This is the fifth year for Eye on TDWI, and it has come time in my career to turn over the pen to a fellow professional in the field, Mike Lampa. I have been actually been attending TDWI
World Conferences in some capacity or another for about ten years and have probably attended more TDWI classes than anybody on the face of the planet. However, I have never come away from a
conference thinking that there was nothing new to learn. I hope that Mike will find his stint as the author of “Eye” equally rewarding.
Introducing Mike Lampa (New Author of Eye on TDWI)
Note from Mike Lampa:
When Tim propositioned me to take over his column, I was met with emotions of honor and humility in equal doses. I appreciate the confidence of TDAN.com and TDWI and will strive to fill the
big shoes Tim has left behind.I currently run a data warehouse consultancy practice, TeamDNA, Inc., along with my partner Dave Davis. Yes, I’m one of those guys that has been “doing data” for a long time
(28 years). The majority of my implementation experience is in data integration, data warehousing, business intelligence, strategic planning and program management. I believe in the value of IT
solutions that are founded on a strict adherence to model based/model managed architectural disciplines. I also believe in the power of reuse and the development reference patterns using common
sense methods.You’ll see me around at TDWI conferences as an instructor and as an
attendee. I look forward to sharing conference experiences with you all
in the future. My best wishes to Dave Wells and Tim Feetham as they pursue their next personal and professional life-chapters. A couple of great gentlemen that have left a legacy for us to carry
on!
Orlando Conference
Given the occasion, I thought that my last column should include a bit of a retrospective as well as a look into the future. As it turned out, the Orlando conference turned out to provide a number
of opportunities to weave that retrospective into this report. Even my time the week before the conference when I went kayaking with friends on several nearby rivers fit in, but more about that
later.
I usually give a chronological review of each conference, but this time, there were several threads that stood out enough to warrant following each one individually over the course of the week. The
recent vendor consolidations in the business intelligence (BI) space drew me to Cindi Howson’s Evaluating BI Toolsets and BI Tools in Action. I followed up her class with a visit to
the BI toolset vendor exhibits. While at the vendor hall, I dropped in on a new database demo and that triggered visits to other vendors and a night school session on evolving database
technologies. I followed up this thread with Dan Linstedt’s VLDB (Terabytes to Petabytes): Concepts and Architectures. Each of these threads contained significant developments.
However, the most rewarding train of thought to evolve out of the conference, and perhaps the decade, came from Dave Wells’s Putting the Business Back into BI, Anthony
Politano’s CPM and MDM, and Mark Madsen’s keynote, Outside In: The Next Generation of BI Innovations.
BI Tools: Something Old, Something New, Something Borrowed…
Cindi Howson has been offering her analysis of the BI vendor space for a number of years. However, Howson continually updates her class. Focusing on the information delivery and analysis aspects of
these toolsets, she has always been thorough, tough, even-handed, and current in her evaluations. Vendors have come to respect her, and students are sure to learn valuable insights in her classes.
Howson reviews all of the major BI players in the morning, and then picks three vendors (different for each conference) for a scripted shoot-out in the afternoon. This time she selected Business
Objects, Information Builders, and MicroStrategy.
Howson gave us evaluation sheets so that we could keep score during the afternoon session. I expected the market leader, Business Objects (recently purchased by SAP), to walk away from the others.
However, when I had totaled up the numbers, I had scored Information Builders in a dead heat with Business Objects, and MicroStrategy tallied only a point behind. All of the vendors showed products
that supported basic and advanced reporting as well as data analysis. While Business Objects showed well in all of these areas, MicroStrategy got top marks in data analysis. However, Information
Builders was impressive with its new developments in portable reporting databases, Excel integration, unique search engine integration, scalability, and totally customizable data access development
platform. Indeed, Information Builders would have come out on top overall on my score sheet except the demo team conceded the data analysis part of the shoot-out to MicroStrategy and ended up
nailing down third place in that section.
Later on, I saw Information Builders’ OLAP viewer in the vendor area. Although, in my view, it was not as strong as MicroStrategy’s tool, had the Information Builders’ demo team
put a little more effort into that section, I would have scored them tops overall. However, I will give Information Builders credit for focusing on its core strengths and demonstrating leadership
in developing mass market reporting.
Database Technology Revisited
While I was in the vendor area looking over BI toolsets, I stopped by the ParAccel booth to check out this new database vendor. ParAccel offers a high performance, columnar database that supports
standard SQL-based queries generated by most of the leading BI data access tool vendors. Although databases such as Sybase IQ and Sand Technologies, which also use similar columnar storage and
compression techniques to enhance performance, have been around for awhile, ParAccel took these technologies a step further, added massively parallel processing, and recently partnered with Sun to
post a rather spectacular set of TPC-H benchmarks in the 100GB to 1TB range. They also posted some impressive gains in the price-performance arena. (The TPC-H benchmark, found at www.TPC.org, is the standard test for comparing databases designed for typical data warehousing queries.)
ParAccel’s entry into the analytical DBMS space had spurred my interest in advances in the analytical database market in general, so I stopped by the Sybase IQ booth. They have been in this
space for a number of years and have developed a devoted following. I was impressed with some of their customer success stories. Later, I sat in on Sam Madden’s night school session on
columnar database architecture. Madden is an associate professor at MIT, and he gave a good overview of the subject. I came away thinking that any organization looking for a cost-effective, high
performance analytical database should include vendors with columnar architectures such as ParAccel and Sybase on its short list.
Given that I was now fired up about database technologies, I decided to catch Dan Linstedt’s class, VLDW (Terabytes to Petabytes): Concepts and Architectures. Linstedt pointed out
that when it comes to really large data warehouses that approach the petabyte range, they usually have a dramatic impact on visibility and the bottom line. This makes gaining all of the information
one can about working with VLDW’s (very large data warehouses) central to one’s success. Linstedt told us that there are very few database architectures that scale into this area.
Although he reviewed VLDW technology and mentioned Teradata and IBM as leaders in this area, he focused mainly on implementation issues from team makeup to tuning. Linstedt has a lot of experience
in this area, and we all gained from it. It was a great session.
A Significant Contribution to Our Discipline
Dave Wells, one of the chief thought leaders in our field, has focused his attention on the IT-business divide and the continuing challenges of creating robust BI environments. As Director of
Education at TDWI, Wells has brought in a number of keynote speakers and instructors who have expanded our understanding of this divide. He has also authored several courses that focused on
consolidating what he has seen organizations gain from effective BI implementations. He has also built a course around a BI architecture that places the significant parts of the discipline in a
structure that has made it easy for numerous newbies to come up to speed. However, during this conference, Wells went back to square one and presented a class based on a new BI framework that he
created strictly from a business perspective. As he laid out his thoughts, I had the feeling that I was experiencing one of those contributions to our thinking that comes along maybe once in a
decade.
Wells started out with asking us to lay aside our technology-based definitions of BI. He suggested that we consider the separate definitions of business and intelligence. He took us through a BI
value chain that included these actions: reason, plan, predict, solve, abstract, comprehend, innovate, and learn within the context of business drivers, goals, strategies, tactics, and outcomes.
From this value chain, Wells created a framework in the form of a cube that included business management (i.e., strategy, finance, HR, customers, operations, etc.) on one axis, corporate governance
(i.e., policy, compliance, risk, etc.) on the second axis, and measurement (i.e., measures, metrics, references, indexes, etc.) on the third axis. Interestingly enough, Wells was ambivalent about
where IT fit in. Is IT a plane on the business management axis, or is it a part of each of these business activities, and, by the way, I thought, why is any of this important?
However, Wells was quick to give his framework a context for meaning. He quipped that first off, many BI projects create elegant solutions to the wrong business problems. These outcomes arise for
several reasons – i.e., BI is often a hook to sell software and is often offered within the context of delivery systems. (This isn’t just an IT problem. Business frequently couches its
needs in terms of technology.) When IT and business units do try to hammer out the business context for a given BI project, the result can be too narrow in scope (e.g., business wants to track a
certain composite index). IT builds the database and implements the software to report on it. However, once business observes index values that are out of line, it wants to analyze the numbers
making up the index. Does the original system support this? This is a bit over simplified, but the application of Wells’s framework to requirements gathering might give IT and business a
fighting chance of creating more robust systems that support business management, governance, and sound measurement. Perhaps the biggest contribution IT can make is not to offer technology as a
solution by itself, but to help business with a logical business framework for understanding issues that both can address together.
Connections
When Mark Madsen gave his keynote, Outside In: The Next Generation of BI Innovations, on Thursday morning, I saw a couple of connections with earlier sessions. One of his observations
about BI was that that we are still using the language of control and the potential of BI technology will not be realized until it fits our evolving life style – that is, becomes highly
available, served the way we want it, where we want it, and accommodates social networking. One of the things that occurred to me in Wells’ class was that the implications of using his
framework for gathering requirements might lead to a more collegial relationship between IT and business. This seemed to be more of a fit with Madsen’s view of the future of BI than the
control model often incorporated by IT. On the technical front, Information Builders also seemed to be in step with the future of BI, given its service oriented architecture, customizable front
end, and take it anywhere BI capabilities.
That afternoon, I sat in on Anthony Politano’s CPM and MDM. Although I hadn’t recognized a connection between Wells’ class and this one when I had signed up for them, it
quickly became apparent that there was a common thread here as well. While Wells had laid out a logical business-oriented framework, Politano gave us techniques to put it to use. Corporate
performance management is based on identifying and tracking key metrics within the context of different business management functions. Politano led us through a couple of exercises based on
determining definitions and issues and then using analysis chaining to drive out the metrics. He then moved on to a master data management logical analysis exercise and then connected corporate
performance management (CPM) and master data management (MDM). Although the techniques that he showed us were designed to focus on the business, Politano showed us how CPM analysis techniques could
indeed help us identify metrics (facts) and MDM analysis could help us connect those facts to related dimensions.
As the week drew to a close and I reflected upon the conference, I was impressed by the number of emerging developments and the evidence of fresh thinking that I had witnessed. This went for
theory, technique, and technology. I think that experienced professionals in the field of data warehousing and BI will be making serious mistakes if they think that there is nothing more to learn
in our field.
Other Highlights
Arkady Maydanchik’s Monday keynote, The Quality Dilemma: Why Isn’t It Getting Better was thought-provoking. Maydanchik has extensive experience in the field of data quality,
and one of his observations has been that enterprises frequently make the mistake of believing that data quality issues will automatically go away once they migrate to ERP systems. Migrating poor
data guarantees that ERP systems will have poor data.
DataFlux hosted a mellow hospitality suite Monday evening. They had a great guitar duo, extensive tapas, and a tequila bar with a very knowledgeable host. I had a great time.
Michael Gonzales hosted a night school session, HandsOn Statistical Analysis for BI on Wednesday evening. Gonzales is a great teacher, and his session turned out to be a good review of
basic statistics within the context of BI. Gonzales has built the HandsOn brand up over the years and offers a number of classes for various techniques and/or technologies. Look him up next time.
Changes
After five years as Director of Education at TDWI, Dave Wells is stepping down and passing the baton to Paul Kautza. It has been my honor to watch Wells make data warehousing and BI more relevant
by continuous innovation in the courses offered by TDWI. Wells is talking about continuing to do some limited teaching, and my expectation is that he will continue to offer his insights at future
TDWI conferences.
Postscript
TDWI weeks are jammed packed with content. It’s always wise to schedule some down-time before or after the conference. Due our schedules, my friends and I decided to go kayaking prior to the
conference. We established a base at the Crystal River Resort in Crystal River, Florida. From there we explored the Chassahowitzka, Weeki Wachee, and Ocklawaha Rivers over the next three days.
The slow-moving creeks leading into the Chassahowitzka provided the classic Florida swamp experience. We paddled into back water springs, saw lots of birds, the spare ‘gator, and enjoyed some
quiet time away from the press of technology. The “Chass” proved to be a good warm-up.
The next day we headed up the Weeki Wachee against a fairly good current. However, it was worth it. We saw several manatees plus some new territory.
We floated down the Ocklawaha River next. This river turned out to be quite beautiful and very peaceful. Our outfitter told us to be on the lookout for Charlie, a 14-foot ‘gator, but he was a
no-show. We had to make do with a colony of ibis and a curious heron. It was a great way to finish up three days of pretty mellow rivers.
I have found covering TDWI World Conferences to be richly rewarding. I have always learned a lot and managed to work in some great side trips, all in the name of “color,” of course.
However, I am at one of those points in my career when I want to change the mix of what I do. I still plan on showing up at TDWI events now and then, so don’t be surprised if you do see me
there.
