[10]

Geode supports repeatable reads, but its default allows for dirty reads.

Logging[11]

Apache Log4j 2

Query Interface[12]

Geode uses its own type of query language, OQL, which stands for Object Query Language. The main differences between SQL and OQL are: SQL has aggregate functions, while OQL allows querying object attributes, object graphs, and also using object methods.

Storage Architecture[05]

Geode is a data grid, meaning it stores caches of the database, and does so in- memory. This is what allows Geode to be scalable and to have low latency.

System Architecture[01]

Views[13]