The call processing architecture of switch design has always been a subject of extensive interest in the data networking and telecommunication industries. In the early days of switch design, centralized call processing architecture, in which a single control module performed all processing tasks, was largely adopted. With the advent of high-speed switch capacity and increased traffic volume, centralized call processing architecture has become a significant bottleneck to achieving high levels of call processing capacity in switch design. Call processing architectures are evaluated based on scalable performance, high availability, and flexible configuration. An architecture should be scalable to accommodate the number and types of network interface modules to be supported initially and in the future. It should be able to process a high volume of calls and, if properly implemented, it should allow a balanced distribution of processing power across the entire system. It should also be robust enough to eliminate any single point of failure. This paper proposes three alternative call processing architectures — distributed signaling architecture, distributed call control architecture, and distributed routing architecture — all of which provide scalable performance and enhanced reliability. Distributing call processing power to each network interface module in a switch is the key to these proposed architectures, enabling them to overcome the deficiencies associated with centralized call processing architecture.