An online video game company must maintain ultra-low latency for its game servers.

The requirements are UDP support, ultra-low latency, and the ability to handle millions of requests per second in a cost-effective manner. Network Load Balancer (NLB) is the AWS service designed for high-performance Layer 4 load balancing, including TCP, TLS, and UDP. NLB can scale to very high throughput and connection rates while maintaining low latency, which is critical for real-time gaming traffic.

Option C is correct because NLB supports UDP listeners and forwards traffic directly to EC2 targets with minimal processing overhead. This yields lower latency than Layer 7 solutions and is well suited for game protocols that are latency-sensitive and often use UDP for fast, lightweight communication. NLB also provides static IP addresses per Availability Zone and integrates cleanly with Auto Scaling groups.

Option A is incorrect because an Application Load Balancer operates at Layer 7 (HTTP/HTTPS) and does not handle raw UDP traffic; it’s optimized for web applications, not gaming protocols. Option B is incorrect because Gateway Load Balancer is intended to deploy and scale third-party virtual appliances (such as firewalls) and uses Geneve encapsulation; it is not meant for distributing general internet game traffic to EC2 game servers. Option D adds significant cost and complexity by deploying multi-Region fleets. It may help global latency in some scenarios, but the question asks for the most cost-effective way to handle massive UDP traffic volume; multi-Region duplication is typically not the first choice for that requirement alone.

Therefore, C (NLB with UDP) best meets the scale and latency requirements with the most efficient, purpose-built AWS load balancing service.