2023/05/12 - Engineering Blogs on Memcache
This newsletter collects 8 engineering blogs on Memcache. Memcache, also known as Memcached, is a high-performance distributed caching system that is often used to improve the performance of websites and web applications. It stores frequently used data, such as database query results, in memory, allowing applications to access the data more quickly.
[Pinterest] [2022/05/12] Improving Distributed Caching Performance and Efficiency at Pinterest
[Pinterest] [2020/12/10] Scaling Cache Infrastructure at Pinterest
[Netflix] [2018/12/04] Cache warming: Agility for a stateful service
[DoorDash] [2018/08/03] Avoiding Cache Stampede at DoorDash
[Netflix] [2018/07/12] Evolution of Application Data Caching : From RAM to SSD
[Facebook] [2014/12/15] Introducing mcrouter: A memcached protocol router for scaling memcached deployments
[Facebook] [2013/04/15] Scaling memcache at Facebook
[Netflix] [2013/02/25] Announcing EVCache: Distributed in-memory datastore for Cloud
[Pinterest] [2022/05/12] Improving Distributed Caching Performance and Efficiency at Pinterest
TL; DR
The blog shared the observability and performance testing tools that enable Cache optimization work. Then the blog did a deep dive into practical optimizations currently running in Pinterest’s production environment along dimensions of hardware selection strategy, compute efficiency, and networking performance.
[Pinterest] [2020/12/10] Scaling Cache Infrastructure at Pinterest
TL; DR
The blog explained how Pinterest scaled their cache infrastructure to handle increased traffic and data volume. They used a combination of sharding, replication, and load balancing techniques to improve performance and reliability. The blog also described how Pinterest implemented monitoring and alerting systems.
[Netflix] [2018/12/04] Cache warming: Agility for a stateful service
TL; DR
Problem. Netflix had to pay additional cost to scale the cache either due to storage or network demand. Netflix’s approach was to provision a new empty cache, dual-write to the existing cache and the new one, and then allow the data to expire in the older cluster after their Time to Live (TTL) value. This approach doubled the cost, and also didn’t suit well for clusters that had items with no expiration time or were not mutated in between.
Solution. The blog described that Netflix developed a EVCache cache warmer infrastructure with two capabilities, including replica warmer and instance warmer.
[DoorDash] [2018/08/03] Avoiding Cache Stampede at DoorDash
TL; DR
DoorDash used built-in constructs from Kotlin coroutines to solve the cache stampede (cache miss storm) problem that could be caused by a local cache hit miss.
[Netflix] [2018/07/12] Evolution of Application Data Caching : From RAM to SSD
TL; DR
Problem. As Netflix started moving towards the Global Cloud, Netflix had a three times increase in the data that needed to be replicated and cached in each region. Netflix also needed to move this data swiftly and securely across all regions. Supporting these features came at a considerable increase in cost and complexities.
Solution. Netflix provided a global caching solution which was not only fast but was also cost effective.
[Facebook] [2014/12/15] Introducing mcrouter: A memcached protocol router for scaling memcached deployments
TL; DR
This blog introduced mcrouter, a protocol router forwarding requests between clients and Memcached server. Mcrouter provides features such as dynamic configuration updates and failover support.
[Facebook] [2013/04/15] Scaling memcache at Facebook
TL; DR
This blog explains how Facebook scaled their Memcached deployment to handle increased traffic and data volume in the social network.
[Netflix] [2013/02/25] Announcing EVCache: Distributed in-memory datastore for Cloud
TL; DR
Netflix's engineering team introduced Evcache, a distributed in-memory datastore for the cloud on top of Memcached. It is designed to provide high-performance caching for frequently accessed data, such as metadata and session state. The new system improves the efficiency and reliability of Netflix's caching infrastructure.
If you have any questions or comments, please contact plussmart2018@gmail.com.