When its mainframe becomes obsolete, a city must instigate a program to move its departmental IT systems to a new server
BY PETER FRETTY // PHOTOS BY DAVID AHNTHOLZ
City officials in Inglewood, California, USA realized their governmental systems, including the one that handled emergency response, were dependent on a mainframe that was rapidly becoming out-of-date. Although highly resilient, the mainframe increasingly cost more and more to maintain—and was becoming harder and harder to repair.
As the mainframe aged and the old-school programmers and analysts retired, it became evident that migration to a new platform was imminent.
“We needed an upgrade path that alleviated our dependency on the mainframe, took into account the budgetary limitations inherent to the public sector, and still provided a robust and desirable long-term solution without compromising the complex business needs of a multifaceted organization,” says Michael D. Falkow, PMP, CIO and assistant city manager for Inglewood. “We also needed to put the city in a position where its information technology and communications (ITC) department could efficiently and effectively support and maintain the line of business applications without having to rely upon expensive software developers.”
The city decided there was no choice but to migrate to a new platform soon after Mr. Falkow became the ITC director in mid-2005.
He launched a more than US$4 million strategic program to evaluate each and every line of business application that resided on the mainframe.
“From a resource perspective, I was able to convince the city manager and city council to allow me to create a systems analysis and implementation division of the ITC department where the primary mission was to evaluate technology solutions across all city departments, and a public safety systems division focused solely on our police department's technology needs,” Mr. Falkow says. “Inherent in this strategic evaluation was a buy-versus-build-versus-modernize approach.”
ITC officials found that in some cases, it might be cost-effective and more efficient to purchase an entirely new system; in others it could prove beneficial to join forces with other organizations.
DEPARTMENT BY DEPARTMENT
The mainframe decommissioning took over five years and involved a host of internal resources, including computer operators, systems analysts, project managers, developers, business process experts and trainers, as well as dozens of contractors and consultants.
“After we evaluated each application, we created a project for each. In the case of system replacements—for example, a homegrown line of business application being replaced with a commercial off-the-shelf solution—we issued a request for proposal (RFP) and went through the standard procurement process,” he says.
The finance, payroll, timekeeping and personnel system as well as the library solution fell into this category.
“Given our resource limitations, we weren't able to assign a specific project manager to each project,” Mr. Falkow says. “In some cases, the systems implementation specialist or systems analyst became the de facto project manager. In other cases, it was the business process expert within the department who took charge.”
Once the project team acquired the new system, it trained the user base as part of its implementation process.
LIBRARY SYSTEM: The process was one of the easiest systems to convert, and most of the schedule was to account for the demands of the bureaucracy, Mr. Falkow says.
He and the library director visited a site in Phoenix, Arizona, USA to see a sizable and successful installation of the vendor system. After city council approval, the project team extracted all of the data from the mainframe library system, and the vendor converted it over to its system. After a few weeks of testing, the system went on line—nine months total, from RFP to go-live.
PARKING TICKET SYSTEM: The parking ticket system proved much more complex than the library's. The tool processes parking citations, not only for the city but for more than 60 other municipal entities around the state as well. This provides net revenue to Inglewood through a government-to-government shared services model. Each client agency pays for the services it uses—including an administration fee for the program management provided by Inglewood.
That financial incentive led the governmental agency to deem it worthwhile to continue providing those services, Mr. Falkow says.
But the cost of developing a replacement system was estimated at US$1.5 million, and new server equipment would cost an additional US$300,000.
August 2005 Program launch
September 2005 Began discussions on homegrown computer-aided dispatch system
October 2005 Request for proposal (RFP) issued for finance, payroll and human resources system
February 2006 Go-live at neighboring city through government-to-government shared services model for the safety records management system
September 2006 RFPs issued for dispatch and parking ticket systems
November 2006 RFP issued for library system
January 2007 Library system project launch
March 2007 Public-private partnership established for parking ticket system
October 2007 Vendor agreement on computer-aided dispatch system migration
March 2008 Library system goes live
July 2008 Final client agency converted to new parking ticket solution
August 2009 First attempt at the dispatch system project fails
November 2009 New dispatch system project launched
September 2010 Second attempt at dispatch system goes live
October 2011 Residential sound insulation department project ends
November 2011 Mainframe fully decommissioned
The public-private partnership strategy for the parking ticket system avoided an estimated US$1.8 million in capital expenditures required to develop and implement a replacement system, and it shortened the required implementation time by 24 months.
—Michael D. Falkow, PMP
In addition, city officials voiced concerns about the need to improve system backup and disaster-recovery capabilities.
“It was determined that the most cost-effective and timely course of action was to contract with a private-sector outsourced service provider who could provide a hosted system and the related technical and operational support staff,” Mr. Falkow says.
The city issued its RFP in September 2006, and eventually decided upon a vendor to provide a hosted, web-enabled citation management system, as well as the technical staff for the conversion and implementation across all client agencies.
“Upon completion of the implementation project, the city was in a position to terminate the mainframe-based citation management system it had built, resulting in operational savings associated with the maintenance and upkeep of the system,” Mr. Falkow says. “It also allowed for the retirement of the principal software engineer. That position was removed from the budget from that point forward.”
Overall, the initiative resulted in ongoing savings of more than US$200,000 per year in personnel costs, he adds.
The city of Inglewood was the first agency to implement the hosted system, followed by the partner agencies based upon their size, staff availability, configuration requirements and physical location.
“The public-private partnership strategy avoided an estimated US$1.8 million in capital expenditures required to develop and implement a replacement system, and it shortened the required implementation time by 24 months,” Mr. Falkow attests.
RESIDENTIAL SOUND INSULATION SYSTEM: The ITC team's last project in the decommissioning program involved residential sound insulation. This department's mission is to attain, coordinate and manage grant funds provided by the U.S. Federal Aviation Administration and the Los Angeles World Airports, which runs Los Angeles International Airport (known as LAX). These funds are used to sound-insulate homes.
To comply with the grant-reporting requirements, the department must file financial reports.
“Recently, the granting authorities asked for a great deal of data that went back several years. We have much of this data scanned and archived in our enterprise document management system or available in hardcopy files,” Mr. Falkow says, “but the only way to determine what exactly we need to retrieve is to access the legacy finance system on the mainframe.”
This required him to contact a retired mainframe programmer/analyst for assistance. The project closed in October 2011, wrapping up the complex multi-year program.
NO SMALL CHANCE
Looking back on the enterprise-wide systems upgrade program, the biggest lesson was the need to always evaluate how a change might affect end-users before making a decision, Mr. Falkow says.
“In the case of our emergency-response system, knowing that a new system would have an impact on our 911 dispatchers and public safety personnel in the field, we chose to modernize with the help of an application migration specialist rather than buy an off-the-shelf solution,” he says. “You can't really pinpoint a dollar value here, but it is indeed a critical success factor. In project management, success isn't always based solely upon correctly implementing a solution on time and on budget. If the users reject it, it's a failure—usually an expensive one!”
The project teams knew that any major change in the way departments do business would be significantly disruptive to the user base.
Mr. Falkow also learned the significance of facilitating and supporting the project as a team effort—not just internally but externally as well.
“Vendors must be brought in as partners from the start, and they have to want success just as much as the customer,” he says. “When the going got tough and we were trying to overcome hurdles, it was the strength of our partnerships that pushed us over the top and made the program a success.”
The organization also assigned a strategic account manager to provide a focal point to address challenges, help triage severe problems and maintain an ongoing support dialogue.
MAPPING OUT AN IT MIGRATION PROJECT
When the city of Inglewood, California, USA's emergency-dispatch migration project failed, Micro Focus was brought in to save the initiative. Here, Kevin Brearley, director of product management at the enterprise software maker in Newbury, Berkshire, England, offers his advice for IT migration projects:
To stay on track and ensure success with migrations, I follow this seven-step roadmap:
1. Clearly establish delivery ownership and schedule commitments of each migration component.
2. During the analysis and design delivery phase, apply expert advice to technical solution design to ensure the solution fits your requirements and environment.
3. Apply strong project management and solution architecture expertise across your project life cycle and assign strong internal subject-matter experts to ensure solution adoption and success.
4. Prepare robust problem-resolution processes early (toward the end of the “build phase”) and leverage pre-existing test processes to avoid wasting time in post-migration testing and production support.
5. Adopt an incident-tracking solution from the start of your project and organize support processes and operations environment support during the project initiation phase.
6. Limit data conversion planning to only “input files.”
7. Plan for cost savings, then track achievement.
WHAT HAPPENS WHEN A MISSION-CRITICAL PROJECT FAILS?
Because of the mission-critical nature of supplying software to the emergency call center and the city's fleet of patrol vehicles, migrating the computer-aided dispatch system was the most intense project, says Michael D. Falkow, PMP, City of Inglewood, California, USA.
After a two-year failed attempt that cost Inglewood US$210,000 (not including internal staff costs), the city partnered with a vendor in August 2009.
WHAT WENT WRONG: The dispatch system was the most complex system in the migration program—not only because it is a mission-critical public safety system but also because the feature set was superior to any readily available commercial system.
“After all, we built, maintained and enhanced the system over the course of more than two decades,” Mr. Falkow says. “To make matters worse, we didn't have the funding available to purchase a commercial solution, which we estimated would cost us more than US$1.5 million.”
After an evaluation of options, the project team selected a subcontractor with the understanding that a code migration would take an estimated six months.
“After 19 months, we had to abandon the project after a simple operating system security update took the development system down for nearly eight hours,” he says. “We couldn't risk taking a system as critical as a dispatch system into production and have a simple patch break it.”
WHAT WAS DONE ABOUT IT: At this point, the city switched to a vendor that could emulate the mainframe environment on the new server platform.
The failure, while disheartening, didn't disrupt endusers because they were still operating on the mainframe.
“As a team, we were certainly unhappy, as a great deal of effort went into the project,” Mr. Falkow says. “I took special care to work with my team to ensure that this defeat wouldn't permanently disable them from putting in the same or more effort going forward. Getting your team to accept defeat, not allowing them to wallow in self-pity and instead leveraging the loss as a strength going forward is a skill only learned through experience.”
Not only did he have to report the failure to the city council, but he had to ask the members for an additional US$112,000 for a second attempt.
“Project failure is only as bad as you make it. If you learn from your mistakes, leverage as much of the work you did going forward and mitigate the costs of the next attempt, often the initial failure can be overshadowed by success,” he says.
The project team accomplished a rebound by leveraging much of the time, money and resources that went into the unsuccessful initiative and transferred it to the new partner.
“The work we did with the first vendor involved evaluating the interconnectivities among all the dispatch screens, and understanding the way in which data was stored, manipulated and accessed in and from the mainframe's data structure,” Mr. Falkow explains.
Once the second project successfully wrapped up, the new system allowed the governmental organization to eliminate two full-time employee positions through retirement, saving more than US$210,000 annually in salary and benefits. In addition, the city was able to finally decommission its aged mainframe, another savings of US$120,000 per year in maintenance and support.
On a governmental program of this size, accountability falls heavily on all involved. The mainframe migration was highly visible to the city manager and the city council because it was about more than just money, Mr. Falkow explains. “It involved the way the city operates as a whole, and it set the stage for how it would operate going forward,” he says.
The city's ITC department has achieved a 50 percent reduction in staff headcount, from 36 to 18.
“That, coupled with the increases in revenue from our government-to-government operations and decreases in maintenance fees, is dramatic to say the least,” Mr. Falkow says. “The entire program cost less than anticipated as well. Each year, we came in under budget. Part of this shrinking budget was due to cost-savings measures we intentionally put in place, and part of it was due to streamlining and other efficiencies.” PM
PM NETWORK MARCH 2012 WWW.PMI.ORG
MARCH 2012 PM NETWORK