I've always been a fan of finding ways to be effective and efficient with my resources. I like creative, out-of-the-box solutions that attack a problem in a different way. This past week I ran across a post from Brent Ozar that captured some of my philosophy well. It was titled The Fastest Query is the One You Never Make and it's worth a few minutes of your time.
The post looks at caching and other creative ways of preventing queries against your database. In the vast majority of my career, the database has been a bottleneck that isn't easily scalable. Buying larger, more powerful hardware is difficult to accomplish in many companies, and it only solves the problem so long. Writing much more efficient T-SQL is great, but it can be hard to find, and keep, top-notch coders in lots of companies. Re-architecting the database can be difficult, if not impossible.
However I have found that the front end software is often much more malleable, especially when developers get the opportunity to avoid the database and implement some new technology. Caching is also the type of solution I've found many developers enjoy implementing. It's kind of cool and different, giving them a chance to a) learn, and b) do something unusual. These are both things that often get developers excited.
The lesson here is that the database is a limited resource, perhaps your most limited one. Finding ways to work with it more efficiently, and spread the load to other servers, or remove it, can make your application perform much better.
I'm hoping we can a database service, PaaS, Azure like software installation on our own racks of machines at some point, and we'll be able to scale the database as we need it.