Efficient data center management in educational platforms faces significant challenges in balancing performance and energy consumption, particularly in resource-constrained settings like developing countries. This study evaluates the \(\textrm{CiS}^2\) (Consolidated Index for CPU-Server Saturation) metric as a tool to optimize this balance in technological infrastructure. Employing a mixed-methods design, we compared an original server selection algorithm with an optimized version across seven homogeneous servers equipped with Intel® Xeon® E5-2670 v3 processors and VMware vSphere 7.0. A one-way ANOVA ( \(\alpha = 0.05\) ) confirmed significant differences in \(\textrm{CiS}_{n_s}^2\) across servers \((F(6, 14) = 10.45, p < 0.001)\) , while a survey of ICT administrators (Cronbach’s \(\alpha = 0.87\) ) validated its practical utility (mean = 4.4/5). The optimized algorithm enhanced computational efficiency, reducing complexity from \(\mathcal {O}(m \cdot k)\) to \(\mathcal {O}(m)\) , and consistently selected UFASLQUTICVMW01 as the most efficient server (score = 4.98 ± 0.5 with \(\beta = 1.0\) ), though its advantage diminished under higher saturation penalties ( \(\beta = 3.0\) ). These findings highlight \(\textrm{CiS}^2\) ’s effectiveness in identifying efficient configurations and support sustainable optimization strategies, reducing power consumption by an estimated 15%, for educational platforms in similar resource-limited contexts.