Network Management: The Only Constant Is Change
Network Management: The Only Constant Is Change<@VM>What's on the Horizon?
By John Makulowich
Sit still for just a Web minute, idly pausing to refresh, and a subtle wave of change will pass over you. It is the rush of the Beta, the continual movement that masks itself as progress on the Internet. Nowhere is it more evident than network management.
A recent discussion had just begun about the issues and challenges in network management that federal chief information officers are likely to face, when Ed DeSeve mentioned President Clinton's Dec. 17 directive on electronic government.
DeSeve is partner and national industry director for KPMG LLP's federal practice in Washington. He was referring to Clinton's initiative "to help our citizens gain one-stop access to existing government information and services and to provide better, more efficient government services and increased government accountability to its citizens."
The directive outlined 11 points designed to achieve those goals. One of the most important items in the directive called upon the administrator of General Services, the National Partnership for Reinventing Government, the Chief Information Officers' Council and the Government Information Technology Services Board, among others, to promote access to government information not by agency, but by type of service or information that people seek.
In other words, identify and organize data to make it easier for the public to find.
Another item informed the heads of executive departments and agencies to make available online by December 2000 the forms needed for the top 500 federal government services used by the public.
On one key level, this directive shapes the agenda for network management by focusing on the end user and customer service, the Web's Holy Grail.
More importantly, it moves to the front of CIOs' minds the issues that already were bubbling below the surface, such as security, customer access and intelligent portals.
For DeSeve, who served as deputy director for the Office of Management and Budget and headed several interagency task forces in this administration, including the President's Management Council, Chief Financial Officers' Council and Chief Information Officers' Council, the No. 1 issue is internal security.
"The reason it is No. 1 is so we can begin having very customer-oriented and very publicly available external transactions. The challenge is to put that level of security in place," DeSeve said.
Another major challenge, in line with the presidential directive, is to open up federal systems to customer access. And that means figuring out the terms and conditions of opening the network to outside parties.
"Major questions arise here, for example, what level of encryption is needed? Is it PKI or a PIN-based ID system?" DeSeve said. "And with transaction-based access for individuals, the dangers [of system compromise] are magnified. What is called for are rules, regulations, terms and conditions through which the government will allow external parties to engage in transactions."
To comply with the president's directive, agencies must build intelligent portals, he said. DeSeve also noted that the president said to comply fully, it is not enough to get a form from an agency or send it back. A person accessing the government Web site must be able to engage in transactions.
Thus, the administration is beginning to fashion a specification of the configuration to which any given federal Web must comply, where transactions-based systems are available for use and properly secured. In essence, that is a Web where individuals can transact with any agency and have the information sent where it needs to go.
For the federal CIO and Web master, it amounts to working on system architectures and Web designs that work well within searchable portals.
Looking beyond the president's directive, security remains the primary issue in the arena of network management, at least to another member of the Big 5 professional services firms.
Chell Smith, global leader of critical technologies for Ernst & Young LLP, said security is paramount, especially as organizations open their networks to the outside world, understand who is coming in and what information is being made available.
"But as much as we talk about technology and tough decisions, the bigger challenges for CIOs are from a budgeting and planning perspective. For large organizations, IT represents massive investments. And with these kinds of [network] infrastructure investments, it is hard to do a traditional business case," said Smith.
With IT budgets under scrutiny, Smith's point is that decisions about network management center on financial factors as much as technical ones. She cited the intranet as an example, a key network management initiative for many organizations in every economic sector.
The question revolves around whether you can get a cost-effective solution. It is also one reason why the major players place so much emphasis on total cost of ownership.
Yet another fly in the ointment to coming up with reasonable approaches to the problems of network management is that it is difficult to determine the cycles of technology and the refresh costs. Typically justified on five- to 10-year horizons, this is no longer a realistic time frame. The question remains: How do you justify these expenditures before someone comes up with something else that replaces the technology you just bought?
"As they tackle new technology and implementation, CIOs will have major concerns about reliability and performance, about the right time to make a decision in the maturation cycle to meet and exceed the expectations of the constituency," Smith said.
Exploring beneath the surface, to what she termed an "under-the-waterline" issue, is the need for organizational skills to manage the transitions mentioned by Smith.
From Smith's point of view, the dilemma that emerges must look like an iceberg to a CIO. Outsource by hiring someone to design and implement these technologies and give up internal control. Significantly upgrade staff skills at a significant cost and introduce the risk they will jump ship to the highest bidder for their skills.
On the other side of the fence, in the backyard of a company that focuses on network management, you find Computer Based Systems Inc. of Fairfax, Va., the wholly owned subsidiary of AverStar Inc. The Burlington, Mass., parent company provides software, information assurance services and technology solutions for government and commercial customers.
The 600-employee CBSI focuses on data center management, network operations support and enterprise application development services. Among its list of credits, CBSI built and manages PTO Net, the U.S. Patent and Trademark Office network that serves 6,000 agency employees in 17 locations. Among AverStar's main clients are the Bureau of Land Management, the Environmental Protection Agency, the Health Care Finance Administration, the Immigration and Naturalization Service, NASA, the Navy and the Postal Service.
According to Richard Zins, president of CBSI and senior vice president at AverStar, one of the key issues for federal CIOs is how to handle network management itself. It no longer stands alone but has become more an issue of infrastructure management and includes other computer and computing assets, he said.
Another key issue, Zins said, is outsourcing, contracting out the services connected with managing and maintaining the network.
"We are well-acquainted with GoCo, or government-owned and contractor operated," Zins said. "The next step may be having the contractor own the assets, so called CoCo, or contractor-owned, contractor-operated."
While he said issues of privacy and security would arise in any such arrangement, Zins did not believe they were insurmountable and pointed out that each agency would have different demands for systems security, whether it be the Department of Defense, the Federal Drug Administration or the Patent and Trademark Office.
Another key player in the federal government network management space is Computer Associates International Inc., Islandia, N.Y., with its network and enterprise management products.
Computer Associates just announced that the commander in chief of U.S. Naval Forces, Europe, the operational commander of all U.S. naval forces in Europe, chose the company's Unicenter TNG to manage centrally its widely distributed IT enterprise.
The Unicenter TNG will allow the IT staff at central headquarters in London to capture data from all European naval stations, track the location of IT assets and determine how they are used. The varied IT infrastructure includes thousands of decentralized servers, workstations and desktops, and multioperating systems and applications, such as Windows NT, Sun Solaris, Unix, Novell, Oracle, Cisco and Citrix technologies.
According to Trevor Kemp, Computer Associates' business manager for network and Internet management, while all the issues cited are important, a defining issue is the e-business infrastructure and the demands placed on it.
"One of the dilemmas faced by federal CIOs will be how to use the range of services available, whether VPN- or ISP-type services, and how to start managing them," Kemp said. "From a technical perspective, the question is how to effectively use that bandwidth, a precious resource we never seem to have enough of."
Kemp agreed with other officials that a second key issue is security. This means controlling access not only for external users but also for e-business users through open environments. This will demand a complete umbrella of securities, a complete web of measures to coordinate all aspects of security.
"This brings us to a wider issue: How do we segment management in this environment? There is system management, network management, database management. The majority of company applications, databases and systems are interlinked. No one really acknowledges the importance of this," Kemp said.
What he sees is a trend toward bringing together the network that connects all these management functions. For example, Web sites need management in performance, response times and page hits. These activities must be viewed from an overall service perspective instead of individual functions. By John Makulowich
For those who can tolerate a moment to stop obsessing on the present and the blinding speed of change on the Internet, a noteworthy resting point would be a recent report produced by the National Institute of Standards and Technology.
Apropos that federal laboratory, a unit of the Department of Commerce, the report is titled, "ITL [Information Technology Laboratory] Strategic Planning Input from the Advanced Network Technologies Division [ANTD]," and authored by seven researchers.
The importance of this report stems not only from the central role that NIST plays in developing standards in partnership with industry, but also from the areas identified in its research horizons.
The report opens by noting that economic forecasts project spending by Internet service providers on hardware and software will grow to $11.6 billion in 2002. And by that year, 57 percent of this growth will be due to sales abroad.
The researchers recommend that the Information Technology Laboratory pursue two strategic directions to anticipate the changes ahead in networking: so-called agile networking infrastructures and programmable network services. If you were a betting person and had to wager on the future of networking, you could do worse than follow the advice of these experts.
According to Kevin Mills, director of the Advanced Network Technologies Division and co-author of the report, the strategic plan looks at the networking industry in three time periods, typical for a research organization.
The first covers the next four years, in which industry continues to focus on media convergence and securing the network. One reason for the long lead time is that the industry's attempt to superimpose video, voice and data on the Internet is constrained, because the Internet was not designed to carry that type of load. This opens up issues and challenges in dealing with bandwidth reservation and congestion.
And since commerce is moving steadily and rapidly to the Internet, the need will be intense to ensure the infrastructure is secure. This is likely to stimulate continuing development of virtual private networks and subnets.
In the second time horizon, about five to seven years, Mills and his colleagues see industry working on increasing the agility of the network, accommodating wireless and the range of devices that people likely will be carrying. We also will see the advent of pervasive computing, where computers and sensors are embedded in buildings and clothing.
With the range and number of devices, there will be demand for dynamic discovery by the network as well as automatic configuration.
As Mills noted: "Security will have to change. We will see third-generation cellular phones and devices that can discover the network and adapt how their software works to what they find."
In the third time horizon, eight plus years out, speculation is that we are likely to see applications that can reprogram the internals of the network to make themselves work better.