Parallel execution is used to improve the performance of large queries and DML operations by allocating additional slave processes to complete the task. The extra processes put an added drain on resourses which can affect performance of other statements.
As for the next bit of your connection:
DEDICATED - N/A. This is to do with how the client connects to the server, not the type of processing the server does.
MTS - N/A. This is to do with how the client connects to the server, not the type of processing the server does.
OLTP - Typically OLTP systems service lots of small transactions where data is accessed by it's primary key. Such transactions benefit little from parallel execution. In addition, the extra resources required may actually reduce total throughput on the server. There may be some scope for parallel execution of batch processes on an OLTP system during quiet times.
DSS - This is what parallel execution was designed for since queries against DSS usually require alot of data to be retrieved and agregated. In addition, most DSS do not suffer from such high numbers of transactions as OLTP systems so the resource drain associated with parallel execution does not affect other users to the same degree.
Remember, switching this parameter to true will not automatically make operations execute in parallel. This will only occur if the relevent parameters are set at table level or you are using optimizer hints.