small medium large xlarge

01 Dec 2011, 12:55
Handojo Goenadi (1 post)

Thanks for writing the book. A very interesting book topic. I have not read a book comparing many open source database.

I hope this book will describe in memory database too. Since memory becomes bigger and cheapper.

12 Dec 2011, 02:55
Eric Redmond (21 posts)

Not in memory per se (like the MySQL memory engine), but we do cover databases that function better with more RAM (such as Neo4j, or indexing in Mongo). That said, you can certainly gain performance by increasing the amount of memory used in any database. Beyond that, SSD will certainly help DBs fly as they get cheaper – they’re basically like giant flash drives with crazy fast i/o.

12 Dec 2011, 16:31
Jim R. Wilson (84 posts)

Hi Handojo,

In-memory databases are interesting in that they’re really fast, and consequently quite a bit more risky than, say, a disk backed database. If the server reboots unexpectly, for instance, all the data is gone. Prototypical examples of principally in-memory databases include memcached, Berkeley DB and SQLite.

We don’t cover these in particular because our main goal is to cover different methodologies and styles of databases to introduce the reader to what’s out there. For example, we cover PostgreSQL to address relational databases, and Redis and Riak for modern Key/Value stores.

We do talk about memory-related details from time to time, but, per Eric’s comment, we don’t cover strictly “in-memory” databases as a separate concept. Hope this helps!

You must be logged in to comment