SingleStore is a hybrid transactions/analytics processing (HTAP) database. In order to support both use cases, it makes compromises that hinder performance, ease of use and cost. Rockset, on the other hand, is purpose-built for real-time analytics. It’s both cloud native and fully managed, offering unmatched price-performance.
SingleStore is not designed for high-volume writes.
With SingleStore, streaming data and queries share the same compute, causing compute contention and application slowdown.
SingleStore is immutable.
SingleStore’s columnar store is immutable, requiring compute-intensive copy-on-writes to handle updates and deletes. When handling upserts, ingestion and queries are blocked until a merge operation completes.
SingleStore is designed for structured data and static schemas.
SingleStore’s JSON column type is inefficient for querying semi-structured data, requiring the construction and management of pipelines to flatten nested JSON before ingestion. SingleStore requires ALTER TABLE commands to make schema changes, adding latency.
SingleStore requires index management to achieve good query performance.
SingleStore requires configuring and managing indexes to achieve optimal query performance. Too many indexes cause write contention. Too few indexes slows down query performance. Adding and removing indexes is an operational burden.
Performance for Large Working Datasets
SingleStore is an in-memory datastore.
SingleStore caches hot data for queries and moves cold data to blob storage, causing unpredictable performance when working with large-scale data. When scaling compute, SingleStore takes time to move data from blob storage to the cache, adding significant latency.
If you want flexible and easy real-time analytics, check out Rockset.
Here are four reasons why:
Creation of search, columnar and row indexes at ingest time
SQL search, aggregations and joins on semi-structured data
Efficient inserts, updates and deletes
Independent scaling of storage and compute in the cloud