~/talk

It's Time to Revisit LRU vs. FIFO

github:ohadeytan
links:none on file
name:Effi Ofer
links:none on file
links:none on file
name:Ronen Kat
links:none on file
slides:none on file

Reevaluates LRU vs FIFO by modeling metadata access cost and simulating caches on 99 cloud-object traces plus standard MSR/TPCC workloads, varying cache sizes and latency ratios. Finds that although LRU often wins on raw hit rate, when metadata no longer fits in RAM and persistent metadata access is costly — e.g. 1:50 local:remote latency — FIFO yields lower overall cost in the majority of cases (dozens of configurations, ~30-case example), so FIFO is a serious candidate.

AI-written summary · about

  1. 0:00Start
  2. 0:31The Essence of Caching
  3. 2:15Traditionally: LRU Considered Better
  4. 3:46Motivation - Cloud Object Storage
  5. 6:20IBM Cloud Object Storage Traces
  6. 11:55Conclusions & Discussion
~/search
type to search 6,249 conferences, 138,987 speakers and 185,172 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.