Dan Kusnetzky proposes the following 6 features as part of an ideal transactional database:
- ACID transactions
- Data and application independence
- Geographic distribution
The real question would such a database including all these features be possible? We already know that ACID transactions and support for geographic distribution don’t mix well. SQL was created to work with the relational model, so it’ll be quite limited when considering other data models—think graphs. There are also some (good) arguments why a declarative language as SQL might not be the best fit for large scale databases. Last, but not least, designing a common API to support the different data models is not that realistic either.
Original title and link: 6 Ideal Features for Big Data Transactional Database ( ©myNoSQL)