Redis Enterprise Software
Mongo DB Enterprise
Mongo DB Enterprise
Datastax Enterprise
Datastax Enterprise
Data Model Data Structure Store, Key-Value, Document, Graph, Indexing Document, Indexing Column, Graph, Indexing
Object size Unlimited (each element in the object is limited to 512MB) 16MB 2GB
Strong Consistency and Eventual Consistency Yes Yes Yes
CRDT-based Strong Eventual Consistency Yes
Replication Model Master- Slave and Peer to Peer Master-Slave Peer to Peer
Tunable Durability Yes Yes Yes
Servers needed for 1 million writes/sec 2 1000 1 300
Latency characteristics for above 1 million writes/sec <1 ms 280 ms 1 >20ms
Consistent reads/writes <1 ms Not benchmarked >100 ms 2
Inconsistent reads/writes <1 ms 280 ms 1 >20 ms
Needs a caching layer No Yes Yes
Linear scaling Yes
Topologies available Active-Active and Active-Passive Active-Passive only Active-Active only
WAN latencies for reads and writes No Yes Yes
Conflict resolution method CRDT-based N/A Last writer wins (LWW)
Conflict resolution in simple use cases Yes N/A Yes
Conflict resolution in complex use cases Yes N/A Requires application intervention or data is lost
Failover time Single digit seconds Double digit seconds Immediate, but inconsistent views
Scaling Easy and seamless Complex Complex
Cost for 1 million writes/second $14,832 Not benchmarked $2,226,216
DBaaS on public cloud (hosted) Yes Yes Yes
Serverless DBaaS Yes
DBaaS in customer's VPC Yes
On-prem Yes Yes Yes

1 Estimated based on Benchmark | 2 Based on Benchmark

Key Differentiators

Redis Enterprise provides orders of magnitude higher performance than both Datastax and MongoDB, making the costs associated with scaling to high throughput and latency practically disappear when used instead of both those databases. With its unique Active-Active geographic distribution based on CRDTs, Redis Enterprise brings added advantages of:

-Simultaneous reads and writes to the same shared, distributed dataset with built-in smart conflict-resolution

-Assured local latencies regardless of the number of replicas associated with the dataset, since conflicts are handled in a consensus-free manner

-Assured availability regardless of how many other replicas are available (i.e. no need for quorum).

Add in the development simplicity and versatile data structures of Redis Enterprise, and the choice is even easier.