Although there was snow in the mountains and a cool wind blowing off of the desert, ‘Vegas was hotter than ever and TDWI sizzled. ‘Vegas is good to TDWI. This town is always popular with data
warehousing pros. Last year’s conference posted over 700 attendees and this year, about 800 came looking for the latest action in data warehousing and business intelligence.
Sunday
After a smooth flight down from Seattle, I checked in and caught a class with a catchy title, eXtreme Data Warehousing, taught by Stephen Brobst and Richard Hackathorn. These guys have been around
this business since it started and have seen a lot. Their session focused on performance where response time is measured in seconds, the prospect of petabytes sized data warehouses, and the trend
towards the integration – in real-time – of data warehouses and operational systems. Companies that have modest, batch oriented data warehouses, will need to be cognizant of the technical and
organizational impacts that will come with the demand for data warehouses with high performance, big data, and low latency (real time or active data warehousing). Some companies are already there.
After class, I walked down to Napoleon’s at the Paris. This place has a particularly good house band as I had discovered last year. Sax player, Tommy Thompson led the group. He was backed up by
Chuck Hoover on keys, Teddy Davis Jr. on bass, and Eran Cohen on drums. They sounded even better the second time around.
Monday
Dave Wells, TDWI Director of Education, is constantly on the prowl for keynotes that will shake us up. Peter Fingar’s “The New IT: Lead, Follow, or Get Out of the Way” filled the bill nicely.
Fingar pointed out that offshore services are increasingly becoming more sophisticated. Good programmers here at home will not be able to compete unless they move offshore themselves – or add
business analysis to their skill sets. Fingar also pointed out that IT itself must change to meet the challenges of time-based competition vs. the cost-based competition of the past 50 years.
Fingar’s admonitions drove home the point that conference goers should pay close attention to such sessions as “eXtreme Data Warehousing” that show us how data warehousing and BI are evolving.
The session, “Leading and Organizing Data Warehousing Teams: Improving Individual and Team Performance,” was next on my dance card. Over the years, I have seen more data warehousing initiatives
hit the rocks because of organizational problems than any other reason. I was curious about what insights Maureen Clarry and Lorna Rickard might have to say on the subject. Using Denison’s model
for a high performance organization they used case studies to show how a lack of balance between adaptability, consistency, mission, and involvement can derail data warehousing initiatives. This
lack of balance can almost always be traced to interactions between personality types. Clarry and Rickard then pointed out problems with data warehousing efforts that are too project oriented (tend
to end up with silos), and efforts that focus primarily on the program (fail to deliver in a timely fashion). A good data warehousing initiative includes an ongoing program with integrated
projects, and each of these elements takes a different mind set. As the day progressed they gave us tips on how to make our different perspectives work for us. The session was time well spent.
In keeping with my focus for the day on the organization, I dropped in on Jill Dyche’s night school session, “Launching Your Information Center of Excellence (ICE).” Dyche’ describe how an ICE
fits between business and IT and how its focus on data management including, but not limited to the data warehouse, can make for a more effective enterprise. She showed us how an ICE is set up and
deployed. I immediately thought of several organizations that could use an ICE to break the ice between IT and business.
Tuesday
TDWI offers a session on Business Intelligence Strategies at its conferences where we can hear about best practices in BI and data warehousing. I sat in on a session titled “Case Study: How to
Upgrade to a High-Volume Enterprise BI Infrastructure” given by Chris Gentry. Gentry led us through a well thought out process for transitioning between a data warehouse/BI technical and
organizational infrastructure that an organization had outgrown to a more robust environment. Although it was a short session, it was worth the price of admission.
By Tuesday afternoon, I had taken in a lot and it was time for a break. I rented a car and drove due north for an hour. In response to recent rains, the desert sage was in its full blue-green
glory, and the red-orange sandstone formations in the Valley of Fire provided a magnificent backdrop. I found a quiet place to play my flute, and soon after I returned to the conference refreshed
and ready to go.
I arrived back at the conference as the vendor hospitality suites got into to full swing. I headed for the Microsoft soirée. Microsoft sponsored a number of its partners who had tables and
reps set up around the room – sort of a mini trade show. I took the opportunity to talk to Cizer.NET about their reporting product that supports Web based ad hoc reporting capabilities for
Microsoft’s Reporting Services. Hitachi Consulting’s presence indicated that Microsoft’s server products are being increasingly adopted by larger firms. I also had a chance to talk with Bill
Baker about Microsoft’s up and coming release of SQL Server and their enhanced ETL technology. He claimed that the successor to DTS will have greater capabilities, faster execution times, and
improved management facilities. I later talked with a friend who is in the process of beta testing the new product. He backed up Baker’s assertions. SQL Server 2005 sounds exiting.
Wednesday
Data stewardship is an important element in BI and data warehousing. Robert Seiner was offering a workshop on “Building an Effective Data Governance Program,” so I decided to sit in. Seiner
pointed out that data governance has been getting a lot of attention over the past few years because of the enactment of laws such as the Sarbanes-Oxley Act. However, he also pointed out that a
good data governance program can go a long way towards improving corporate data quality, and improved data quality can result in improved profits. Seiner then went on to show us ways to implement
such a program. The approach that he has developed centers on 3 Ds: recognizing de facto data stewards, creating discipline through formalizing accountability for the data assets of the
corporation, and formalizing the accountability of the stewards by recording information about the stewards in a database. He also suggested goals, metrics, and organizational structures for
implementing a data governance program.
Although the vendor show at TDWI conferences always includes new offerings, I decided to spend some time studying products and technologies that had been around for at least a couple of years.
I had looked at Golden Gate Technology awhile back and decided that their high speed data integration services, which were oriented towards transaction processing, didn’t really have a play in
data warehousing. However, the growing demand for real-time or active data warehousing has changed all that. Sabre, the company that supplies on-line reservations services to airlines is a Golden
Gate customer. Sabre found that Web shoppers do a lot of looking before buying. In order to ease the load on their transaction servers, they integrated a farm of inexpensive read-only servers with
their transaction servers, and Golden Gate Technology. It all works because update transactions are propagated to the read-only server farm instantaneously. That pretty much describes the type of
technical architecture that real-time data warehousing requires.
PolyVista is a tough product to define adequately. It definitely takes both intentional and un-premeditated data discovery to another level through advanced data visualization. It also includes
text mining capabilities. PolyVista, which is pretty easy to use, should make data analysts and savvy managers more productive through a short learning curve, with the reduced time it takes to
analyze a given business problem, and by the value of serendipitous discovery.
It seems like every other presenter talks about the importance of usage management in data warehousing, but I don’t see the supporting technology for it deployed very often. As a result, data
warehousing programs are not likely to get a high level of productivity out of their server platforms, and change management can become problematic. However, with HIPAA and Sarbanes-Oxley, there
are now additional reasons for managing and tracking user access. Teleran has developed a low impact monitoring and management technology that tracks and enforces policy management controls. This
technology is based on expert systems. It monitors who accesses what data and it can be deployed to pre-emptively prevent runaway queries. More companies should be looking at this technology.
Thursday
Wayne Eckerson, TDWI’s Director of Research, gave the keynote for the second half of the conference. He presented a pretty nifty data warehousing maturity model that should prove useful in helping
business sponsors understand where they stand in relation to others in the industry. The model included six stages from management reporting, through spreadmarts, then data marts, data warehouses,
the enterprise data warehouse, and then finally to analytic services. He claimed that as a company moves through these stages costs will decline relative to realized value. Eckerson covered such
measurement items as project/program scope, funding, team growth, and governance. He suggested that executive perceptions would change as the data warehouse moved into a position of strategic
importance.
One thing about TDWI, it attracts teachers who are the top thought leaders in data warehousing. Claudia Imhoff, who along with W. H. Inmon and Ryan Sousa defined the Corporate Information Factory
(CIF), was giving a class called “Six Steps to a Successful Corporate Information Factory.” The session was packed with content. Managers and architects who were just starting a CIF program would
have benefited mightily from it. Over the course of the day Imhoff covered program management, project definition, getting data in, getting information out, deployment, and administration and
operations.
Later on I went out to dinner with friends who were celebrating their impending marriage. We had a great time. The grouper was especially fine.
Friday
I wound up the conference by sitting in on Rick Sherman’s “Establishing Information as a Corporate Asset Using a Data Integration Framework (DIF).” This seemed to be a curious title given that
this is a data warehousing conference. Isn’t the data warehouse supposed to be a single, integrated source of data for the enterprise? Sherman did start off with this assertion, but he then
brought us down to earth.
In reality the average company has two data warehouses, six independent data marts, 4.5 operational data stores, and 28.5 spreadmarts or shadow systems – according to a survey done by TDWI in the
winter of 2004. Whoa doggie! Sherman then outlined a strategy for tying these different data sources together with the DIF. The DIF includes a tailored information architecture, processes,
standards, tools, resources, and skills. He then asserted that companies that adopt a DIF can expect more consistent data, cost savings with hardware, software, and training, plus increased ROI for
reporting and analytic projects. The DIF did sound like a practical solution to the real problem of multiple versions of the truth.
Sherman’s class was a good way to end up the conference. As my plane took off from Las Vegas, I read the Spring TDWI Conference brochure from a new perspective. Fortunately, TDWI does provide a
great mix of classes that reflect theory and practice. See you in Baltimore.
