High concurrency major electricity supplier details page system of large-scale high performance and high availability cache architecture
1, one hundred million electricity supplier website traffic product details page System Architecture
Faced with problems: traffic every day to hundreds of millions, with hundreds of millions of pages of major electricity supplier site, the ability to support high concurrent access, while being able to make second-level system architecture product details page of the entry into force of the latest template design is how?
Solution: Asynchronous multi-level cache architecture + nginx + dynamic localization template cache rendering architecture
2, redis enterprise-class cluster architecture
Faced with challenges: how to make hundreds of thousands of QPS redis clusters support high concurrent availability + 99.99% + TB + class mass data enterprise-class data backup and recovery?
Solution: redis enterprise-class backup and recovery program replication architecture + + + separate read and write to deploy Sentinel architecture + redis cluster cluster
3, multi-level cache architecture
Faced with challenges: how to design a caching architecture can support high performance and high concurrent to the extreme? But also to the cache architecture of a final layer of security?
Solution: nginx + anti-hot data redis large-scale anti-offline request + three cache architecture ehcache anti-redis crash
4, double write consistency cache database + Solutions
Faced with problems: high concurrency scenarios, how to solve the inconsistent database cache double the time to write data situation?
Solution: asynchronous serial queue of dual-write consistency cache database + Solutions
5, the dimensions of the split cache solutions
Faced with challenges: how to solve the great value low volume full buffer update efficiency?
Solution: The dimension of the split cache data product solutions
6, the cache hit rate increased Solution
Faced with challenges: how to improve the cache hit rate to the extreme?
Solution: Double nginx + lua script deployment architecture to achieve consistency hash traffic distribution policy
7, rebuild the cache concurrency conflict resolution
Faced with challenges: how to solve under high concurrency scenarios, when distributed cache concurrency conflict reconstruction reconstruction?
Solution: Based Clash zookeeper distributed lock the cache concurrency solutions
8, cache warming solutions
Faced with challenges: how to address the high concurrency scenarios, the cache cold start problems caused MySQL load is too high, even instantly killed?
Solution: Distributed cache storm fast real-time statistical data preheat thermal solutions
9, hot caching scheme automatically downgrade
Faced with challenges: how to solve the hot cache results in a single ultra-high machine load moment?
Solution: Based on the real-time hot spots storm discovery + millisecond real-time load balancing hot cache downgrade
10, high-availability distributed system architecture design
Faced with challenges: how to solve the problem of high availability services in a distributed system? Avoid multi-tier service because a small amount dependent failure caused a system crash?
Solution: Based hystrix high availability caching services, resource isolation + current-limiting fuse + + + downgrade timeout control
11, complex high-availability distributed system architecture design
Faced with challenges: how for complex distributed systems will be one of the high availability architecture is designed to serve?
Solution: hystrix based fault tolerance of multi-level downgrade + + + manual downgrade parameter optimization of the production environment + visual experience with the operation and maintenance monitoring
12, caching solutions avalanche
Faced with challenges: how to solve the problem of cache avalanche of terror? Avoid the company has brought huge economic losses?
Solution: the whole network exclusive things in advance + + after the three-level cache avalanche perfect solution
13, caching solutions to penetrate
Faced with challenges: how to solve the problem of penetrating the buffer at high concurrency scenarios? MySQL to avoid bringing too much pressure?
Solution: Cache penetrate Solutions
14, cache invalidation Solutions
Faced with challenges: how to solve the cache under high concurrency scenarios failures? Redis clusters to avoid bringing too much pressure?
Solution: random based cache expiration time of failure Solutions