Apache Cassandra

Benefits:

Easy to install and connect from applications. User friendly to navigate the data with utility comes with it.

Drawbacks:

Most of the things are self explanatory, took very less time to use the product and integrate with applications.

In sum:

We are in the need of to use NO SQL database, and we thought Cassandra would be best choice to our needs.

Benefits:

The query language cql is quite intuitive specially for sql users. Setup is also very easy since they have extensive documentation. Cassandra also provides very good and easy to use APIs that can be easily integrated in any application.

Drawbacks:

It is a column oriented storage format which takes some time to get used specially for sql users.

Benefits:

It is very easy to get started with Cassandra because of the vast community of developers and resources available. It is flexible for any course project or production application

Drawbacks:

Configuring it to work with Tableau on a Macbook Pro is not that easy.

In sum:

If you are starting a data intensive project it is a very good datastore for the computing needs

Benefits:

csql syntax is powerful lightning fast read/writes smart client to reduce hops to the correct node easy to install

Drawbacks:

cluster management complexity problematic node recovery

In sum:

we needed to build a scalable object store with high write load (write is the main pattern of usage)

Benefits:

The is awesome tool for No SQL DB requirements. We use this to store transactions so that we can reprocess in case something goes down.

Drawbacks:

The Queries are little tough and most of the features of sql queries are not supported.

Benefits:

Opensource license provides an easy cost entry point Programmatic support for our purpose was very good.

Drawbacks:

There is only one commercial version of this product and licensing is high. It can consume far more resources to handle the workload you give it than spec data suggests

In sum:

Looks good on paper but has been very costly in terms of required resources to finally manage the load. Training material is limited and people with real experience are scarce.

Benefits:

I like it scales very easily when we see increase it traffic

Drawbacks:

Cassandra does not support joins so it is difficult to use as complex OLTP system

Benefits:

It was the best NoSQL solution for our organization. decent performance and cross-Datacenter replication.

Drawbacks:

It's written in JAVA. GC was killing us. at some point we had a server on GC for 30 min every 15 min. we had to add more and more servers to deal with the situation. DataStax's PHP driver was using the first server available as a 'Contact point' when it was on GC, the second server on the list was hit and sometimes caused to crash; as result - we had a domino effect and an entire cluster would have crashed. (that, btw, was fixed and now Contact point is random). There is a better alternative ***SCYLLADB**** Since we moved to Scylla - we were able to reduce the amount of servers by 75%. Performance and stability is magnificent. client side hasn't changed - running with the same drivers.

In sum:

We were able to offload tables from our MySQL to NoSQL.

Benefits:

- High availability as no master slave concept - Partition tolerance of the system - Replication across multi data centers

Drawbacks:

- Upgrade to a higher version is non-trivial and requires a lot of testing and evaluation - Cqlsh query results not formatted well which creates confusion in reading

In sum:

Overall it has been a wonderful experience using Cassandra with very low/none downtime to High availability and partition tolerance behavior

Benefits:

Cassandra is good choice from NoSql database when you need to store data which is coming at high velocity. As write operation in cassandra is very fast so it make it reliable to use it for real time scenarios where you have to process and modify the data in real time. And as Cassandra is built using Java, so it is easy to integrate this with other Java base tools like Spark, Hadoop etc.

Drawbacks:

Only cons for me is that it's unpredictable performance because cassandra has lots of asynchronous jobs and other background job which are not in control. In such cases, troubleshooting some issue is big problem.

In sum:

Just very good NoSql database in scenarios where you have to deal with large and hig velocity data in real time.