Check out how ScaleOut In-Memory Database running open-source Redis compares with other implementations of Redis:
| Features | ScaleOut In-Memory Database | Other Redis Implementations |
|---|---|---|
| Clustering Servers | “One button scaling; up or down:” Just connect a server to the cluster; it automatically joins the membership, creates replicas, and rebalances the load. Same for leaving a server. | Clustering requires that you manually create and configure shards on each server and distribute hash slots. This process is complicated, time consuming, and error prone. |
| # of Servers Required | Servers host both primary and replica objects within a single service process. This minimizes the number of servers needed. | Redis recommends separate master and replica servers for high availability. That requires more servers for a given data set. |
| Failure / Recovery Process | After a server or network failure, ScaleOut automatically promotes replica objects to primaries AND creates new replicas to “self-heal.” The cluster automatically re-balances the workload. No manual intervention is required. | After a failure, Redis promotes replica objects to primaries but does not create new replicas. Recovery requires manual intervention to restore replicas. If both a primary shard and its replica fail, the entire cluster becomes inoperable. |
| Data Consistency | Provides full (strong) data consistency with patented technology; ScaleOut never serves stale data. | Redis uses eventual data consistency to updates replicas and can serve stale data to applications. |
| Multi-Threading | Runs Redis commands using multi-threaded processing to automatically take advantage of all processing cores using one process per server. | Redis uses single-threading and runs one command at a time. It requires you to license multiple shards per server to use additional cores. |
| Supported OS | Runs natively on Linux or Windows. OSs can be mixed in a single cluster. | Redis has no native Windows support and runs select versions of Linux. |
| Backup / Restore | “One button backup and restore:” Provides fully parallel backup and restore. Allows a different cluster configuration when restoring data. | Users must separately save each node’s database to disk and must restore them to exactly the original cluster configuration. |
Many of the limitations described above apply to both open-source Redis and Redis Enterprise. For example, users need to manage shards and hashslots, and they need to deploy additional shards for hi-av and multi-core speedup. ScaleOut In-Memory Database automatically takes care of all this and delivers higher performance with lower TCO. In fact, ScaleOut In-Memory Database has up to 10X lower TCO than Redis Enterprise.