Oracle Vs Ms Sql Server to handle around 6-7 Terabytes of data

Question

I am considering Oracle 11G to handle around 6-7 Terabytes of data. This data will get accumulated over a period of 8 Months. Any data older then 8 months is purged out of the database.
Some of my MS SQL DBAs are saying SQL server offers similar performance at a very low cost. Is this really true? On What basis or parameters I can compare Oracle with MS SQL Server to nullify the price advantage?
I am not looking for ease of development, I am purely looking in terms of handling the huge data and still be able to provide good performance. Apart from performance if there is any parameter which makes any difference then please highlight.

Answer 1

As with any database requirements tend to drive the best technology to utilize.
I have run Oracle and MS SQL Server as well as MySQL and PostgreSQL.
If you are looking to support a fairly large user based on a database that is 6-7 TB is size where a good portiion of the data is subject to access then there is really is no comparison. Oracle does a much better job of handling large user bases with large data sets then MS SQL Server or MySQL at least based on my experience with them.
MySQL works very well for average user bases and smaller databases where the amount of data accessed is small. In other words you can have 6 TB of data and if you only frequently access 5GB of it then MySQL will work pretty well. If a large portion of the data is accessed all the time by many users you will struggle with MySQL. Most databases in MySQL I ran had issues at user bases of a few hundred with a database size of around 100G-200G, now my main expertise is in Oracle so there may be ways in MySQL to address this, but not easily as normal research and experiment channels did not yeild good results for me.
I have had a MS SQL Server databases with thousands of users fall on it face with a database only 20GB in size, its was completely CPU bound with 8 CPUs, then moved the database to Oracle with minor code changes to accommidate the differences between Oracle and SQL server using the same hardware specs and CPU was a steady 50% for the same work load and basically the same application.
I have found the capabilities of Oracle for large data sets and large users provide advantages that the other databases just did not offer or did not do well enough to overcome. Does not mean that Oracle is perfect, or the other databases are bad, but there are a lot of options and capabilities in oracle the I can take advantage of not to mention if I need to I can run on high end Unix hardware and goto 64CPU dozens of I/O channels, etc. I can not do that with MS SQL Server as Windows is my only option and the intel platform just does not hit the capabilties of higher end Unix Hardware.
When deciding the database to use cost should be the last factor used. If the database will not do what you need it to do then low cost does not help you. Deteremine the needs and requirements and then examine the database features and capabilities to determine what might work best for your situation. If it is a critical business function and you need the caddie then your business will pay for the caddie, it if needs a malibu then pay for the malibu or if it needs the yugo then pay for the yugo. Implement the right technologies where you need them. However there is something of an advantage to limit the number of technologies you will need to support as a jack of all trades never becomes an expert and that in of itself can be costly when trying to overcome issues developed by limitations of a database technology.