I’ve created another video of the demonstration that we used in the second Ask TOM Office Hours session where we focused on showing why column store scans are faster than row store scans. I’ve had many questions from customers and conference attendees who ask, “What’s the big deal, if I just cache my entire table in the buffer cache, won’t it be just as fast as being in-memory?” We’ve even done two different blog posts that discuss this:
In-Memory Column Store versus the Buffer Cache
Memory Usage with Oracle Database In-Memory
In this demonstration I show the same query accessing the In-Memory column store and the buffer cache or row-store. I use the same Star-Schema Benchmark (SSB) schema that we’ve used in our other blog posts and fully populate the tables in the IM column store and, using the KEEP pool, fully cache the tables in the buffer cache.
I then briefly discuss why querying the IM column store is so much faster and how the features of Database In-Memory enable orders of magnitude faster queries.
The video is available at the link below and is located on the Database In-Memory YouTube channel. While you’re there, check out some of the other videos we have available as well.
Original publish date: May 11, 2018