Best in-memory scale out engine on the market for wealthy customers
December 04, 2020

Best in-memory scale out engine on the market for wealthy customers

Anonymous | TrustRadius Reviewer
Score 8 out of 10
Vetted Review
Verified User

Overall Satisfaction with SingleStore (formerly MemSQL)

We use [SingleStore] [(formerly MemSQL)] to address the ingestion of fast data streams and to answer analytical queries with tight timeout requirement.
  • Fast analytical queries with in-mem and on-disk columnar store.
  • Scale out capability
  • Good dashboard for managing the cluster
  • Licencing model needs to be more flexible, users need to be able to do pay-as-you-go
  • Data ingestion is still old-style. Integration with e.g., spark lagged behind.
  • In-mem and col-store does not share the same language compat. Table type transition is hard.
  • HA is no good with 6.0. Recovery time is very long. (and non-predictable)
  • Positive: Suites / overkills most projects at the beginning level.
  • Negative: Cost is high, under pressure of optimize app. and move away from memsql.
The in-memory tech is the best in class. If perf. is the hard limit then this is the No.1 to go.
When data scales up, the cost of maintaining a big cluster with memory keep increasing (both the infrastructure cost and licensing cost) becomes high. Also the balance between using memory for table-store and query execution is hard to achieve with limited budget.
This allows streaming real-time analytics to happen. Overall [SingleStore (formerly MemSQL)] gives very good throughput and above average query latency. When it is used in a multi-tenant environment. it becomes less easy to make sure workloads don't interfere each other.

Very good for small-sized dedicated system that requires very fast response.
Not much. The licensing model limits the "scaling" operation. Basically once you buy the capacity you will always use to that limit even at the beginning you don't need (as you already have to pay for it). And when you need more later, The option of scaling it out on demand is stopped by the license. E.g., I cannot scale out to 150% use on "busy days" while in the rest of the time using 75% of the capacity.

Do you think SingleStore delivers good value for the price?

No

Are you happy with SingleStore's feature set?

Yes

Did SingleStore live up to sales and marketing promises?

No

Did implementation of SingleStore go as expected?

Yes

Would you buy SingleStore again?

Yes

[SingleStore (formerly MemSQL)] is a good OLAP scale-out system with mysql compat. for rich customers. Suitable for fast analytical queries.
The single-main arch. (with read-only backup) limits it's OLTP concurrency as well as the OLAP query throughput.

Good to build small to medium sized (near) real-time analytical systems with. The cost and inflexible licensing model limits large scaled use.
Technically the best in-mem distributed RDBMS in class.

Management-wise although the dashboard and studio is a good start, to make it a 24-7 reliable database system there is still room to improve.
HA cost is very high as it also eats the licensed mem-space. The costly mem-based table and free on-disk columnar tables don't automatically play will together. E.g., auto offloading old data from in-men to on-disk is not supported, making it hard for DBA to manage the cluster.