According to the networking best practices, which network device should be used for optimization and rate limiting?

Provider core devices are designed for high-speed forwarding and typically avoid extensive per-customer QoS or rate limiting to maintain maximum throughput.

While provider edge devices may apply some policies, the customer has direct control and visibility over their own customer edge device for specific internal optimization and rate limiting needs.

Customer core devices primarily focus on high-speed switching and routing within the internal customer network, not typically on policing or shaping traffic at the network ingress/egress.

The customer edge device is the most appropriate place for implementing optimization techniques like QoS and rate limiting. This device sits at the boundary of the customer's network, making it the ideal control point to manage traffic flow, enforce bandwidth limits, and prioritize critical applications as traffic enters or exits the customer's domain, thereby preventing network congestion or over-utilization of links.