Add data structure to handle in-memory records for the Gora-DynamoDB module

While writing the GoraCI module for the Gora-DynamoDB module an in-memory data structure was not envisioned because there was not a clear advantage in terms of number of IO operations, and "cost" in general. Some time ago, Amazon Web Services released the batchFuncionality which we can use to persist and read groups of items in a single operation (25 for writes and 100 for reads). Now we can also decide on the consistency level used for these functionalities [1].

[1] http://aws.amazon.com/dynamodb/