Consideration of reliability in the design of Water Distribution Networks (WDNs) requires performance evaluation of the network under various conditions that affect reliability, which makes the design methodology tedious. Several surrogate measures of reliability have been suggested, the improvement of which shows improvement in reliability and obviates the tedious performance evaluation step. Variance of Pipe Flows (VPF) is one among the different surrogate measures. The reduction of VPF improves the uniformity of flows in different pipes. Therefore, failure of any pipe and its isolation results in lesser impact as compared to a network in which flows are concentrated in a few major pipes. A two-step methodology was earlier used in which flows in pipe networks were obtained first using genetic algorithm to minimize the VPF. Then, in the second step, the network was designed to minimize the cost of the network considering the assigned flows. This article presents a single-step methodology using Non-dominated Sorting Genetic Algorithm (NSGA) for the optimal design of a WDN considering the VPF as a surrogate reliability measure. The pipe diameters are taken as decision variables and a Pareto front is generated between network cost and VPF. A solution that minimizes both VPF and network cost is considered the best solution. The methodology is illustrated with a simple two-loop network taken from the literature. The obtained solutions are compared with those obtained through a two-step methodology. The relative advantages and limitations are discussed.