CATS: A Cache Time-to-Live Setting Auto Adjustment Strategy for an Air Ticket Query Service

Chunhou Liu, Jian Cao, Yudong Tan, Shiyou Qian

Abstract: A cache is an efficient way to speed up the response of services, which can also be applied in air ticket query services. However, the time-to-live (TTL) setting is a challenging task for the caching mechanism especially when the validity periods of the records change with time, which is the case for the air tickets. In air ticket query services, the validity of air ticket information is verified when a user places the order. It provides a chance to adjust the initial TTLs. In this paper, we provide a CAche Time-to-live setting auto adjustment Strategy (CATS) for air ticket query service based on the verification results during order placement. In addition, by considering the associations between cached records, CATS can also adjust the TTL settings for related cached records. Experiments are conducted to test and compare different implementations of CATS and the results show that CATS can maintain both the hit rates and verification success rates at a high level.