Q&A with Davide Mauri
Building the Agile Data Warehouse with SQL Server 2012
Q: If your pre-con had a theme song, what would it be and why?
A: Wow, good question. Hmm… I think I would choose "One Vision" by Queen. It's just perfect: It’s fast, it’s powerful, and it’s full of energy! Plus, a data warehouse offers one (big) vision of a company's data. Yes, it's just perfect.
Q: What excites you most about the data warehouse?
A: There's a lot of data that has to be processed very quickly. And transformations are complex, so brute force cannot be the only solution. So you have to come up with clever solutions, which makes working with data warehouses very stimulating.
I started my IT career back in the ‘90s, in the "demoscene" subculture (http://en.wikipedia.org/wiki/Demoscene). Everything had to be done in real time, and when the most powerful computer was a 486, you had to be really clever to optimize your code. I've transitioned this passion to databases, and building and growing a data warehouse provides similar challenges. It’s also very rewarding, from a professional perspective, seeing how much your efforts can impact and improve others’ work. This is the Information Age, and information is exactly what people are asking for. The more information you can provide and the higher quality you can deliver, the better. Give them what they want, and they’ll be happy! And I'll be happy, too.
Q: Where does your workshop take attendees beyond what you might cover in a 60- or 75-minute session?
A: Well, we'll start from zero, defining why and how a data warehouse is useful and what its added value to the company is – driving through team management, the agile approach, and best practices to make everything work smoothly and efficiently. In this first part of the pre-con, attendees will understand the theory behind everything we’ll talk about. This is an important aspect for me, because understanding what lies behind the curtain lets attendees personalize what they see, so they can best fit it into their own organizations.
Of course, theory and practice must go hand-in-hand, so we'll also spend a lot of time solving real-world problems. By the end of the day, attendees will have a deep understanding of a data warehouse, how it's made, and how it works because we'll have built one using SQL Server 2012. Doing all that in less than 7 hours would just be impossible.
Q: What's the most surprising statement attendees might hear you say during the pre-con?
A: Ah, that's easy: "We don't have to build a solution that does everything the customer requested." It's a very shocking mindset, but it's what experience has taught me. In several of my projects, some of the initial requirements were never delivered. Yet the customers were so happy they agreed to make their solutions public case studies. And not because they were crazy. They simply realized that with the final solution, some of the original requirements weren’t needed anymore because what we gave them was even better.
Data warehouse projects are very long, and requirements will almost surely change during development. So the key is building an infrastructure that can quickly deliver the highest priority information and that can easily be changed and re-shaped to evolve with the business. This means we have to change the way we work, accepting changes as a matter of fact and not fighting them. Of course, this also brings into play important practices such as unit testing – which is well known in the development field but much less so in the BI universe.
Q: What’s the biggest myth surrounding data warehousing that you’d like to debunk?
A: The biggest myth nowadays is quite incredible to me: There's the idea that you can avoid using a data warehouse when building a BI solution. Today, managers want results very quickly. Of course, building a data warehouse – at first glance – goes in the opposite direction. Many companies failed to build a proper data warehouse in the first place, and someone came out with the idea that now with in-memory databases, it's not needed anymore - just connect to your source data, and that's it. They believe the brute power of today's hardware will make the magic.
That idea may sound attractive, but it is like going to a car dealer and asking to have all the individual parts of a car so you can put them together yourself, just because you don't want to wait for your car to be built by the manufacturer. It doesn't make sense to me. Of course, even if building a data warehouse is a complex process, we cannot simply say “wait 6 months” to people who need data to make important business decisions. We need to change the way we deliver data warehouse solutions. That's why the Agile approach works so well.
Q: What still trips you up in your adventures with the agile data warehouse?
A: Every project is different. It allows me to learn a lot, and learning is the most exciting thing for your mind, isn't it? It's surely difficult, but it's never boring. In addition, the Agile approach is still in its infancy in this field, so there's a lot to discover, improve, and shape. And discovering new and better ways of doing things is a challenge that I never tire of.
Q: If attendees could start putting into practice just one thing after your pre-con, what would you want that to be?
A: I would like them to focus on all the best-practices and rules I'll explain and use during the workshop. They are the result of several years of research and practice, and I use them in all the projects I lead. In all cases, they helped the development teams to be much more productive and to deliver very high quality solutions, all with the maximum efficiency. The reality is that in the IT field, there's a lot of architecture and very little engineering. And like it or not, the latter discipline is what makes a project cost-effective in the medium and long term. That's the whole point of this workshop: to define an engineered framework to make building a data warehouse as smooth as possible.
Check out our other Q&As with PASS Summit 2012 pre-con speakers.