230 likes | 554 Views
Berk Atikoglu, Yuehai Xu , Eitan Fracthenberg , Song Yiang , Mike Paleczny. Workload Analysis of a Large-Scale Key-Value Store. Analyze Memcached at Facebook. +284,000,000,000 requests 5 different use cases Workload characteristics, locality, cache effectiveness.
E N D
Berk Atikoglu, YuehaiXu, EitanFracthenberg, Song Yiang, Mike Paleczny Workload Analysis of a Large-Scale Key-Value Store
Analyze Memcached at Facebook • +284,000,000,000 requests • 5 different use cases • Workload characteristics, locality, cache effectiveness
Why Is Caching Important? Database Cache Servers Web Servers
Motivation • Understand workload characteristics • Identify factors affecting performance • Provide a benchmark for future studies
Memcached • Distributed memory caching system • Key-value store for small objects Key Hash Function Memcached Servers
Tracing Methodology • Capture traces through a Linux Kernel Module (LKM) • Process traces with Hive LKM
Facebook Deployment Contains server related information Anything that doesn’t belong to a specific pool goes to ETC
Analysis • Workload Characteristics • Locality, Cache Behavior
Request Composition > 99.8% GET GET:UPDATE = 30:1
Key Size Distribution 90% of VAR keys are 31B USR keys are 16B or 21B ETC is heterogeneous
Value Size Distribution USR values are only 2B 90% of values are smaller than 500B
Value Size Dist. By Overall Weight 90% of data is generated by values of 500B or smaller except ETC 90% is 10KB or smaller values for ETC
Request Rate Over Time All pools show diurnal pattern except SYS
Request Rate Over Time (ETC) North America starts its day Night time in Western Semiphere
Analysis • Workload Characteristics • Locality, Cache Behavior
Repeating Keys 0.0003% of keys in 10% of requests in ETC 1% of keys in 55% of requests in ETC Least frequent 50% of keys in 1% of requests in ETC
Reuse Period of Keys 99.9% of SYS keys are reused in 1hr 88.5% of ETC keys are reused in 1hr 96.4% of ETC keys are reused in 6hr
Hit Rate Why? 98.2% 92.9% 81.4% 93.7% 98.7%
Conclusion • Analyzed 5 different memcached use cases • Different applications of memcached have extreme variations in access patterns • Answered pertinent questions to improve Facebook’s memcachedusage
Thank You • Questions?