Snapipeline: Accelerating Snapshot Startup for FaaS Containers

Due to the frequent starts and stops of numerous services in FaaS (Function as a Service), reducing cold start overhead is a core issue in improving the performance of container-based FaaS services. Snapshot and restore-based mechanisms effectively reduce the cold start time of containers by transforming container initialization overhead into restoration overhead. Consequently, this mechanism has become a research hotspot in accelerating the cold start of FaaS containers. Researchers introduce snapshot compression and decompress the snapshots to reduce the storage cost before starting instances. However, existing works have the following shortcomings: (1) File-mapped memory pages are not processed during snapshot compression, resulting in a significant amount of redundant data in memory; (2) The serial execution of snapshot decompression and instance restoration leads to high instance startup latency.

To address these shortcomings, we propose the Snapipeline mechanism, which implements the following optimizations: (1) In the snapshot compression phase, it extends deduplication to file-mapped memory; (2) In the restoration phase, it leverages the hot and cold distinction in FaaS application memory to prioritize the memory pages restoration, pipelining snapshot decompression, memory pages restoration, and function execution. This mechanism hides the expensive snapshot decompression latency behind the instance restoration and function execution latency and removes the snapshot decompression from the critical path of instance startup. Evaluation on real-world FaaS application datasets shows that Snapipeline reduces memory usage by up to 28% and decreases end-to-end latency by an average of 53%, compared to the baseline.