Senthil Babu Malli Jayaraj
Public sector enterprise applications hosted on Amazon Web Services (AWS), spanning benefits eligibility portals, tax filing systems, and unemployment insurance platforms among others, experience demand patterns fundamentally different from typical commercial enterprise applications. Public program demand frequently spikes in response to policy changes, filing deadlines, natural disasters, or economic conditions, producing surges that can exceed baseline traffic by an order of magnitude within a short window, followed by extended periods of comparatively low, stable utilization. This research article presents a capacity planning and performance optimization framework tailored to AWS-hosted public sector enterprise applications operating under this fluctuating demand pattern. The article examines multi-tier scaling architecture, cost versus performance tradeoffs across common capacity strategies, latency and utilization patterns observed across time of day and day of week, and a five stage capacity planning maturity model. Findings are illustrated throughout with complex charts, including a demand and capacity overlay, a latency heatmap, a cost versus performance scatter plot, a stacked cost composition chart, and a multi-dimensional optimization radar chart, alongside extensive grid style data tables. Drawing on AWS architectural guidance, cloud cost optimization research, and public sector digital service performance literature published through June 2023, this article finds that agencies combining elastic auto scaling architecture with proactive, demand-informed capacity planning report substantially better cost efficiency and materially lower latency degradation during demand spikes compared to agencies relying on static, manually provisioned capacity.
Keywords: AWS, capacity planning, performance optimization, auto scaling, public sector cloud, cost optimization, latency, elastic infrastructure, demand forecasting