Abstract:
Techniques for facilitating high availability in a device (e.g., a network device) comprising redundant processing entities (e.g., one or more processors, one or more cores, etc.) and a transactional memory system. The transactional memory system comprises a memory that is shareable between the redundant processing entities and ensures consistency of information stored in the memory at the atomicity of a transaction. A first processing entity may operate in a first mode (e.g., active mode) while a second processing entity operates in a second mode (e.g., standby mode). Operational state information used by the active processing entity for performing a set of functions in the first mode may be stored in the shared memory. Upon a switchover, the second processing entity may start to operate in the first mode and commence performing the set of functions using the operational state information stored by the transactional memory system.