Apache GEODE - A Wonderful Low Latency & High Consistency In-Memory Data Store
Use Cases and Deployment Scope
We are trying to use it as an experimental alternative to redis as a low latency, in-memory server to save copies of static dataloads (configuration data, data that does not change frequently etc) that is saved in our Database (Mysql). Everytime when a request is made for the above mentioned data, instead of loading the data from DB we pull it from Apache Geode
Pros
- Super Fast data pull/push
- Provided ACID transactions, so it works like a SQL Database
- Provides replication & partitioning, so our data is never lost and extraction is super fast. NoSql like properties
Cons
- Needs more supporting languages. Out of box Python, Nodejs adapters would be wonderful
- Currently it supports just KV Store. But if we could cache documents or timeseries data would be great
- Needs more community support, documentation.
Return on Investment
- Still experimenting so difficult to quote
- For a small size project/teams might be an overkill as it still has certain learning curve
- For Medium to large projects with complex Data Structures that need to be queried with a fast o/p it definitely works
Alternatives Considered
Redis™*, Aerospike and Amazon ElastiCache
Other Software Used
Redis™*, Aerospike, MySQL, Elasticsearch, Apache Kafka, Amazon Kinesis, Amazon Aurora
