True Universal Access: Fantasy or Modern Reality?

What's the real story on universal access? Even the so-called "casual" users can access a company's informational resources via their new friend, the Internet.

• 06/01/1999

The widespread adoption of Internet technologies by businesses today allows them to do something that the businesses of yesteryear could only dream about: Deliver content to users anywhere, anytime and in any number of different ways. Informational resources are no longer abstract and far away; they are available right now, at this very moment. More so now than ever before, people are connected to the world around them. Electronic mail has become the norm, and anyone with a phone line and a browser-equipped PC can instantly access close to a billion pages of information via the World Wide Web.

Within the scope of the business enterprise, the arrival of the Internet era dramatically changed the economics of data warehouse technologies. In the past, the use of such technologies was limited to a relatively small handful of power users who were charged with analyzing and poring over data to make important, far-reaching decisions. In fact, the weighty significance of these decisions used to be the sole justification for the great cost associated with maintaining and implementing the data warehouse to begin with.

Today the situation is somewhat different. The fusion of the Internet with powerful database technology has allowed an increasing number of "casual" users to access and use the company’s informational resources. This phenomenon has changed not only the profile of the "typical" user, but also how that typical user goes about getting the information that he or she needs.

It is an accepted fact that informational capital does not realize its true value until it becomes intellectual capital, that is, until it is absorbed by knowledge workers and put to use. Clearly, achieving successful deployment is of utmost importance within the data warehouse scheme. This basic truth is especially relevant today, given the rapidly changing user base to which the data warehouse must cater. Deployment is a fairly simple task when the intended audience is a small group of power users who are making similar, strategic decisions, but it becomes increasingly complex as that audience broadens to include literally thousands of workers.

"The initial focus of nearly all data warehouse projects is on providing access to a handful of power users, such as product managers or business analysts," says Richard Tanler of Information Advantage, a firm specializing in online information analysis and delivery tools. "While this alone is frequently enough to justify the investment in the data warehouse, it leaves untapped the value that the organization could realize if the same capabilities were extended to its thousands of other decision makers."

Targeting these previously untapped users – the personnel who are involved with the company’s "day-to-day" decisions – has dramatically changed the methods and goals of the deployment process. It is no longer enough to have a "functioning" data warehouse for the purpose of formulating broad strategic plans; rather, the successful data warehouse must be diverse, dynamic and it must be accessible all the time, from anywhere in the world.

As an industry buzzword, OLAP is perhaps too broad and too vague, but it is still the primary model for providing true universal access. This is hardly surprising, given the fact that the Internet itself is really nothing more than one gigantic data warehouse. Even if they aren’t familiar with the technicalities involved, most casual users are already familiar with the various URL formats, such as Hypertext Transfer Protocol (HTTP) and File Transfer Protocol (FTP).

At their most basic level, these protocols are really nothing more than a means of issuing pointers to specific resources residing somewhere within the Internet’s vast storehouse of information. With a single interface tool (the browser), any user in the world can issue these pointers and retrieve information.

Given the basic similarities between this model and that of the localized data warehouse, the integration of OLAP capabilities was, and continues to be, the next logical step.

Since OLAP is primarily concerned with providing an online bridge between a remote user and an active data warehouse, most of the problems and challenges presented by the technology aren’t much different from the challenges presented by most other online applications. For example, there are the persistent questions of network security and reliability, as well as the unique concerns these issues raise within the broader context of universal access. Providing both universal access and a high degree of security are sometimes incompatible goals, and the benefits of one must be weighed against the drawbacks of the other.

It is also accepted that security risks are a built-in feature of virtually all Internet applications, but in the case of data warehouses the problem is compounded by several factors. Besides the somewhat contradictory goal of providing broad, open access to many users, there is the problem of having to re-validate users each and every time they wish to perform some function.

Since the Web operates under a "stateless" protocol, meaning that the connection between client and server closes automatically if there is no activity, the OLAP architecture must provide a way for the data warehouse to "remember" that a particular request is coming from an authorized user. This problem is usually addressed with unique validation tokens, or "cookies," which get passed back and forth between client and server as needed.

But the need for security precautions becomes even more acute when the data being accessed is of an extremely sensitive or proprietary nature. In cases such as these, robust encryption schemes should be employed in conjunction with conventional tools like firewalls and password/cookie validation. A security approach that includes real-time encryption will not only thwart intruders, but eavesdroppers as well. In today’s world, you can never be too careful.

"Network security is, and always will be, a major concern," says John A. Robertson, Senior Computer Scientist at Boeing Information Services. "It’s gotten to the point now that it’s no longer a matter of if your data is being threatened, but merely how and by whom. If you’re housing or transferring valuable information, you can bet that somebody, somewhere is trying to get at it."

In the case of Boeing I.S., an organization that frequently performs contract work for government defense projects, that "valuable information" often takes the form of secret or highly classified data. But security concerns are just as vital and just as pressing in the private sector.

Information warfare and corporate espionage are unfortunate realities of our time, and the organization that ignores these dangers is almost certain to become a victim. Thus, the best security plan is one that is strong, proactive and preventive, yet still permeable enough for the data warehouse to remain useful.

Besides the increased security risks, perhaps the greatest challenge faced by OLAP technology is the requirement that it make the data warehouse available to many users, simultaneously. Again, this is a task that has grown more difficult as the user base of the typical data warehouse has expanded. It must be remembered that the OLAP architecture is much more than a simple gateway for accessing the warehouse’s contents; in many cases, it specifies entire sets of data objects from which the user’s requested report or information can be constructed.

Since all of the tasks associated with this construction occur on the server side of the transaction, it is important that the OLAP engine be designed in such a way that it will not slow down when "crunched" by many users at the same time.

An offshoot of the many-user problem is the effect that an expanded user base has on the size and scope of the data warehouse. According to Richard Tanler, extending access to "the masses" necessitates an expansion of the data warehouse along two dimensions: depth, to provide for a greater variety of analytical requirements, and breadth, to incorporate a greater number of subject areas.

Tanler points out that such an expansion can quickly enlarge the size of an average data warehouse to several hundred gigabytes, and frequently to terabytes or more. If not managed properly, data stores of this magnitude run the risk of crumbling under their own weight, so to speak, and they frequently require specialized, unconventional handling methods in order to realize their maximum potential.

Of course, from an OLAP standpoint, the greatest challenge of multi-terabyte data warehouses lies not so much in their size per se, but in their resistance to data extraction. The old "needle in a haystack" analogy is more than appropriate here. Since it is the job of the OLAP architecture to find and then transfer information from database to user, extracting and assembling the correct pieces of information (and doing it in real-time, no less) can be an extremely difficult task.

In response to this situation, increasingly sophisticated algorithms have been developed which allow users to "mine" or "drill" through layers of data with amazing efficiency. As these sorting algorithms continue to improve, so too will the ability of users to find and access precisely what they’re looking for. As long as the size of data warehouses continues to increase, such improvements will be necessary if true universal access is to stay a reality.

There can be little doubt that widespread online processing tools have revolutionized the way data warehouse technologies are used. Such tools have made the process of accessing information convenient and instantaneous, and they have made this access available to an unprecedented number of users. But even with these accomplishments in mind, the fact remains that many casual users – the very individuals who stand to benefit the most from these developments – have neither the time nor the motivation to actively seek out informational resources on their own.

In order for the data warehouse to be fully useful, the OLAP system must address this issue and integrate itself seamlessly into the user’s environment.

One partial solution to this problem lies in background processing, or "push" technology, which takes the search burden off of the user and places it on the system. Push technology enables the OLAP system to automatically deliver information to users on a need-to-know basis. With a push-based system, users get the benefit of increased access, but without the hassle and effort. Online subscription services, which automatically deliver information like news articles and special advertisements based on a subscriber’s interests, are crude examples of customized push technology at work. In some cases, where the delivered information is simply a small part of a larger ongoing application, the user isn’t even aware that any delivery is taking place.

Push technologies that require some minimal amount of effort on the part of the user, such as online search engines, have also contributed to the Internet’s usefulness as a means of accessing data warehouses. Search engines can be configured to search through data in any number of ways, but most often do so through a combination of simple keywords.

However, search engines lose a margin of their usefulness when applied to the kinds of dynamic content that OLAP is known for providing. For example, a user wouldn’t want to retrieve a static (and possibly outdated) report that had been catalogued by the search engine, but would instead want to retrieve a generic object template that included instructions for building the report using up-to-the-minute data. The report could then be assembled and delivered to the user as a standard Web document without he or she ever being the wiser.

The growing popularity of wireless technology also promises to further close the gap between user and information. In conjunction with OLAP technologies, the spread of cheap, reliable satellite communications has made it possible for users to access information even when there is no hard phone line available. Thanks to laptop computing and portable satellite uplinks, users can browse through informational resources at any time, day or night, whether they’re in the middle of the ocean or the middle of the Sahara desert.

Will the connectivity of the world’s information networks ever reach a point where there is no gap between user and information? A point where information is everywhere, alarm clocks can read e-mail and microwaves automatically download instructions to fix themselves when they break? Some on the fringe of information research think it’s possible, albeit unlikely.

Besides the potential technical problems inherent in a world saturated with connectivity, there are the practical considerations: For example, would anyone really want their microwave to be hooked up to the Internet? After all, just because something is possible does not mean that it will be useful. Fortunately for us, OLAP technology happens to be both.

About the Author: John Pitcock is Senior Manager for AmQUEST, Inc. (Atlanta), a managed operations provider. For further information, please call (404) 264-5641, or visit the company Web site at www.amquest.com.