Aging information systems are expensive to maintain and most are eventually retired and replaced. But what determines (in the choices made by managers) whether and when a system reaches end-of-life? What shapes managers' judgements about a system's remaining life expectancy and do these judgments influence the maintenance effort itself? System maintenance and prospective replacement are examined here in new terms, positing that managers 'equilibrate' (balance) their allocation of maintenance effort with their expectations of a system's remaining life. Drawing from data on 758 systems among 54 organizations, support is found for an exploratory structural equation model in which the relationship between maintenance effort and remaining life expectancy is newly explained. A 'portfolio effect,' reflecting a system's familial complexity, is also found to be directly and positively related to the maintenance effort. A further finding is that a system's size is directly and positively associated with its remaining life expectancy. Notwithstanding normative research suggesting the contrary, larger systems may tend to be longer-lived than smaller systems. Practically, the suggestion is made that better documented and monitored portfolios, together with regular, periodic performance assessments, can lead to better management of systems' life cycles.