My question will be about appropriate NoSQL solution rather than asking Cassandra related questions.
In our case:
- We have more than *denormalized* 4 million rows data and at the end of this year we are expecting 5-6 million rows
- Every minute maybe more than 1000 rows can be updated in some situations.
- These rows has more than 4 index in different columns.
- The system we want to build does not has transactions.
As far as I know from the things I learnt in last two months in this mailing list, Cassandra seems to fit for this operation. But Katta, MongoDB and other solutions made me little confused.
In such a case what is the appropriate NoSql solution?