אוהד איתן, הרצאה סמינריונית לדוקטורט
יום שני, 18.10.2021, 13:00
Although caching is a well-studied topic and a widely used approach, modern software cache systems still struggle to optimize their decisions for many types of environments and workloads. This talk will briefly present four of our studies trying to tackle several on-going caching challenges.
The first study reexamines the FIFO vs. LRU battle, showing that modern cache systems can often benefit from the simpler FIFO policy.
The second study presents an adaptivity mechanism for software cache management schemes that offers tuning parameters targeted at the frequency vs. recency bias.
The third study deals with size-aware caching policies and the difference between hit-ratio and byte-hit-ratio in these settings.
The fourth study considers a multi-level cache environment and the importance of writes saving, especially for hybrid DRAM/SSD caches.
Some of our ideas are implemented as part of the popular open-source Caffeine (Java), Ristretto (Go), and Moka (Rust) caching libraries.