The year ahead ... in technology
From business intelligence to wireless, progress abounds but work remains
- By Brian Robinson, John Moore
- Jan 09, 2005
Some of the technologies that grabbed information technology managers' attention 12 months ago still attract their interest in 2005. Only the focus will change as technologies such as business intelligence, open-source software, Web services and wireless mature and users grow more comfortable with them.
Support for open-source projects by major companies such as IBM and Hewlett-Packard, along with government officials' growing use of Linux, could help spur real solutions this year. "Interoperability" will be the operative word for Web services and wireless technologies as developers seek to make wireless devices talk to one another and Web services components communicate faster.
Security will still top many agency officials' priorities list with an emphasis on tightening defenses on all systems connected either directly or remotely — known as endpoint security — to the corporate local-area network (LAN).
Still, as some of the same technologies continue to demand attention, there is room for the unconventional, such as using the Internet as a backbone for enterprise wide-area network (WAN) traffic.
This article offers a snapshot of how the focus on some technologies is changing and what obstacles still need to be addressed.
Business intelligence products have acquired an elitist reputation because only expert users could figure them out.
The tools that constitute business intelligence include products for gathering business data, analyzing that data and churning out reports. The idea is to search for patterns and trends in an organization's data that can bolster the decision-making process.
Business intelligence products have taken root among IT departments and power users. But attempts to serve a broader audience have been mixed at best, according to industry watchers.
The issue: Business and casual users have not been inclined to learn the nuances of sophisticated business intelligence software.
"Our clients are saying, 'We want more' " business intelligence, said Keith Gile, a principal analyst at Forrester Research who follows business intelligence. "The problem is they don't have the time to become [business intelligence] power users."
In short, users want the results of business intelligence analysis but not the learning curve.
Help may be on the way. Gile said vendors have begun to heed customer concerns. The upshot is a business intelligence transformation that's beginning to unfold.
Gile said business intelligence will shift from being a distinct solution set to a capability embedded in a variety of applications. "Customers ...don't need to know that they are doing" business intelligence, he said. "They want it as a complement to the job they are already doing."
In 2005, business intelligence will be pushed to the operational level, Gile said. He cited call center operations as one area that might see an embedded business intelligence capability.
"You're going to have [business intelligence] everywhere," said Neil Raden, founder of Hired Brains, which conducts market research for the business intelligence industry. "The best [business intelligence] systems are the ones you don't know are there."
But that vision of business intelligence will take years to fully realize, Raden said. For embedded business intelligence to happen, the major vendors will have to rebuild their products as components within a service-oriented architecture. Vendors such as Business Objects and Cognos have begun the process, "turning themselves inside out," Raden said.
Cognos' ReportNet product, designed to serve the needs of a broad range of users, already features a Web services architecture. Company officials' task now is to "re-engineer all of their products to be in the same architectural mode as ReportNet," Gile said.
The radical architectural shift now under way will be far from easy. But a successful transition could dramatically enlarge the business intelligence market. "Once the components are embedded with and in line with what users are doing, [business intelligence] becomes rather invisible yet indispensable," Gile said.
No matter how good IT managers think their enterprise security is, it's only as good as the least secure system on the network — a fact that should make endpoint security a hot topic in 2005.
Endpoint security solutions ensure that systems attached to the corporate LAN, whether directly or remotely, have the appropriate security patches and antivirus software and that host firewalls and other devices are configured correctly before systems are allowed to connect to the network.
Pete Lindstrom, research director at Spire Security, said endpoint solutions will also give enterprise managers a greater degree of flexibility in managing their security.
"You can isolate [the endpoint system], but you don't necessarily have to take it off the wire," he said.
That's been a problem with security solutions such as intrusion prevention, he said, which look for and block problems.
But the effect is that users, who may not be aware that their systems are compromised, flood the help desk with messages saying their computers stopped working. With endpoint security, Lindstrom said, a popup screen can tell users the system has been compromised and offer suggestions for fixing it.
Several major industry projects are trying different approaches to endpoint security. Cisco Systems officials announced the Network Admission Control (NAC) program, and Microsoft came out with Network Access Protection (NAP). Officials at well-known security vendors have begun attaching themselves to one program or the other as a way of ensuring that their devices and solutions will interoperate with a range of other endpoint security products.
Trusted Computing Group (TCG) officials are also working on an open, multivendor industry specification for endpoint security called Trusted Network Connect. It could be available by the second half of 2005.
Burton Group analyst Phil Schacter thinks that product is the leading candidate for an industrywide solution because the Cisco and Microsoft programs may not be broadly deployed for another three or four years.
"We hope soon to have some form of marketing schema, as well as a program for compliance testing" of endpoint security products, said Anne Price, a TCG spokeswoman.
In addition to contributing to overall security, the impact endpoint security will have on patch management regimes should be enough to push it to the forefront in the coming year, Lindstrom said.
Enterprises must provide security patches to systems when a threat emerges. With endpoint security, however, they'll be able to isolate and repair only those systems that need attention with just-in-time security, he said.
The Internet as a WAN backbone
The public Internet has become a core part of modern culture, but it's never been good enough for the kind of backbone communications major organizations need — until now, perhaps.
Eric Siegel, senior analyst for network and telecommunications strategies at the Burton Group, thinks performance improvements of the Internet make it a viable option organizations could use for backbone WAN traffic.
"Just five years or so ago, the Internet still had significant performance problems, with around 5 percent IP packet loss," Siegel said. "Now the measured packet loss is less than 0.4 percent." Also, average roundtrip latency for data packets moving via the Internet across the United States has dropped to about 55 milliseconds, he said, which compares favorably with the approximately 50 millisecond latency of leased-line communications.
And the availability of the Internet — the average time it's available vs. down during any given period — is now better than 99.5 percent, Siegel said. What problems remain, such as security and provisioning, are more perception than reality, he said.
Ben Rushlow, consulting manager at Keynote Systems, a leader in Internet performance measurement, said reliability of the public Internet has hit 100 percent in some months. But he questioned whether the Internet could match corporate requirements for backbone WANs, which are typically used for connecting data centers, for instance.
"Costs are much lower when using the Internet, for sure, but what about security," which requires data to be encrypted on the public Internet? he asked. "When you have a secure [leased-line] WAN, you don't have to encrypt things, and that opens up performance options, such as being able to do database synchronizations and high-end [input/output] operations."
The encryption requirement would slow network performance, he said. Also, using the Internet to meet service-level agreements for corporate customers would be difficult for IT managers because they don't own the network links, he said.
Nevertheless, Siegel said Burton officials are convinced the Internet is at the point where organizations could use it for many WAN requirements, and company employees have begun advising clients to consider using it for WAN backbones, at least in the United States.
Open source moves into the mainstream
The next few years could finally see Linux, and open-source software in general, pass into the mainstream of commercial and government IT markets.
It's not that the drivers — lower costs plus the ability to mix off-the-shelf systems and applications with existing systems — have changed much, said Bill Weinberg, architecture specialist at Open Source Development Labs (OSDL). But the inhibitors are starting to fall away.
"There's likely to be a resolution of the [SCO Group's Linux copyright infringement] case against IBM and others, and that will liberate those [who] have been reluctant to disclose their Linux use," he said. "It will put a whole tier of potential adopters at ease."
SCO Group officials filed a multibillion-dollar lawsuit in March 2003 alleging that IBM had unlawfully contributed SCO Unix code to open-source Linux, and followed it up with claims for license fees from other companies.
The independent software community now also seems to be coalescing in its support of Linux, Weinberg said. OSDL has been building a group of developers to more closely examine their requirements, and Weinberg said independent software vendors' plans for Linux would become much clearer this year.
"Once we get end users and [independent software vendors] more involved, we can begin to move from where the issue is Linux itself to actual solutions," he said.
Users have also been convinced by the growing professionalism of open-source software, said Richard Soley, chief executive of the Object Management Group, which develops standards and specifications in support of open-source software development.
"The primary reason for that is the network effect — the fact that we are past the early adopters [of open-source software] and enough people are using it," he said.
Along with the support of major companies such as IBM and HP, people can see there is a more serious approach to open-source projects, which is backed by the high quality of documentation and training and the availability of consultants knowledgeable about open source, Soley said. That could boost Linux's popularity, particularly in government.
Surprisingly, said Tony Stanco, executive director of the Center of Open Source and Government, about 30 percent of federal government servers already appear to be Linux-based.
"One-third of Web servers with Linux is not enough, and that share could still slip," he said. "But the move from Unix to Linux is a natural move, and I think that migration will push Linux adoption to 50 percent."
Getting another 16 percent or so of the market could take some time and effort, he said, but if that happens and Linux gets to two-thirds of the server market, "the rest drops into your lap," Stanco said. "Microsoft is fighting this hard, and they are always formidable. But my guess right now is that Linux will win."
Changing focus of Web services
Web services aren't exactly old hat, but the technology is riding higher on the maturation curve.
Fundamental concerns about interoperability and security have eased in recent months. The Web Services Interoperability Organization's WS-I Basic Profile provides some assurance that Web services built with different tools will work together.
The Web Services Security standard, ratified in April 2004, also boosts confidence in the technology. The focus of those guiding Web services' evolution is now shifting to other areas.
Brand Niemann, a computer scientist at the Environmental Protection Agency, cited semantic interoperability as one such field. Earlier interoperability work was "all about making applications talk to each other," he said. "But now we are at the stage where we want to do information-to-information Web services."
That's where semantic interoperability comes in. "The goal is not just to connect systems but also to make the data and information resident within these systems interoperable," according to a white paper published by the CIO Council's Semantic Interoperability Community of Practice. Niemann is co-chairman of that group.
The implication for semantic interoperability is that Web services can be orchestrated in something closer to real time. Today, Web services may stumble over the differences in how organizations define the same term. "It can take as much as 20 minutes to harmonize your concept with the other," Niemann said.
The faster linkage of Web services components is important for health and emergency applications. The EPA and the Centers for Disease Control and Prevention, for example, are conducting pilot tests of semantic interoperability technology to link information related to environmental toxins.
This year may see further exploration of semantic interoperability. The World Wide Web Consortium (W3C) is advancing Web Ontology Language (OWL) and the Resource Description Framework (RDF), both of which became W3C recommendations in 2004.
OWL and RDF provide standard formats for data sharing, according to W3C officials. The Semantic Web Applications for National Security conference, which will be held in February, is intended to further stimulate interest in semantic interoperability, Niemann said. Among its objectives is to spur projects employing OWL and RDF. The Defense Advanced Research Projects Agency is sponsoring the event.
Web services' semantic integration push is evidence of increasing maturity. Another signal comes from the world of systems integration.
Professional services companies are gearing up for a significant year in Web services, said Ronald Schmelzer, senior analyst at ZapThink, a market research firm focusing on Web services. He said consultants and integrators are increasing their skills to build Web services and service-oriented architectures.
Generally, integrators won't plunge into a particular technology until they are confident it will be widely deployed.
Progress and obstacles to wireless
Wireless interoperability isn't the oxymoron it used to be.
Indeed, no one disputes that progress has been made in making wireless devices work together. In October 2004, Cellular Telecommunications and Internet Association members announced an agreement among wireless carriers to provide interoperability for Multimedia Messaging Services. MMS interoperability, which covers photo and video messaging services, is expected to be available this year.
But industry watchers say plenty of work must still be done. J. Gerry Purdy, principal analyst at MobileTrax, underscored that point during a recent talk at Stanford University. He listed interoperability as the top challenge facing the wireless industry.
Purdy said one issue is making wireless phones and handheld computers work regardless of where they are used worldwide. Today, quad-band phones rely on multiple radios so people can use them in the United States, Europe and Asia, he said. That's only the beginning, however.
Code Division Multiple Access and Global System for Mobile Communication standards, however,
are not available in the same phone, he said.
The WLAN arena shows a similar pattern of improvement and lingering obstacles. Richard Alena, a computer engineer at NASA's Ames Research Center, called WLAN technology reasonably mature. Interoperability, however, may break down when deployments push the technical envelope.
Alena, group lead of Ames' Intelligent Mobile Technology team, conducts interoperability tests in the organization's lab. He said he has discovered that most products work well in a "nominal mode that supports 90 percent of users." Beyond this area, however, various combinations of products may not work together.
Interoperability can be particularly troublesome in specific applications such as public safety. First responder radio systems often fail to communicate across jurisdictions, a situation the Sept. 11, 2001, attacks thrust into the spotlight.
Last spring, Homeland Security Department officials issued a statement of requirements addressing interoperability. Some movement on that front is expected this year. In the meantime, industry and government continue to experiment with such approaches as mobile ad hoc networks and mobile IP routing.
Progress is slow, however. One obstacle: Public safety officials are leery of IP-based solutions in mission-critical situations. "They are concerned about latency," said Leonard Miller, an electronics engineer in the National Institute of Standards and Technology's Wireless Communications Technologies Group.
Public safety users are particularly concerned about mission-critical voice communication, added Nader Moayeri, manager of the group.
Miller said those concerns will diminish as public safety users learn more about IP.
Moore is a freelance writer based in Syracuse, N.Y.