You can download the complete installation here at Oracle Site, or as Patch 17694377 from Oracle Support.
This DB release is known for it's In Memory option (No, this is not Times Ten) covered by Oracle White paper, Oracle Page, Oracle Blog and some nice blogs like test on counting 1 Billion records by Antony Heljula from Peak Indicators or the post by Mark Rittman.
Actually there are some additional new features in this release, such as:
- JSON support in Database (links: Oracle Docs, Blog and slides).
- Approximate Count Distinct (APPROX_COUNT_DISTINCT() function).
- Advanced Index Compression.
The new features guide is here.
The Upgrading to 184.108.40.206 white paper is here.
Finally, a personal experience:
It seems that the "In memory" term is used very often by different vendors, with various meanings.
I had a conversation with a customer few days ago. He assumed the in memory option of MS 2014 DB and Oracle 220.127.116.11, described here, are the same. They are not! As far as I understand, The Microsoft option is oriented on OLTP processes, you need specifically define tables as In-Memory and (If you don't like loosing data) make sure the log is created for them (there are many other limitations).
The Oracle In-Memory is designed for Analytic, the existing relational tables are updated as they always were, in the regular method. We can define additional in-memory columnar data storage on any table or partition we want. The DB optimizer can work with both. As a result every thing that worked before, keeps working. Just faster.
Microsoft has a different tool for analytic data, that is the Tabular, but it seems to me more of a Microsoft Analysis Services replacement. What I like about the Oracle 18.104.22.168 In-Memory, is how open and flexible, though powerful this option is.