I’m disappointed. Of all the questions and comments that I’ve received from this column so far, no one has asked the one question that I expected to be faced with right away. Hasn’t anyone wondered what a column on pervasive computing is doing in a publication devoted to business intelligence (BI) and analytics?
On the surface, the promise of pervasive computing seems rather remote from business intelligence. Wearable computers with circuitry woven into their very fabric may offer significant potential to portable technologies such as global system for mobile communication (GSM) phones and MP3 players but seems remote from business intelligence. Similarly, active spaces that detect someone’s arrival or proximity at a location, whether it be their home, office or an airport, have tremendous potential for safety, comfort and entertainment but again seem equally distant from BI.
Even the challenges of pervasive computing seem to be of a purely engineering nature. Before wearable computers make it to prime time, significant advances in power management and miniaturization need to be made. Realizing the goal of active spaces requires significant leaps in artificial intelligence so that computing devices can better sense and perceive the physical world. And ultimately the prerequisite keystone of pervasive computing a ubiquitous, wireless, always-on internetworked world needs to be achieved.
If your area of focus sits squarely in the traditional business intelligence space, why would you care about pervasive computing? The answer is because pervasive computing is changing the world we live in. It not only will change the world of BI (its influence is already being felt), but pervasive computing will indirectly drive BI developments, because BI is a critical building block in the practical application of pervasive computing.
Pervasive computing and business intelligence are linked in perhaps not-so-obvious but very fundamental ways. It is a push-me, pull-you relationship with pervasive computing pushing BI along at times while other times pulling BI directly into the pervasive computing systems themselves.
Pulling Business Intelligence into Pervasive Computing
The Star Trek-like engineering feats that enable pervasive computing may grab the headlines, but underneath the surface, these systems are dependent on an underlying data stream of information that continually adjusts and adapts to deliver relevant intelligence in the current user’s context. If this information is delivered successfully in context when it is needed, it transforms from being raw data into intelligence.
An example of this that is on its way to becoming pervasive in the marketplace is the navigational systems that are appearing in more and more car models every year. What started out as a GPS with a rudimentary geographic information system (GIS) is fast evolving in a number of directions. General Motor’s OnStar system is a prime example. The OnStar system extends the typical navigation system with an "information/convenience service." This service utilizes OnStar’s built-in cellular phone to link the driver to OnStar’s customer service center. The customer service center knows the real-time location of the automobile calling in and can cross-reference that location with their data warehouse of places of interest, hotels, restaurants, etc. Want to find the closest ATM, the nearest hotel of your favorite chain or directions to the airport from your present location? OnStar’s underlying data warehouse contains that information and provides it to you in real time.
Pushing BI into the Real-Time Digital World
Business intelligence may be a component of most pervasive computing systems, but BI is also being driven by pervasive computing. Pervasive computing is challenging BI systems with their requirements for real-time data access and access from a variety of interfaces. Over the past few years BI systems have moved to a Web interface or to a Web-portal interface. Now, they will be challenged even further as BI systems are pushed down to mobile devices, handhelds, tablet PCs and cellular phones. If you need real-time information, these alternative devices become a necessary interface because you cannot guarantee that the user will always be sitting in front of a networked PC.
While the user interface will change, the real driver is the need for real-time information. In all business intelligence implementations I’ve worked on in the past, the business requirement for information latency starts out as zero. Business users always want the most current and up to date information. Now in the past, the cost and ROI associated with it did not justify real-time BI. The difficulty associated with zero or near zero latency made daily, weekly or even monthly updates of BI systems the standard.
Now BI is being challenged on two simultaneous fronts. Real-time enterprise integration architectures are not only possible but are proving themselves in corporations around the world. Business intelligence systems are being integrated into these architectures. Business intelligence, primarily ETL, vendors have developed and are now selling connectors into these enterprise integration systems. Concurrent with the implementation of these real-time architectures is the user’s familiarity with real-time information. Even if the need is not there from an ROI perspective, there is an increasing perception that the information in BI systems should be available in real time.
From Enterprise Data Warehousing to Global Data Warehousing
Enterprise data warehouses have always been pushing the envelope in the very large database (VLDB) space. An enterprise data warehouse for a large Fortune 1000 firm will easily be in the hundreds of gigabytes with terabyte-size databases becoming more common. For those of you who believe, you are on the edge with your VLDB data warehouse, I hate to tell, but you ain’t seen nothing yet.
For retail systems, the UPC bar code has been pervasive for years. It exists on virtually every retail product. These UPC bar codes are the front runners of what is referred to in pervasive computing as coded tags. In order for smart spaces or mobile systems to be able to perceive their environment, the physical environment must be encoded. The next generation Internet protocol, IPv6, with its address size moving from 32-bit to 128-bit has enough naming and address capacity to uniquely identify every square meter of the earth’s surface. How big will your data warehouse be if it needs to contain the entire physical world?
Much of this technology exists today. The largest issues surrounding its adoption are privacy issues, not the underlying technology. However, real-life examples already exist. Proximity cards are beginning to be adopted as an alternative to having personnel remember usernames and passwords. A proximity card is worn by a user and uniquely identifies that user to a PC by transmitting an encrypted code as the user approaches the PC. The PC is adapted with a radio receiver and uses the transmitted code to authenticate and authorize the user. Later, if the user gets up from the PC, as the PC loses the radio signal, the user is automatically logged out. How’s that for single sign on?
Maybe not Today, but Tomorrow is Right around the Corner
For those of you who are not interested in pervasive computing right now, you probably do have some time. However, if you get a call from one of your business clients who say they need a real-time business activity monitoring (BAM) system, and maybe, just maybe, the system needs automated publishing of alerts to a variety of devices, take a moment to pause and reflect. Pervasive BI may be closer than you think.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access