Why MongoDB Struggles in China: Stuck Between Caching, OLTP, and OLAP
The article argues MongoDB's positioning is awkward: early versions lacked persistence and competed with Redis, but Redis now dominates caching with alternatives like TiKV and Hazelcast; MongoDB's JSON advantage is eroded by PostgreSQL and Redis ReJSON; for OLTP it cannot match MySQL-compatible distributed databases like TiDB; for OLAP it is outclassed by modern lakehouse engines such as ClickHouse and Doris.
MongoDB's market position in China is described as fundamentally awkward across multiple technology layers.
Early History: A Cache Contender Without Persistence
In its earliest versions (0.x), MongoDB had almost no persistence and was positioned directly against Redis, Memcached, and Tokyo Tyrant/Tokyo Cabinet as a caching solution.
Feature Expansion: Becoming a High-Capacity Distributed Database
Over time MongoDB added GridFS, transactions, and distributed multi-replica capabilities, evolving into a system that can function as a high-capacity distributed database.
Caching Domain: Redis Remains King; New Alternatives Emerge
In caching, Redis is still the absolute leader. For small-scale deployments (under 1,000 nodes) Redis Cluster suffices; larger scales use multi-level cache sharding middleware such as Camellia. One cited project runs 2,300 Redis nodes (64 GB each). Two more promising large-capacity distributed caching directions have appeared:
KV storage with automatic cold/hot separation and unlimited capacity: represented by TiKV, Apache Kvrocks, and Apache Pegasus.
In-memory grids based on distributed protocols: represented by Hazelcast and Apache Ignite, which are true distributed replacements for Redis-style caches.
MongoDB is not competitive in this space.
Document Database Domain: JSON Support Is Now Table Stakes
MongoDB's document model and BSON/JSON handling used to be a differentiator, but JSON operations are now baseline features. MySQL offers basic support, PostgreSQL provides significantly better JSON functionality, and Redis ReJSON is already widely adopted. MongoDB's advantage here is no longer obvious.
Distributed OLTP: No Clear Niche, No MySQL Compatibility
When positioned as a distributed OLTP database, MongoDB's schema-free design is a natural disadvantage for strong-transaction, constraint-heavy domains like finance or trading cores. It cannot beat MySQL-compatible distributed databases such as TiDB, OceanBase, PolarDB, TDSQL, GoldenDB, or GaussDB. Gaining an ecosystem foothold requires compatibility with MySQL, PostgreSQL, or Oracle — MongoDB has a JDBC driver but its SQL support is fragmentary and transactions are not a strength.
OLAP: Outclassed by Modern Lakehouse Engines
For analytical workloads, MongoDB does not integrate with the traditional Hadoop/HDFS/Hive or Spark/Flink stacks. In the modern lakehouse landscape it appears too simplistic compared to ClickHouse, Hudi, Iceberg, Paimon, Doris, and StarRocks.
Conclusion
Across both traditional and distributed systems, in caching and database roles, and for both OLTP and OLAP, MongoDB's positioning remains awkward. A Chinese vendor, SequoiaDB (巨杉), is pursuing MongoDB localization; its progress remains to be seen.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
