Today, most available parallel environments support multiple users executing their jobs in a common shared infrastructure. In such environments, when scheduling individually important jobs from different users, the problem of fairness arises. In this work, we propose a fair scheduling algorithm that handles this problem by adopting processor fair-share as a strategy for user fairness. The user’s workload follows the Campaign Scheduling model, in which each workload is characterized by the submission of successive sets (or campaigns) of sequential and independent jobs. The algorithm guarantees that the flow time of a campaign is proportional to campaign’s size and the total number of users. In contrast to the classic queuing systems, the flow time does not depend on the total system load. Internally, the algorithm maintains a virtual time-sharing schedule in which each user is assigned the same share of processors. The order of completion in the virtual schedule determines the execution order on the real processors. We analyse the performance of the algorithm by simulation on a real parallel machine workload trace. Our approach shows that a parallel machine can give a similar type of performance guarantee as a round-robin scheduler, no job preemption been required. Consequently, processors are shared in a proportionally-fair manner among users.