What is Cassandra?
What is PostgreSQL?
Want advice about which of these to choose?Ask the StackShare community!
We use postgresql for the merge between sql/nosql. A lot of our data is unstructured JSON, or JSON that is currently in flux due to some MVP/interation processes that are going on. PostgreSQL gives the capability to do this.
At the moment PostgreSQL on amazon is only at 9.5 which is one minor version down from support for document fragment updates which is something that we are waiting for. However, that may be some ways away.
Other than that, we are using PostgreSQL as our main SQL store as a replacement for all the MSSQL databases that we have. Not only does it have great support through RDS (small ops team), but it also has some great ways for us to migrate off RDS to managed EC2 instances down the line if we need to.
PostgreSQL combines the best aspects of traditional SQL databases such as reliability, consistent performance, transactions, querying power, etc. with the flexibility of schemaless noSQL systems that are all the rage these days. Through the powerful JSON column types and indexes, you can now have your cake and eat it too! PostgreSQL may seem a bit arcane and old fashioned at first, but the developers have clearly shown that they understand databases and the storage trends better than almost anyone else. It definitely deserves to be part of everyone's toolbox; when you find yourself needing rock solid performance, operational simplicity and reliability, reach for PostgresQL.
Stitch is a wrapper around a Cassandra database. It has a web application that provides read-access to the counts through an HTTP API. The counts are written to Cassandra in two distinct ways, and it's possible to use either or both of them:
Real-time: For real-time updates, Stitch has a processor application that handles a stream of events coming from a broker and increments the appropriate counts in Cassandra.
Batch: The batch part is a MapReduce job running on Hadoop that reads event logs, calculates the overall totals, and bulk loads this into Cassandra.
Relational data stores solve a lot of problems reasonably well. Postgres has some data types that are really handy such as spatial, json, and a plethora of useful dates and integers. It has good availability of indexing solutions, and is well-supported for both custom modifications as well as hosting options (I like Amazon's Postgres for RDS). I use HoneySQL for Clojure as a composable AST that translates reliably to SQL. I typically use JDBC on Clojure, usually via org.clojure/java.jdbc.
PostgreSQL is responsible for nearly all data storage, validation and integrity. We leverage constraints, functions and custom extensions to ensure we have only one source of truth for our data access rules and that those rules live as close to the data as possible. Call us crazy, but ORMs only lead to ruin and despair.
Tried MongoDB - early euphoria - later dread. Tried MySQL - not bad at all. Found PostgreSQL - will never go back. So much support for this it should be your first choice. Simple local (free) installation, and one-click setup in Heroku - lots of options in terms of pricing/performance combinations.
Cassandra is our data management workhorse. It handles all our key-value services, supports time-series data storage and retrieval, securely stores all our audit trails, and backs our Datomic database.
While we experimented with Cassandra in the past, we are no longer using it. It is, however, open for consideration in future projects.
We are using Cassandra in a few of our apps. One of them is as a count service application to track the number of shares, clicks.. etc