Question-6: Is it a good idea to have heterogeneous nodes in the Cassandra cluster?

Answer: No, it is recommended that you use the Homogeneous nodes while creating or updating the cluster.

 Question-7: What exactly is the difference between Memtable and commit log, the way data is stored?

Answer: Memtable is an in-memory data structure and it stores the data in the same format as SSTable and indexed as well for faster retrieval. But in Commit log data stored sequentially which is useful when node crashes and you have to replay all the writes. 

Questions-8: why it is considered that Cassandra database is not a single point of failure database?

Answer: Cassandra database considered not a single point of failure database because there is no master node concept. Each node in the cluster is equal, find any node in the cluster can serve the request.  

Questions-9: what do you mean by coordinator node in the Cassandra cluster?

Answer: The node which is responding to read and write request would become coordinator node, whether it has a data or not. Coordinator nodes responsibility is to find the node which can serve the token range for read and write operation. During read it finds the node and return data from that node.  

Question-10: How the token range is distributed in the Cassandra cluster?

Answer: In the Cassandra cluster each node is responsible for particular range of tokens to be hold the data. Let's assume there is a token range from one to hundred. And in the Cassandra cluster there are 10 physical nodes with equal configuration. So, you can say that each node is responsible for handling data with ten tokens. For example, node-1 would take care for token range 1 to 10 and similarly last node would take care for token 91 to token 100.