Sunday, September 20, 2009

MS SQL 2008 or Oracle 11g


I was invited by an associate to attend a RFI (Request for Information) meeting. It is for a tender which is to be awarded soon.

One of the customer's requirement is for a scalable, highly-available database. At the same time, it has to be a cost-effective solution.




To be on the safe side, most bidders will usually opt for Oracle 11g. This is especially so if the customer is from the defense industry.

However, I'm not very keen to quickly jump into a conclusion. I wanted to find out more ...



I know that Microsoft SQL Server 2008 has improved over the years and it's really worth giving it a chance.

Important selection questions:
1. Performance
2. Data Size (especially historical data)
3. High-Availability
4. Scalability (concurrent access)
5. Data Warehousing Requirement

I proceeded to show them the diagram below - "Cost vs Performance".



I then asked - 

"Are you willing to accept a lower performance in return for a less costly solution?"

The key is that the performance must be at least acceptable. We should not compromise on this just for a cheaper solution. 

Personally, I know of a listed Australian logistical company that is in the process of migrating their backend databases from Oracle to Microsoft SQL Server 2008. 

I know the decision maker and I was curious to know why. His reply:

1. Oracle refuses to cut down their price
2. Microsoft SQL is slowing catching up with the release of version 2008
3. We can accept the lower performance as long as the business users do not complain when they access the web applications
4. I save a lot of money by not having to send my developers to learn Oracle database

Surprisingly the customer that I talked to kept nodding her head when I mentioned the last point. Her developers are junior (mostly fresh graduates) and they are not comfortable with advanced technology. 


So the point is when you adopt a technology, put aside cost/performance, ask yourself whether or not the people you have is able to fully utilize that new technology. 

Otherwise, it's pointless to waste so much money on expensive product only to realize that your developers do not know how to utilize it.




No comments:

Post a Comment