The pattern we are using now is for direct key value to be cached directly with memcached. Redis is used for collection types. For the content of the big data volume stuff, we are going to try to use MongoDB. Also learning neo4j to deal with deep search, recommended features.
1.Memcached single Key-value size Limited, one value maximum only 1MB, Redis's maximum supported 512mb2.memcached is just a memory cache, no requirement for reliability, and Redis prefers a memory database, so it is 3 more reliable. Essentially, memcached is just a single key-value memory cache, while Redis is a Data structure in memory database, support five types, so redis in addition to simple caching, but also can handle some simple logic operations, Redis can not only cache, And can also be used as a database with 4. The new version (3.0) of Redis refers to cluster distribution, which means that the cluster itself balances client requests, each node can communicate, can expand the line, maintainability more powerful.
Redis memcached MongoDB