State of Alaska saves 10 percent on mainframe software
Published on December 30, 2003
on CNET's TechRepublic.com
The State of Alaska provides public services to its 640,000 residents who occupy 570,374 square miles, or more than 300 million acres of land, making Alaska the largest state in the union—about 2.3 times the size of Texas and about one-fifth the size of the lower 48 states.
The State of Alaska has a single mainframe computer located in Juneau. Many of the state's 15,000 employees tap into the mainframe from its three main population centers in Anchorage, Juneau, and Fairbanks as well as outlying areas spread out over hundreds of miles.
Twelve government agencies rely on this single CPU for all of their data storage and processing needs, including the Department of Labor and the Department of Health and Social Service. These agencies connect to the mainframe through a multiplicity of servers and telecommunication systems, each through TCP/IP protocol.
The dilemma: No program of asset management and no data from which to benchmark
Businesses and government organizations are constantly facing higher mainframe software costs for a variety of reasons These may include vendor pricing hikes, as well as corporate mergers and acquisitions, divestitures, data center consolidation or relocation, or hardware upgrades, all potentially resulting in expensive mainframe software redundancies.
Most computer executives suspect they have inflated mainframe expenses, but are unable to tackle the challenge of mainframe asset management because they lack hard industry data.
The challenge of measuring true mainframe software costs at state governments is no different. Although individual departments at the State of Alaska had made stops and starts at benchmarking mainframe software assets, the state had no overarching program in place to do so, and any efforts to do asset management were stuck in information silos or simply short-lived.
After attending a software asset management session at a Share Conference for IBM computer users in 2002, Allen Morse, a technical services manager for the State of Alaska, started to consider what a comprehensive program of software asset management could do. "I had the feeling that it could save us significant money by taking a close look at the software portfolio, but we had no way to really demonstrate it or qualify it," said Morse.
The solution: ISAM'sSCORcard and Software Portfolio Review and Analysis
After defining the need for a multiple-departmental asset management program, the State of Alaska issued a request for bids. Only two companies stepped up: St. Paul, MN-based ISAM, and another organization whose services merely offered discovery and cataloging the state's mainframe software portfolio. The latter didn't go far enough to identify the potential cost savings Morse sought to uncover.
ISAM offers services that prime Blue Chip companies and government agencies with cost comparison data that can significantly reduce mainframe software expenses. The services analyze mainframe software costs and compare them to industry averages.
SCORcard is a free assessment that gives executives a high-level picture of where they may be burdened with excessive mainframe licensing costs. The more detailed Software Portfolio Review and Analysis benchmarks the actual licensing costs paid against 500 vendor price books, highly confidential data that demonstrates the price variations of more than 15,000 products. An analysis of the average industry costs against the actual cost a company pays for its software can help companies identify overhead and unnecessary redundancy. Coupled with standard contract language, companies can use the data to realize real cost savings.
Following an initial SCORcard assessment from ISAM, which compared how much the State of Alaska was paying for its mainframe software products in comparison to its peers, the State selected ISAM to do a complete rundown of where the State could trim the fat.
The 90-day project began with a review of how each of the State's 12 agencies used its software portfolio. Ordinarily, a PILS review (Product Image Licenses) would yield the needed information. But as the State ran only a single CPU, the PILS was rendered next to useless because it traditionally examines CPUs dedicated to single departments and then measures the utilization of each application on the image.
Instead, the technical leads at each State agency completed a granular survey to identify the perceived value of each function, how frequently it was used, and how many people were using it. ISAM then fed this data along with the State's overall software portfolio and licensing expenses into its analysis pack. The results were surprising.
From the ISAM assessment, Morse quickly identified software that could be eliminated or replaced based on its minimal use or perceived lack of functionality. For example, a handful of users had substituted TSO (Time Sharing Option) software, an application used for programming activities, in favor of Roscoe, a TSO replacement editor. But the low number of users didn't justify the expense of having two applications that preformed basically the same function, so the State decided to eliminate Roscoe from its portfolio.
Decisions like this one to cut out unnecessary redundancy and, in some cases, replace software with less expensive options helped the state eliminate 10 percent of its mainframe software overhead from its total portfolio cost.
Since the ISAM report was issued a year ago, the State of Alaska has significantly altered its software portfolio, partly as a result of the ISAM report and partly due to a planned mainframe upgrade. The upgrade to the mainframe software has resulted in near zero budget growth for the next three years, in large part due to the ISAM report, said Morse.
Going forward, Morse plans to use ISAM's Software Portfolio Review and Analysis again to establish the delta from the baseline he established a year ago. Also, because the State has altered its mainframe portfolio significantly, Morse wants to create new baselines from which to judge future asset management efforts.
The overall experience has been so positive that Morse says he plans to extend the asset management program beyond the mainframe to examine the network databases and server utility. He won't be able to use ISAM, however, to benchmark non-mainframe packages. Instead, he'll have to turn to vendors such as Compass or Gartner, companies that benchmark datacenter costs related to hardware, software, and labor at a macro level.
|Time on the project||90 days|
|Program staff of 12 state agencies||Two hours each 24 hours total|
|Project leader||Half time for 90 days 1,200 hours total|
|Two project members as needed||2,236 hours total|
|Total project hours||3,460|
|One-time costs||Not more than $75,000|
|Cost savings||10 percent of total mainframe software portfolio costs|
|Projected long-term costs||One more annual benchmarking|