Compare Tair (Enterprise Edition) and Redis Open-Source Edition across performance, cost, data structures, persistence, security, and advanced features to select the series that fits your workload.
Selection guide
Tair (Enterprise Edition) comes in three series, each with a different balance of performance, persistence, and cost. Redis Open-Source Edition is included for reference.
Choose your series based on your primary requirement:
-
Maximum throughput, enterprise features → Memory-optimized: up to 3× the performance of Redis Open-Source Edition, full Tair extended data structures, TLS and TDE encryption.
-
Cost savings with near-standard performance and command-level durability → Persistent memory: ~30% lower cost, 90% of Redis Open-Source Edition performance, per-command persistence.
-
Large datasets at low cost, where performance is secondary → Disk-based: up to 85% cost reduction, hundreds of terabytes of capacity.
-
Standard Redis compatibility and migration → Redis Open-Source Edition: no trade-offs, standard Redis behavior.
| Series | Summary | Key features | Best for |
|---|---|---|---|
| Memory-optimized | Highest throughput, full enterprise feature set |
|
Performance-centric, mission-critical workloads |
| Persistent memory | Near-Redis-OSE performance at ~30% lower cost, with command-level durability |
|
High-performance caching and storage where data durability matters and cost is a secondary concern |
| Disk-based | Up to 85% cost reduction, hundreds of terabytes of capacity |
|
Large-dataset storage with low access density and cost as the primary constraint |
| Redis Open-Source Edition | Standard Redis compatibility | Compatible with open source Redis. High performance. | Standard Redis workloads and migration scenarios |
For a detailed decision framework, see Product selection reference.
Feature comparison
In the following table, ✔ indicates that the feature is supported, and ✗ indicates that the feature is not supported.
The table columns map to the following instance types:
| Column | Instance type |
|---|---|
| Memory-optimized | Tair (Enterprise Edition) — memory-optimized series |
| Persistent memory | Tair (Enterprise Edition) — persistent memory series |
| Disk-based (ESSD) | Tair (Enterprise Edition) — disk-based series, ESSD variant |
| Disk-based (SSD) | Tair (Enterprise Edition) — disk-based series, SSD variant |
| v2.8, 4.0, 5.0 | Redis Open-Source Edition versions 2.8, 4.0, and 5.0 |
| v6.0, 7.0 | Redis Open-Source Edition versions 6.0 and 7.0 |
| Redis for ARM | Redis Open-Source Edition — ARM architecture |
Baseline performance
| Metric | Memory-optimized | Persistent memory | Disk-based (ESSD) | Disk-based (SSD) | v2.8, 4.0, 5.0 | v6.0, 7.0 | Redis for ARM |
|---|---|---|---|---|---|---|---|
| Performance baseline (relative to Redis Open-Source Edition) | 300% | 90% | Read: 40%, Write: 30% | Read: 60%, Write: 40% | Same | 120% | 120% |
| Maximum connections per data node | 30,000 | 10,000 | 10,000 | 40,000 | 10,000 | 10,000 | 10,000 |
| Single-key QPS reference ① | 450,000 | 130,000 | 30,000–60,000 | 50,000–60,000 | 140,000 | 160,000 | 160,000 |
| Maximum bandwidth (MB/s) | 96–2,048 | 96–2,048 | 187.5–1,000 | 187.5–2,048 | 10–2,048 | 48–2,048 | 96–2,048 |
Instance attributes
| Metric | Memory-optimized | Persistent memory | Disk-based (ESSD) | Disk-based (SSD) | v2.8, 4.0, 5.0 | v6.0, 7.0 | Redis for ARM |
|---|---|---|---|---|---|---|---|
| I/O and worker model | Multi-I/O (Real Multi-I/O) ③ | Single-threaded | Multi-I/O + Multi-worker (Real Multi-I/O) | Multi-I/O + Multi-worker (Real Multi-I/O) | Single-threaded | Single-threaded | Single-threaded |
| Cost per unit (relative to Redis Open-Source Edition) | 117% | 70% | 15%–20% | 15% | Same | Same | 51%–67% |
Data structures
All Tair (Enterprise Edition) series support basic Redis data structures and commands, though the supported command set varies by edition — see Tair (Enterprise Edition) command support and limitations. Redis Open-Source Edition series also have unsupported commands — see Redis Open-Source Edition command support.
| Feature | Memory-optimized | Persistent memory | Disk-based (ESSD) | Disk-based (SSD) | v2.8, 4.0, 5.0 | v6.0, 7.0 | Redis for ARM |
|---|---|---|---|---|---|---|---|
| Tair extended data structures | ✔ | ✔ (Partial) | ✗ | ✗ | ✗ | ✗ | ✗ |
Disk persistence
Disk persistence consistency ④| Metric | Memory-optimized | Persistent memory | Disk-based (ESSD) | Disk-based (SSD) | v2.8, 4.0, 5.0 | v6.0, 7.0 | Redis for ARM |
|---|---|---|---|---|---|---|---|
| Primary/replica consistency | Eventual consistency | Eventual consistency | Eventual consistency | Eventual consistency | Eventual consistency | Eventual consistency | Eventual consistency |
| Write Back | Write Through | Write Through | Write Through | Write Back | Write Back | Write Back | |
| Persistence level | Second-level | Command-level | Command-level | Command-level | Second-level | Second-level | Second-level |
Security
| Feature | Memory-optimized | Persistent memory | Disk-based (ESSD) | Disk-based (SSD) | v2.8, 4.0, 5.0 | v6.0, 7.0 | Redis for ARM |
|---|---|---|---|---|---|---|---|
| TLS encryption | ✔ | ✔ | ✗ | ✗ | ✔ | ✔ | ✔ |
| TDE (Transparent Data Encryption) | ✔ | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ |
| IP whitelist | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ |
Performance analysis
| Feature | Memory-optimized | Persistent memory | Disk-based (ESSD) | Disk-based (SSD) | v2.8, 4.0, 5.0 | v6.0, 7.0 | Redis for ARM |
|---|---|---|---|---|---|---|---|
| Top key statistics | ✔ | ✔ | ✗ | ✗ | ✔ | ✔ | ✔ |
| Query historical hot spot keys | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ |
| Offline analysis of large keys | ✔ | ✔ | ✗ | ✗ | ✔ | ✔ | ✔ |
| Audit log | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ |
Advanced features
| Feature | Memory-optimized | Persistent memory | Disk-based (ESSD) | Disk-based (SSD) | v2.8, 4.0, 5.0 | v6.0, 7.0 | Redis for ARM |
|---|---|---|---|---|---|---|---|
| Point-in-time restore with data flashback | ✔ | ✔ | ✗ | ✔ | ✗ | ✗ | ✗ |
| Semi-synchronous mode | ✔ | ✗ | ✔ | ✔ | ✗ | ✗ | ✗ |
| Proxy query cache | ✔ | ✔ | ✗ | ✗ | ✗ | ✗ | ✗ |
| Global Distributed Cache | ✔ | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ |
| DTS one-way synchronization | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ |
| DTS two-way synchronization | ✔ | ✔ | ✔ | ✗ | ✗ | ✗ | ✗ |
Footnotes
-
① QPS is measured using O(1) commands. Higher time-complexity commands yield lower QPS.
-
② Performance depends on the distribution of hot and cold data access. A higher memory hit ratio brings performance closer to the Redis Open-Source Edition baseline.
-
③ Unlike the I/O multi-threading in Redis 6.0, the Real Multi-I/O model of the memory-optimized series fully accelerates both I/O and command execution. It provides stronger resistance to connection spikes and linearly scales throughput.
-
④ Two persistence strategies are used across series:
-
Write Through: The operation returns only after data is written and synchronously persisted to disk.
-
Write Back: The operation returns as soon as data is written to memory; data is then flushed to disk asynchronously.
-