EXPERT RESPONSE
As do we all! Below are my current favorites.
Once upon a time there were many RDBMSs. To put a figure on it, back in 1992 I reviewed 19 engines in one bunch test and those were limited to engines that ran on Microsoft operating systems. In 2007 the only engines with a significant share on the market are Oracle, IBM and Microsoft. This doesn't mean that niche players like Teradata are not important, they are very important, but not as volume players.
So, one trend has been shrinkage, but I don't think that this will continue, all of the small players have now gone. However, if pushed as to which of the remaining three looks the weakest, I would have to say Oracle. This does NOT mean that I think that Oracle is about to go, it clearly isn't. And yes, I do know that Oracle is a huge company (they all are, which is why I don't think any of them WILL go soon) but Oracle's problem is that it is a one trick pony. It does databases. And it does them very well. But essentially it remains a data handling company. It can't cross-fund its database business from its other revenue streams; but the other two can. And Oracle also sits in the middle, being squeezed from below by Microsoft and above by IBM. Given intelligent leadership I'm sure it can survive -- and Oracle is led by Larry Ellison, the man who built the company in the first place. (On the other hand, now that I think about it, this is also the man who brought us the Network computer (NC) and Raw Iron as the technologies that would revolutionize the internet. You remember those, surely……?)
So, I don't think the shrinkage trend will continue but I think that the current focus on business intelligence (BI) will continue to be important. BI is essentially about the extraction of information from data. All of these vendors are adding BI functionality as we speak and the money they are pouring into this tells us that it is a major trend in the DBMS world. (On the day I was writing these words came news that IBM had bought Cognos; a major BI player.)
Multi-dimensional data (MDD) also continue to grow in importance. Hyperion was a leader here, but has just been bought by Oracle. Rather oddly, Oracle used to manage MDD in its relational engine which always seemed like an odd idea (like trying to teach a Penguin to fly). So maybe Larry has finally noticed that a separate MDD engine is a good idea.
Just in case this is beginning to sound like a bash-Oracle session, let's remember Grid computing, at which Oracle excels. This brings us neatly to the general topic of increasing engine speed and scalability. The need for speed has led to a merging of massively parallel processing (MPP) and symmetrical multi-processor (SMP) technologies. Effectively, this is allowing us to treat commodity hardware as nodes in a cluster, create very large, multi-processor systems very cheaply and make them very scalable.
The good news is that there are multiple ways in which this can be done. Oracle's Grid computing is an excellent example. Another way is by deploying our next trend, the data warehouse appliance. There are multiple offerings of DWAs but many now make use of commodity hardware and can be blisteringly fast. And of course, HP has jumped on the bandwagon with Neoview, which is shaping up to be a very interesting product.
Other important trends? Better visualisation of data, third generation BI (where we use the information coming out of the BI system to directly control the incoming transactions), developments in MDX (the analytical language for MDD), developments in geospatial data types (another area where Oracle was first to market) and, of course, the developments in XML.
So, nothing much happening in DBMSs then...
Also read Mark Whitehorn's two-part series about data warehouse appliances:
Business intelligence and data warehouse trends: Appliances and "third-generation" BI (Part 1)
Business intelligence and data warehouse trends: The challenges of third-generation BI (Part 2)
|
