http://community.ingres.com/forum/blogs/stephenb/ en Tue, 09 Feb 2010 15:27:40 GMT vBulletin 60 http://community.ingres.com/forum/ingres4/misc/rss.jpg http://community.ingres.com/forum/blogs/stephenb/ http://community.ingres.com/forum/blogs/stephenb/4-whats-happening-dbms-group.html Wed, 10 Jun 2009 06:00:55 GMT I thought a blog might be a good way of letting people know what's happening with the Ingres DBMS, and hopefully we can get some feedback. So here are some of the things we're working on right now:

• MVCC - For those that aren't familiar with this acronym, it stands for Multi-version Concurrency Control. It adds the ability for the DBMS to create multiple versions of the same data in a controlled fashion, allowing readers to see the old version of a piece of data as of a given point-in-time; this allows for consistent reads without locks, if another thread updates a page, they simply create a new version of it and the reader continues to read the old version. Once the data is committed the reader may see it or not, based on the point-in-time of their consistent read. So readers don't block writers (and vice-versa), making it a very powerful feature for OLTP since it increases concurrency tremendously. See here for details on that we're doing (http://community.ingres.com/wiki/MVCC)
• data at rest encryption (column based) - this is something we need for a number of standards, and it significantly enhances our security story. MikeT is the point man on this, check out the project page here (http://community.ingres.com/wiki/Dat...est_Encryption)
• support for long object names - This feature will help us port third party partner applications to Ingres. Some other vendors support longer object names than Ingres, and it causes a lot of porting work in certain applications. The current thought is that we will support object names up to 128 bytes (number of actual characters may vary if you're using UTF8 or DBCS). This feature is available in the source now and can be built with a special flag if anyone wants to try it out. We'd really love to get feedback on how this affects existing users and interfaces. e.g. what happens when you try to access a table with a long name from an ABF application. You can check you the details here (http://community.ingres.com/wiki/Long_Identifiers)
• scalar sub-queries - this feature allows you to use other queries within the "SELECT" cause of a query. We have this in a limited form already; this project will implement full ANSI compliance for this feature. The project has just started, so there's not much to say, but you can follow the progress here (http://community.ingres.com/wiki/Scalar_Subqueries)
• Better error handling for offline recovery - some parts of this feature are in Ingres 9.2 (unsupported) and we plan to round it off in the next release and support it properly. Details are here (http://community.ingres.com/wiki/RCP...Error_Handling)

There are a bunch of other things in the works which I'll try to talk about in the next post.

You also might be interested in the DBMS features some community members are working on. Marty Bowes is close to being ready with his check digit functions, see here (http://community.ingres.com/wiki/Check_Digit_Functions). Karl Schendel is working on a bunch of things, including build system improvements (http://community.ingres.com/wiki/Bui...rameterization) and partitioned global temporary tables (http://community.ingres.com/wiki/Par...mporary_tables).

I'm also trying to bootstrap a project to improve the performance of LOBs by inlining them into the base table (and getting rid of the etab). We need this for our geospatial work (http://community.ingres.com/wiki/IngresGeospatial), but it would be generally useful to anyone using LOBs extensively . I have designed a prototype; if anyone wants to help me code it just drop me a note. ]]> stephenb 175 http://community.ingres.com/forum/blogs/stephenb/4-whats-happening-dbms-group.html