Building Year 2000 Interfaces
- By Tracy Mayor
- Nov 30, 1997
You've educated your managers, drawn up your methodology. You've inventoried your systems, assigned project leaders and started code remediation. In short, you're well on your way to fixing your Year 2000 problems before the Big Deadline, and if some other information technology types haven't touched a line of code yet, well, it's not your problem.
Or maybe it is. What if even one of those procrastinating, understaffed, disorganized departments had the power to mess up your well-laid plans and pristine databases? Welcome to the second phase of Year 2000, affectionately known as "The Interface Problem."
Demands to downsize government, as well as to automate and coordinate the delivery of services, led to an explosion of data sharing among local, state and federal agencies in the 1990s. Now those interconnections are coming back to haunt already overburdened IT departments. In addition to ridding their own programs of bad code, government organizations must now somehow ensure that all digital data accepted into their systems from outside sources is not only fixed as well, but fixed in the same-or at least a compatible-way.
"It's like the [United Parcel Service] strike. Suppliers were disrupted, then customers were disrupted, then customers' customers were disrupted," said Carolyn Purcell, executive director of the Texas Department of Information Resources. "If an interface fails to perform or fails to perform correctly, we're all going to have to do some triage."
Data interfaces that are out of sync could stop a system in its tracks or, worse, return incorrect or incomplete data without notification. For example, unrepaired data fields could fail to flag certain suspects as juveniles, causing them to be booked and processed through the justice system as adult offenders.
Such connections also could cause insidious damage downstream. "What happens if a state trooper stops somebody on the highway?" asked Larry Olson, chief information officer for the state of Pennsylvania and a leader in coordinating states' Year 2000 efforts. "He does a check through the state system, which is clean, but if the information doesn't go any further-to the FBI and other states-we're still not in compliance."
Worse, in the eyes of some IT managers, an organization's already cleansed and certified databases could become compromised all over again.
"If you start to allow data that is not corrected into large databases, you'll have the hugely expensive effort of weeding out wrongly calculated data," said Steve Kolodney, director of the Washington State Department of Information Services and the chairman of the National Association of State Information Resource Executives (NASIRE) Year 2000 committee. "And if there's downtime, and you have to revert to manual processes, you lose confidence there as well."
Even when an interface simply fails, workers on both sides will be left scrambling to identify and fix the offending fields while simultaneously performing workarounds to accommodate downed systems. In Texas, one state agency already experienced a portent of what might come, when a federal agency transmitted data in the new four-digit-year format without alerting state workers to the change. "The state spent several days trying to repair the problem," Purcell said. "If you multiply that by hundreds of thousands of similar problems across the nation, the scope gets to be huge."
Whose Problem Is It?
Compounding an already messy situation, a meta-debate is taking place over where the bulk of these interfaces lie and where resources and attention should be allocated: between the federal and state governments, between states and local organizations or equally among all three?
State CIOs are preoccupied primarily with the progress-or lack of progress-from key federal agencies, so much so that in October, NASIRE and Pennsylvania Gov. Tom Ridge sponsored a one-day data-interface summit attended by staff from 21 federal agencies and all but nine states. "Our concerns are less with local than with federal government," Kolodney said. "Federal agencies, particularly [the departments of] Health and Human Services and Transportation, are lagging."
"Is the federal government further behind than the states?" asked Lou Marcoccio, a research director with Gartner Group, Stamford, Conn., who specializes in Year 2000. "That depends on which agency you look at." Of 24 federal agencies monitored, one-third have reached what Gartner calls Level 2 in their Year 2000 efforts-that is, they have taken a detailed inventory of all internal and external software, systems and hardware and have conducted a detailed impact analysis. Three agencies have reached Level 3, where they've divided work into hourly tasks, identified all needed resources and gained approval for those, and audited outside vendors on their progress.
By comparison, Marcoccio said, about 50 percent of state governments are at Level 2 or 3, with the remaining states running behind.
Kolodney is among a group of state leaders who are more sanguine about making intra-state regional repairs. In Washington state, fixing shared state-administered systems, such as welfare administration or criminal justice, will automatically adjust systems downstream. "In states like ours, once you've fixed it, it's fixed all the way down," Kolodney said. "We have autonomous local agencies, but they're using a statewide system."
But not every state is in a similar position, and some government watchers argue that more attention needs to be paid to state and local government linkages. "If you talk to the state CIOs, most don't see [local involvement] as a big issue," said Bob Greeves, a Vienna, Va., consultant on intergovernment issues. "But I don't get that same impression when I'm talking to associations that represent local governments."
Greeves suggested three ways local governments could be assisted with their Year 2000 efforts. First, state and local IT managers need to find ways to raise awareness among elected officials of the complexity of the interface problem. Currently, "county commissions are more concerned with potholes than [Year 2000]," he said. Second, state and local leaders should pool resources to develop standard solutions that can be used and reused across government agencies. Third, "local government needs to be brought to the table in the intergovernmental community," Greeves said.
Organizers of the Pittsburgh summit said the meeting was convened with no time to reach out to local governments, a situation Ridge and others hope to remedy by convening a national gathering for local governments and schools in April 1998. In the meantime, states are by no means indifferent to local Year 2000 needs. States with large metropolitan areas or those that already have established entrepreneurial relationships with their local governments are more progressive in solving the interface problem.
In Pennsylvania, for example, the state has joined with various universities to sponsor regional awareness conferences, and Year 2000 experts are available to speak to township supervisors, school districts and local chambers of commerce. "The idea is to use every local resource we already have in place," Olson said, pointing out that no state has time at this point to invent new communications channels.
Tackling the Problem
Fortunately, government organizations don't have to reinvent their Year 2000 strategies to solve the data-interface problem either, but they do need to modify or expand some strategies already in place. In trying to flush out interface problems-as well as flesh out solutions-attendees of the Pittsburgh summit divided their work into three broad areas of concentration: inventory and compliance (that is, finding and fixing); mutual certification and testing; and the building of electronic bridges when things go awry.
In some areas, the meeting produced usable, ground-breaking conclusions, such as settling on a four-year contiguous date standard to be used in data exchanges. In other cases, findings affirmed what many Year 2000 practitioners are already doing.
In terms of inventory, for example, managers have long advocated that organizations prioritize their systems to ensure the most critical will run unimpeded after the turn of the century. That same prioritization should be applied to data interfaces as well, the summit panel concluded. But state and local IT managers said that goal may not always be easy to reach because one agency's priority may well be another's back-burner issue, particularly in the case of state/federal interfaces.
In addition, forward-thinking managers should take the opportunity not just to prioritize systems and programs but to use the Year 2000 as a "sunset" date for certain systems. "Have you ever moved from one house to another?" Kolodney asked. "You don't need to move everything. [Year 2000] is a chance to re-evaluate systems and to get rid of or streamline a significant number." Such housecleaning requires more employee labor up front but reduces the Year 2000 burden and may save an agency money in systems support, he said.
Beyond prioritization, state governments need to "make sure the assumptions we've made in identifying interfaces are the same ones that the feds are identifying," Olson said. Detail work is important at this stage: The search for data interfaces should be conducted on a systems and a program level, and agencies sharing an interface should identify to one another exactly which employee on either side is responsible for the interface repairs.
When it comes to certification and testing-a phase that some industry watchers have estimated takes up fully 50 percent of most Year 2000 project resources-the summit subcommittee urged data partners to agree on a drop-dead date and to publish and exchange "exception lists" of systems that will not meet that date. The group further suggested Dec. 31, 1998, as a preferred date for compliance of mission-critical applications.
A third subcommittee confronted the potentially politically charged issue of electronic bridges. When one partner's data is compliant by either an agreed-upon date or zero-hour itself but the other partner's is not, someone has to develop an electronic bridge to accommodate the unfixed data.
Although the summit group resolved that "it is the responsibility of the noncompatible partner to provide the bridge," in reality, the "someone" much more likely will be the department that has already done its homework and has fixed its interfaces.
Cooperative Bridge Building
National Year 2000 leaders have gone to great pains to avoid assigning blame and to emphasize instead the importance of a cooperative spirit in reaching mutual solutions. But for many data managers in the field, the motivation behind building bridges will likely be less charitable and more focused on protecting their own hard work from data corruption and systems downtime.
"You hope your partner will get across the line the same time that you do, but if that doesn't happen, you need to find contingencies so you don't corrupt your database," Kolodney said. A mutual drop-dead date allows organizations time to assess the situation and begin remedial action-anything from erecting a firewall between the two systems to capturing the data offline and manually re-entering it into the compliant system.
States could potentially get caught twice: bridging to federal agencies that aren't as far along in their remediation while simultaneously monitoring and helping agencies along. Local and municipal players, meanwhile, might suffer from a lack of clout with larger agencies.
In San Diego, the city's information technology and communications department is well under way with its 12-person, $10 million, Year 2000 program, according to Rod Moyer, the department's Year 2000 project manager. The city is finished with its assessment phase, has begun remediation and is on track for meeting its self-imposed end-of-1998 deadline.
Data interfaces were identified as part of the inventory and refined during project assessment, but the city has not yet contacted directly any other agencies about fixing specific interfaces. "The calls haven't been made yet, but we're prepared to be flexible," Moyer said. "If [other agencies] have a good idea, we're hoping we can match it."
Moyer said he has heard of some state agencies and particularly federal agencies that are behind, but he is too busy concentrating on his own Year 2000 efforts to monitor progress in a general way. "We'll make the calls when we get to the programs one by one," he said. As for reaching agreed-upon standards ahead of time, the city would rather plow ahead and take the risk that some systems will have to be adjusted more than once. "We felt we had to be unwilling to wait [for data partners]. If we have to, we'll just have to fix a few things twice."
-- Tracy Mayor is a Beverly, Mass.-based free-lance writer specializing in IT. She can be reached at [email protected]
* * * * *
Fellow Year 2000 Comrades Advise:
* Executive buy-in is crucial to the success of any Year 2000 project. * Adopt a proven methodology or consider hiring consultants to supply one, rather than starting from scratch. * Consider hiring a third party to more accurately monitor progress and enforce deadlines. * During prioritization, identify systems that can be consolidated or eliminated.* To gain more clout with larger agencies, join with government groups at your level of government. * Don't reinvent the wheel. Use existing channels to communicate with local governments.* Establish action items to avoid "analysis paralysis."
* * * * *
Year 2000 in the Back Room
An invitation to October's Year 2000 data-interface summit probably wouldn't have carried much water with Dick Brich, data proc-essing supervisor for the city of North Platte, Neb.
"Is that the meeting 100 people went to?" Brich asked. "If they think 100 people can straighten this mess out, we're in serious trouble." Brich doubts he'd have had time to fly to Pittsburgh even if he had been invited; he and one other data processing employee are in the middle of fixing 1,578 Cobol source code files by hand-in between their other duties. "I haven't been devoting enough time to it lately," he lamented. "I had to quit doing this to write a new payroll application."
North Platte's Year 2000 strategy involves a minimum amount of planning and a maximum amount of work. Brich's inventory process consisted of printing out a list in job control language of every program the department maintains. Prioritization was similarly straightforward: External programs take precedence over those used only in his department, and revenue-related programs are fixed before all others. His methodology? "We just take a system and fix it, then take the next one and fix that."
The data interfaces Brich worries most about aren't with state or county governments but with the Omaha bank with which North Platte has an electronic-commerce arrangement.
To date, the bank hasn't fixed its own systems, and Brich has no idea how or when it will, but he knows, as a small customer of a large bank, he'll have to be ready to adjust his own systems once again when the time comes.
So how is North Platte doing? With 70 percent of his programs remediated already, Brich is far ahead of the national curve. But he's not resting on his laurels just yet. "Am I going to make it? Call here the week after New Year's . If I'm not here, you'll know what happened."