Skip to main content

OOW 2013 : The new Oracle In-Memory Database

Today the new In-Memory technique for the Oracle database has been introduced. This is not something like TimesTen(+) or so, but a new way of querying data. The concept is like this: All tables that are marked for keeping data in-memory, will be "stored" using both row and column format. The good old row format will be used for the regular OLTP systems, while the column format is used for queries. To speed up those queries, the column format will be held in-memory - and in-memory only. Only the row format goes to disk, just like nowadays. 
To keep all data in sync, the in-memory data will be marked stale during a DML option that affects that data. Therefore the overhead added to DML statements is very small. Upon request the in-memory data will be updated. 
This in-memory data, a.k.a. the "Column Store" is an optional component of the SGA. So you have to switch it on. And you can pick which tables, which partitions, which columns and whether you want it distributed or not (over different RAC nodes) should be kept in-memory. 
The columnar data is loaded in compressed format. So it just takes a fraction of the space that the same total number of rows use (as usually a lot of columns contain the same values and can be more compressed). And this column store can / will be indexed as well.
So you can have an OLTP and a Data Warehouse / Reporting on the same structures. And depending on your query and data , the optimiser will either use the buffer cache or the column store. So the optimiser is "fully aware" of the in-memory database.
The In-Memory database is planned as an option (so additional licensing required) for the 12.1.0.2 release of the databae - expected next calendar year.
Post a Comment

Popular posts from this blog

Showing a success message after closing a modal dialog

APEX 5 comes with Modal Dialogs out of the box. Very neat. Especially for adding and changing data. And to minimise the number of time a user has to click, it could be useful to add a "Close Dialog" process after the actual data processing. When the data processing fails, the Dialog stays on top showing the error. When data processing runs fine, the Dialog is closed ... without any confirmation. And this might be scary for a shaky user.

So how can we provide the user some feedback? On Page 4 of the Sample Dialog Application you can see one solution: up on a Dialog Closed Event on the parent page it does a redirect to refresh the parent page appending the success message of the "Close Dialog" process. This has two drawbacks. First, it probably refreshes more than necessary. And second, if you're using multiple layers of dialogs (dialogs that open other dialogs) the message appears in the "parent dialog".
As an alternative you could follow these steps: 1…

APEX 5 New Static File Features

In APEX 4 you could upload files - like CSS files, JavaScript files, Images and whatever else you like - into the APEX Repository. When you navigate to Shared Components, there is a Files section that offers three different options:
CSS Files are always uploaded (and changed !) for the whole Workspace. For Images and Static Files (usually JavaScript) you could choose whether they should be available for the whole Workspace or for a specific Application only. And if you had a lot of files - e.g. a lot of images - then you had to go through the upload process one-by-one. But that's usually a one time only thing. If you make changes to the CSS and JavaScript files - and that's a continuous process in development - then you had to delete the existing file and upload the new one. Over and over again. And meanwhile fighting the cache of the webserver and your browser.  And another irritating issue: You couldn't use relative references in your CSS or JavaScript files as they just…

Using LDAP for Authentication and Authorization within APEX

One of my current customers would like to use their LDAP (Microsoft Active Directory) server for authentication and authorization of APEX applications. Of course we tried to set up a standard LDAP Authenication that's available within APEX. But we couldn't get that to work. Maybe it has to do with the fact that the client stored their Users within Groups within Groups within .... . Or maybe it doesn't do a full tree walk in the directory. Or maybe it is just because it is Microsoft - and not Oracle Internet Directory (OID). So we moved to a custom Authentication using the DBMS_LDAP functions (and some examples from the Pro Oracle Application Express book and Tim Hall - a.k.a. Oracle Base).

One of the issues we encountered that we wanted to use the user's login name, like "jdoe" and not his full name ("John Doe"). And the login name is stored in the "sAMAccountName" attribute. But authenticating using just "jdoe" didn't work. …