Here, my boss, is asking me and also the people in the office which Database engine may be the perfect for a really very large amount of information. We've been talking about a whilst in the office with no conclusion came... But when i believe that this can be a pretty interesting question for talking about
Like a DBA, here's the actual answer.
The very best database platform on the planet could be not able to handle tiniest of loads without having staff that understands how to utilize it.
A reasonably low-finish database platform can scale as much as huge loads for those who have staff that knows it inside and outside and works around its weak points.
For this reason many people say Oracle can't scale, and why others say MySQL can rule the planet. A great music performer can continue to have beautiful sounds having a crappy instrument, and crappy music artists can't get the best instruments seem good.
When you are determining what platform for the next project, don't go tossing dice and taking advantage of an answer you've got no knowledge about. If you will need to scale up, either make use of the solution you realize inside and outside, or hire someone who knows the brand new solution inside and outside.
A few of the biggest databases on the planet are operate on Microsoft's SQL server: http://www.microsoft.com/sql/prodinfo/compare/wintercorp.mspx
And here's all of the the ten biggest databases on the planet (although I've no clue how accurate it's):
The very best may be the World Data Center for climate with 220 terabytes of web data and 6 petabytes of more data
To obtain a significant answer it can help to know how you feel a really very large amount of information. It might also help to be aware what it might be employed for and just how it might be organised.
Whether it's gb of information most databases could be fine.
Whether it's terabytes of information i am confident Oracle or SQL Server will not have trouble with it. If they are set up and designed properly.
Whether it's bigger... something similar to bigtable.
Most likely more essential than understanding how large is big would be to know give me an idea related to the database. If it's an OLTP you will then be fine with the primary gamers as clarified before but when it's for any confirming tool then you might want to investigate Alterian and Sybase IQ rather.
Really is dependent on which your concept of huge is, and what you would like related to it. For SQL-like access:
- Gb of information may be easily handled by any FOSS or commercial product.
- 100s of gb+ results in something similar to Teradata
For additional specialized processing, Hadoop and HBase work. (Several similar items exist including Sector/Sphere and GridGain to title a few.) Hadoop is really a cloud-computing architecture patterned on Google's filesystem, and may hold 100s of petabytes. HBase is really a "database" which operates on Hadoop, concentrating on the same abilities. I only say "database" since it is column-oriented, a really different model from row-oriented databases like MySQL, PostreSQL, Oracle, etc.
Hadoop/HBase tend to be more suited with the idea to data warehousing, or situations where one can precompute the queries you will need to run, and also have them performed out-of-band via MapReduce.