write a Java-compatible RDBMS-like adapter for AWS DynamoDB

This is a crazy, ambitious project but my budget is only $300 USD. As a bonus, the chosen freelancer who implements this project will be granted the full copyright to the work. That's right, the copyright will go to the freelancer, not the employer. The only condition is that the final code must be released to the public under a GPLv3 dual-license, i.e. the freelancer will maintain the copyright for commercial purposes, but also release and upload to freelancer.com a GPLv3 licensed copy of the final code. The freelancer may maintain their real name or their company name in all the source code and copyright notices and include this work in their curriculum, advertisements, etc, and the freelancer will be able to create future commercial proprietary versions which do not derive from the GPLv3 license and make themselves insanely rich :-) The project requirements are also very simple to describe, but hard to implement ("easier said than done"): I want a Java-compatible, RDBMS-like adapter for DynamoDB. Here are the exact requirements decomposed: - "I want" : meaning, this is the core requirement for this project, in order for you to get paid, you must give me what I want. - "Java-compatible" : meaning, I can use a JDBC driver to connect to this thing like a normal database. - "RDBMS-like": meaning, I can use SQL and DDL language to define, create, update, read, delete items. The SQL implementation needs to be fairly complete, like an average RDBMS database such as H2, Derby, MySQL, or postgres. You only need to implement the least common denominator feature set among the examples mentioned. This requirement also includes indexes. When a "CREATE INDEX" statement is executed, a new DynamoDB table must be created to represent the index, unless a specific optimization enables it to be created as a GSI on the primary table. The requirement also includes using the aforementioned indexes when optimizing queries, in the same way that a normal RDBMS would (it would make sense for you to borrow the optimizer from an existing codebase, instead of writing your own). - "adapter": this is the piece of software that you will be writing. You (the freelancer) will decide exactly how this adapter is written. For example, you could start with H2 Database and replace the FileStore abstraction, or use Carbonado ([login to view URL]) or Apache Calcite, or rewrite a MariaDB Custom Engine, or you could write it as a database proxy, or just do something else entirely. - "DynamoDB": is the NoSQL database service from AWS (Amazon Web Services). This requirement means that your adapter will utilize DynamoDB as the backing store for individual table rows, index tuples, and relational definitions. DynamoDB does not support this out of the box. You will need to build the abstraction, such as a "tables" table in DynamoDB which defines virtual relational tables for your adapter. That's it. No further documentation will be provided. Employer may choose to contribute to the codebase at his discretion, via github. Project deadline will be 60 days. Weekly status reports with meaningful progress, must be provided.