Department of Health and Human Services

Optimizing the Computer Security and Update Process

Solving two computer issues would increase government efficiency and save money. Presently, HHS computers receive software updates overnight, and sporadic reminder emails for the updates (1 day to 1 month in advance) result in individuals leaving computers on overnight. Also, security scans, which substantially reduce computing speed, occur throughout the day. Today, I was able to open only one webpage in 35 minutes due to a scan. A poll of 12 employees in my office suggests that 83% of workers lose productivity due to scan-related slowdowns, productivity losses average 1.92 hours per worker per week, and 58% of workers do not regularly shut down at night. Three reforms are needed. First, update teams should send mass emails to leave computers on overnight at 3pm on update days. This generates substantial energy savings by reducing the computers unnecessarily left on overnight. SAVE proposals advocating mandatory shutdown estimate $10 million in annual savings. Second, security scans should be queued for users to accept during lunch, a meeting, or other convenient time. If users do not accept scans by COB, they can run at computer shutdown or with major updates that evening. Third, workers should receive an educational email regarding the changes and be offered a going-green reward – perhaps a green leaf screen saver – for pledging to shut down their computers. The reward publicly recognizes employees saving power, nudges individuals to pledge to shut down, and serves as a cost-free, daily reminder for employees to shut down. Reducing morning boot times, which are up to 15 minutes, also would increase shutdown rates and should be explored. This low-cost proposal increases worker efficiency by roughly 4.8 percent (1.92 of 40 hours per week) and reduces agency energy expenditures. It also avoids one-size-fits-all policies like mandatory shutdowns that preclude overnight computing capacity, which saves worker time on tasks like complex data analyses.



Idea No. 14184