UniForum '97 Keynotes Span Key Open Systems Topics

By Don Dugdale

 

Ruettgers Proclaims Importance of Enterprise Storage
Anyone who thought storage was boring can think again. Storage, ladies and gentlemen, is a happening thing and it's happening right now, in an enterprise near you. Before their very eyes, the attendees of UniForum '97 heard that message from storage guru Michael Ruettgers, president and CEO of storage vendor EMC Corp. in a March 12 keynote address.

In sometimes dramatic fashion, Ruettgers laid out the basics of today's storage world and EMC's central role in it. His key message: an intelligent storage solution is no longer peripheral but vital to the functioning of enterprise information systems.

In fact, it's the emphasis on that word information-including not only how information is stored but how it is accessed, that Ruettgers believes is the key to the future of enterprise storage. "I believe that management won't be able to dispense large volumes of critical information until it grapples with putting information first and technology second," he said. For stored information to be useful, he believes it must be consolidated into what he calls enterprise storage, where it is readily available throughout the organization. For support on the issue of consolidation, he cites an estimate from The Gartner Group that 70 percent of all storage will be consolidated in the data center by year 2000.

One information capability that Ruettgers feels will accelerate the adoption of consolidated storage is the free sharing of data generated on different CPU platforms, and he says solving it has been called the 'holy grail.' In a survey of IT executives, 87 percent favored shared storage over CPU-based storage, he noted. "If you don't have it on your calendar for 1997, you're already behind the curve because this is clearly where everybody wants to go. Just by its sheer growth, the storage factor looms large." Another indicator of storage's beefed-up role is the emergence of the position of "chief storage officer" in a number of companies, he said.

"Information sharing will allow users from one CPU environment to access and share data from other operating environments, utilizing the central storage system," he said, noting that EMC is "finalizing our next major step in this direction, and it will be soon."

I predict that enterprise storage will bring us to yet another computing model-the diskless server. And why not? For many applications, there is simply no need for you to be held hostage by local host-bound storage that can't be accessed or shared by other groups in the organization.

One step in the direction of enterprise storage, Ruettgers believes, would be to move internal storage out of the server. "I predict that enterprise storage will bring us to yet another computing model-the diskless server. And why not? For many applications, there is simply no need for you to be held hostage by local host-bound storage that can't be accessed or shared by other groups in the organization." And it's possible that the advent of diskless network computers may just make them "the perfect desktop accessory for an enterprise storage system," he added.

Ruettgers should know about such trends. In the past seven years Hopkinton, MA-based EMC has boosted its sales from $190 million in 1990 to $2.3 billion in 1996. In the process it has become not only the leader in external Unix system storage but also the source of over half the annual mainframe storage sales, surpassing IBM, which held 90 percent of that market in 1990. "EMC took over the lead in storage from a company that invented storage," he boasted. "I don't think this has happened before in any industry."

Sybase's Kertzman Warns Against Software Industry's 'Geopolitical War'
The American software industry is engaged in a geopolitical war that puts it in danger of repeating the mistakes of the U.S. automobile industry of the 1950s and 1960s. So said Mitchell Kertzman, president and CEO of Sybase, Inc., in his UniForum '97 keynote address.

By trying to sell customers a "magic bullet" every three to five years, the software industry is doing much the same thing the auto industry did when it believed people wanted to buy a new car that often, and so built cars that would fall apart after three years, Kertzman said. "It took the Japanese auto industry to teach American industry that what people wanted were cars that work and work for a long time."

Now, the software industry is "trying to encourage customers to throw out their software every few years and buy something else," Kertzman said. "I believe we are potentially vulnerable to that sort of thing."

For example, he said, the coming of the Internet is now being proclaimed a paradigm shift--one that dictates a shift to the thin client/fat server model. That's one magic bullet that doesn't necessarily need selling, Kertzman said. In the first place, the Internet does not qualify as a paradigm shift because paradigm shifts always take us by surprise and are accompanied by wealth and power changing hands, big companies falling by the wayside and smaller companies taking over. This time, he said, "The big companies are doing a good job of staying up with the technology." In addition, rather than taking us by surprise, the Internet is being actively promoted as a paradigm shift, Kertzman asserted.

It's all the promotion that's leading to industrial warfare, which is centered in the architecture of computing applications he said. "What causes us to be in a geopolitical war is the opportunity to dislodge Microsoft as the leader in the software industry." To the combatants, "This may be the time when the mighty stumble and the weak will inherit the industry." In this war, Netscape is viewed as the next Microsoft and has been thrust into the role of the giant killer. The rest of the industry is focusing on creating this so-called paradigm shift that will put Microsoft at a disadvantage.

"There should be no religion around application architecture. Centralization is better for some and distributed computing is better for some. We should not become involved in the geopolitical war and say one way is right and the other is wrong, but we should listen to customers and let them decide what is better for their situation."

Supposedly, the shift to thin client/fat server will bring Microsoft down because "Microsoft is wholly based on distributed computing power" and thin client/fat server architecture will recentralize computing power. The weakness is that not everyone needs this new architecture, Kertzman said. "There should be no religion around application architecture. Centralization is better for some and distributed computing is better for some. We should not become involved in the geopolitical war and say one way is right and the other is wrong, but we should listen to customers and let them decide what is better for their situation."

Kertzman, who took over the top job at Emeryville, CA-based Sybase after Sybase acquired the company he founded, Powersoft, said he has seen power shifts in the software industry and they have all been unpredictable. What worries him about the Internet is that it's being touted as the next mass medium after radio and television, but involves only an individual interacting with his computer screen. "It seems like a logical extension. But imagine browsing the Web with someone who has a mouse in his hand and the attention span of the average male. Imagine everyone coming home, eating dinner, and then all going off to surf the Net by themselves. I worry about that. I worry about whether that's the direction we're going."

Platinum's Filipowski Names Security as Top Issue
The need for a comprehensive security solution has emerged within the past six months as the number one problem facing managers of heterogeneous IT systems, UniForum '97 attendees were told in a CEO address on the show's opening day. The speaker was Andrew "Flip" Filipowski, chairman and CEO of Platinum Technology, Inc., who was quick to point out the basis for concern: no single, comprehensive security solution exists for heterogeneous environments and none is on the near horizon.

"We are at the Wordstar, or the VisiCalc level in solving our security issues," said the 46-year-old Chicago native, referring to early but now-faded stars in the word processing and spreadsheet software arenas. "We have very, very massive holes in our security, and for that reason, security is in the forefront of IT folks' minds today."

The security problem has been solved for homogeneous environments because that's a much simpler problem than the one facing heterogeneous systems, Filipowski said, adding that a single vendor will need to provide the end-to-end security solution now demanded. He said Platinum is working on that solution, but that it's at least two years away. "We can provide enough security to prevent attacks on some of the easier targets, but we can't plug every hole every time," he said.

"You can lock up the Unix kernel reasonably well, but you also need a Web security strategy. Can you implement a solution that keeps out the intruder but doesn't make it too difficult for the legitimate user? You can't, because the solution is not available."

Some of the pieces of the security puzzle have been put in place, he noted: "You can lock up the Unix kernel reasonably well, but you also need a Web security strategy. Can you implement a solution that keeps out the intruder but doesn't make it too difficult for the legitimate user? You can't, because the solution is not available." To get to that point, authentication methods need to be refined, he said, because current methods such as thumbprints and retinal scans are off the mark. "People may not necessarily want to put their eyeballs up to a device where it's not going to necessarily be guaranteed that you'll have eyesight for the rest of your life," he said. Voice recognition or "smart cards" coupled with personal passwords are more likely to work as authentication devices, he added.

Filipowski put the heterogeneous security problem in the context of the phenomenon of "disappearing" software companies, pointing out that only an established company that is perceived to have a future will be able to inspire the confidence necessary to sell its product. And Filipowski should know-he has been in the forefront of company acquisitions for Platinum. Within a five-month period in 1995, Platinum gobbled up 10 smaller companies-part of a plan by the CEO to put together a unified series of systems management products across the computing spectrum from PC to mainframes.

Typically, a company's life cycle works like this, Filipowski said: A company starts with a product idea, gets its employees very excited, works very hard to produce the product and gets it to the marketplace. If it's a good product and timely, the produce succeeds, and then the company becomes identified with the product. Then the company gets into a race for survival. Unless the company can continuously update the product, as well as expand its product offerings, the company and the product will be acquired by a "bigger fish" or the product itself may die from neglect. In that respect, Filipowski said, software is a lot like the music business. "There are only six big music distributors. There will always be startups, but there are no medium-sized companies in the music business. In software, there's only going to be five or six of us [large companies] by the end of the decade. Every medium-sized and smaller software company is doomed to extinction."

For that reason, it's the big companies that will rule in heterogeneous security. As an IT manager, he said, confidence in the software vendor becomes more important to you than the product itself. "You end up having to decide whether you have confidence in the company." To emphasize the point, he asked the audience to name the largest independent software vendor of 1985. Only three were able to identify Cullinet Software (a former employer of Filipowski) as the company. "A dozen years from now, who will remember Microsoft?" he asked, to giggles from the audience. "I'm not saying we'll forget Microsoft. Chances are we're not going to do that. But there is this cycle. There is an explosive growth of individual point solutions, then consolidation. You can learn so much from history."

 

Back | Table of Contents | Next