Many NoSQL databases are rather simple concepts compared to traditional databases. With NoSQL databases you typically don't pay to run an instance. You only pay for what you use or provision.

Everything is stored in the NoSQL DBs. No information resides permanently in the Columno program. This makes Columno more fault-tolerant and scalable.

Columno is written in C++ and we currently build releases for Windows, Linux and macOS.

Many modern scalable databases exist, often categorized as NoSQL or NewSQL. Most of them are awesome. Columno is a different and novel approach — we seamlessly let you choose where to store data (any Key-Value Store works). Our algorithms are build for few large read/writes so data can reside outside the instances — even in a different cloud environment.

The transformation from relational data into a key-value store doesn't come for free. But key-value stores are fast, very fast, and so are our algorithms — and with a lot of effort we can get extremely good performance.

Are you (and your colleagues — and the one taking over from you ) experts in Key-Value Stores/NoSQL DBs and do you know how to easily map all of your data and objects so it's all efficient and scalable — then Columno is a waste of your time. But give us a call when you don't have the time anymore :)

It is never a good idea to have tightly coupled components — especially not in a Cloud Native Server-less world. It makes architectures complicated, expensive and hard to maintain. SQL/NewSQL all have tightly coupled storage layers that are close to their processing algorithms (on the same instance). Columno runs completely seperated from the storage layer — no algorithms reside near the storage layer (simple Key-Value Store). This enables clients to choose any server-less storage provider for their database.

Our algorithms are built from the ground up to enable safe intensive multi-threading and multi-instance workloads. We have local mutexes to guarantee ACID and our storage model is built for concurrency. We can support (and have tested) thousands of concurrent users/threads.

NoSQL (or better: NoREL) was pretty hot 10 years ago when people wanted scalability and distributed data. However, SQL (and relations) is here to stay — just look at all the new databases (NewSQL), they all want to fuse SQL, relations, scalability and distributed data.