Choice Is Good, Provided Its Manageable

Opinion: The embedded space is home to a wide variety of database options. If vendors offered application appliances including hardware and software, then the only choice becomes which application.

Download the authoritative guide: Big Data: Mining Data for Revenue

Sometimes its easy to think that the database universe is made up of only five or six players, depending on who you talk with. But that is far from the reality. The market is bursting with dozens of database choices, but most of us go happily about our daily tasks, confident that we have a good handle on who the players are in this space.

Dont get me wrong: Limited choice does make life easier. I mean, I have covered this space for almost six years and have spent virtually my entire 20-year career in database-related fields. Too much choice is messy. Its a lot of work to know you have the one best solution. That the database choice you have made is scalable, portable, secure, highly available, fast, cost-effective blah blah blah blah blah.

Databases like Intersystems Cache and Progress Softwares OpenEdge are two examples of very widely used databases that almost nobody has heard of. You find these databases used in the healthcare, financial services and gambling industries among others. There are in-memory databases and XML databases. There are still a great number of applications that use other approaches, such as network models, hierarchical models. There are innovators like Lazy Software and its Sentences database, built upon something they call the associative model.

Other types of innovation can be found, such as ANTs Software, which follows the relational model but has virtually eliminated the need for locking. There are even examples of people still quite happy with their multivalued (remember Pick) database, such as IBMs U2 database line it acquired from Informix which in turn acquired it from Ardent …

/zimages/3/28571.gifClick here to read more about ANTs.

So just below the surface in critical areas we may not be aware of lies a database doing its job without much fuss. This, of course, is what separates the embedded database from the pretenders. If you are aware of it, its probably not doing its job. The point here however is that you do have alternatives. ISVs that develop niche applications are certainly aware of the alternatives that highlight the freedom one has when the database is only meant for a very specific job.

The embedded space and its focus on the SMB market is desirable because it is supposed to be a breeding ground for new license growth. IBM, Oracle and Sybase have all released editions of their databases to appeal to this space. Microsoft has a built-in audience because of the success of Windows.

Still, if it is a breeding ground, why havent we seen and heard more about some of these alternative databases? Well, just maybe, the embedded space is pointing towards a future that is in many ways a flashback to our past. Imagine applications sold as a standalone solution including hardware and software. Preconfigured and certified. Supported remotely and linked to other application appliances via Web services standards.

If such a future were possible think of the efficiencies that could be gained. IT in its early years was a highly integrated stack of hardware and software. Client/server forced a period of disintegration. There of course were strengths and weaknesses to both approaches.

Perhaps in the future we can have our proverbial cake and eat it too. With powerful and cost-efficient application appliances we could simplify licensing and support. We no longer wrestle over decisions like whose Intel-based hardware we should buy, what operating system is best and what database is our standard. We simply decide which application solution offers us the most and let the ISV worry about those other decisions. Their appliance provides the storage, the server, the software and the support for it all. Open standards provides the interoperability.

/zimages/3/28571.gifIBM is building the Information Virtualization Server to enable customers to change data sources or rules without affecting applications. Click here to read more.

Now its true that every corporate application need will never be fully satisfied by an ISV, but many can be. If that day ever happens we should expect to see more and more database options. Many ISV developers will even have the freedom to revive older technologies or explore newer ones that, while useful, might have never survived. Remember that the packaged application was supposed to save us development costs, time and effort. Perhaps that can still be achieved.

Anyone who had to implement SAP or some other ERP application in the last 10 years would have welcomed an SAP appliance. I imagine SAP itself would have loved to be able to sell an SAP appliance. It helps you and them. Its never been successful before because of a number of issues involved, not the least of which was the best-of-breed mantra that many bought into. Perhaps people are a bit wiser now. Perhaps the IT industry as a whole has matured enough to take a page from its past, coupled with some of the current trends, standards and technology to finally give us what we all desire: a simpler approach to IT infrastructure.

Charles Garry is an independent industry analyst based in Simsbury, Conn. He is a former vice president with META Groups Technology Research Services. He can be reached at

/zimages/3/28571.gifCheck out eWEEK.coms for the latest database news, reviews and analysis.