Virtual Service performance is driven primarily by Service Engine data-plane capacity, not Controller resources. If a Service Engine is busy, performance can be improved by reducing the workload on that SE, vertically scaling SE resources, or horizontally scaling traffic across additional Service Engines. Migrating other Virtual Services away reduces contention for CPU, memory, and packet-processing capacity. Increasing CPU and memory on Service Engines increases data-plane capacity, although the SE VM may require reboot depending on the environment. ECMP scale-out can distribute traffic across multiple Service Engines and improve throughput. Increasing Controller CPU and memory helps management or analytics capacity, not Virtual Service data-plane performance. Moving from EC to RSA would usually reduce SSL performance because RSA processing is generally more expensive. Therefore, the correct choices are A, B, and D.