https://memcached.org/blog/ Recent content on Memcached Hugo -- gohugo.io en-us Sun, 07 Jun 2026 16:21:31 -0700 https://memcached.org/blog/cache-cost/ Sun, 07 Jun 2026 16:21:31 -0700 https://memcached.org/blog/cache-cost/ Judging the Cost of Replacing Cached Items - Dormando (June 3, 2026) Introduction Cache systems differ from databases: when full they must (quickly!) remove something to make room for new data. They do this by assigning a cost (or value) to items already in or new to the cache. Items are organized to quickly discover the lowest cost item to remove as new items enter. Academic papers describe algorithms in terms of performance: achieving a high hit ratio (rate of cache hits vs misses), low latency, and so on. https://memcached.org/blog/how-long-for-real/ Fri, 22 May 2026 14:07:28 -0700 https://memcached.org/blog/how-long-for-real/ How Long Does That Response Take... For Real? - Dormando (May 8, 2026) Introduction Why does memcached not have response time metrics? This is a frustrating question with an unsatisfying answer: the metrics would be misleading. Kicking off with a spoiler: memcached response time is best measured by sampling response times from the client. This takes the entire round trip into consideration and gives the most actionable information, most of the time. https://memcached.org/blog/proxy-intro/ Wed, 27 Mar 2024 12:44:59 -0700 https://memcached.org/blog/proxy-intro/ Introduction: Memcached Proxy - Dormando (March 27, 2024) What is this and why? A generic cache pool management proxy, built into the memcached daemon. It supports all of the text protocol. It is designed to abstract clients away from pools of memcached servers, for ease of management, scale-out, and portability. It uses Lua to create a highly flexible configuration system, allowing a company to fully customize how requests are routed to fit their specific needs. https://memcached.org/blog/paper-review-memc3/ Sun, 23 Feb 2020 12:10:51 -0800 https://memcached.org/blog/paper-review-memc3/ Paper Review: MemC3 - Dormando (February 23, 2020) Memcached’s simplistic in-memory key/value data structures make it a common target for academic works. Hundreds of papers either modify or compare with vanilla memcached. Here we analyze one of the earlier papers which made significant performance and memory efficiency modifications to memcached. We focus on how the paper applies to real-world industry and what it provided at the time of original publication. https://memcached.org/blog/persistent-memory-2/ Wed, 08 May 2019 02:50:21 -0700 https://memcached.org/blog/persistent-memory-2/ The Volatile Benefit of Persistent Memory: Part Two - Dormando (May 8, 2019) In part one we thoroughly explored the performance and configurations of Intel® Optane™ DC Persistent Memory Modules (referred to here as PMEM). During this process we ran tests and worked through thought experiments on how to best work with PMEM modules and memcached’s volatile cache. Given this entirely new (to us) hardware platform, we hoped to find new directions to explore. https://memcached.org/blog/persistent-memory/ Tue, 30 Apr 2019 08:32:41 -0700 https://memcached.org/blog/persistent-memory/ The Volatile Benefit of Persistent Memory - Dormando (April 30, 2019) Non-volatile memory modules and persistent memory are poised to be the next big thing for datacenter computing. With high storage densities and DRAM-like performance, with a DRAM-like interface, entirely new algorithms and systems architectures are possible. This is important for memcached, a high speed key/value cache service which is reliant on DRAM to meet its performance guarantees. https://memcached.org/blog/nvm-multidisk/ Mon, 04 Feb 2019 10:30:00 -0700 https://memcached.org/blog/nvm-multidisk/ Caching beyond RAM: Riding the cliff - Dormando (February 04, 2019) Key value stores are no longer confined to caching small objects over a local network. memcached is today deployed in many different environments, including geographically diverse datacenters, data warehouses and machine learning training systems. These new layouts have items sized kilobytes to megabytes, giving value to stitching devices together or tiering devices of different speeds. Our previous post introduced Extstore, an extension to memcached which allows moving cache data onto flash storage. https://memcached.org/blog/modern-lru/ Mon, 15 Oct 2018 11:00:57 -0700 https://memcached.org/blog/modern-lru/ Replacing the cache replacement algorithm in memcached - Dormando (October 15, 2018) In this post we delve into a reworking of memcached’s Least Recently Used (LRU) algorithm which was made default when 1.5.0 was released. Most of these features have been available via the “-o modern” switch for years. The 1.5.x series has enabled them all to work in concert to reduce RAM requirements. When memcached was first deployed, it was typically co-located on backend web servers, using spare RAM and CPU cycles. https://memcached.org/blog/extstore-cloud/ Wed, 15 Aug 2018 11:42:09 -0700 https://memcached.org/blog/extstore-cloud/ Extstore In The Cloud - Dormando (August 15, 2018) In our introductory post about memcached external storage, we talk about persistent memory, expensive NVMe devices, and terabytes of cache. Is extstore relevant to commodity VMs-with-SSDs from various cloud vendors? In this post we answer this by installing and configuring extstore on a cheap DigitalOcean VM. Use Case: Session caching For better or worse, a large use case for key/value stores in the cloud are session caches. https://memcached.org/blog/nvm-caching/ Tue, 12 Jun 2018 00:00:00 +0000 https://memcached.org/blog/nvm-caching/ Caching beyond RAM: the case for NVMe - Dormando (June 12th, 2018) Caching architectures at every layer of the stack embody an implicit tradeoff between performance and cost. These tradeoffs however are constantly shifting: new inflection points can emerge alongside advances in storage technology, changes in workload patterns, or fluctuations in hardware supply and demand. In this post we explore the design ramifications of the increasing cost of RAM on caching systems.