Redis "Sharding" Tech Scales NoSQL Performance To Over 1M TPS

Commercial database-as-a-service provider Redis Labs has released new sharding technology that is fully compliant with open source Redis Cluster APIs. ProgrammableWeb spoke with CEO Ofer Bengal about the benefits of using the new features to enable more than a million requests per second.

Redis is an in-memory, open source NoSQL database and is used extensively to carry out high-speed transactions, real-time analytics, metrics and ranking, job management, and social capabilities. Redis Labs is a commercial Redis provider, offering the Redis Cloud database-as-a-service.

This week, Redis Labs launched sharding technology to enable developers to customize how their data stores are spread across multiple servers in order to overcome the limitations in the number of transactions typically able to be carried out within large data sets stored over multiple cores and cloud instances.

“Our new shards management technology allows users to shard their database and run each database to get up to 10 times more value. With this technology, you can get 1.2 million operations per second, more than 10 times regular Redis performance,” says Bengal.

Sharding is a technique that lets database administrators define how their large-scale data sets should be split across multiple servers. This is crucial for ensuring performance of APIs at scale, for example, when APIs may need to ensure that millions — perhaps billions — of calls can be transacted without lag. For example, transport, retail, ticketing, social media and increasingly Internet of Things sensor data all need to be able to power global transactions via API and sync data in real time.

“The secret here is the way to shard. The problem before is that sharding didn’t allow control,” explains Bengal. “The way to do it is to put all the keys required in a transaction on the same shard. With the technology we are announcing, the user can do all of that on our platform and very easily define how to shard the database. The way to define the sharding is with regular expression commands, which is very popular amongst developers. The user simply uses regex commands to instruct us to change the sharding.”

The new features are also fully compliant with Redis Cluster APIs. “With Redis Cluster, users can replicate the same database across data centers and across the cloud so that they can serve transactions from replicated databases without losing end data and being able to avoid downtime of any one server. Our feature works with the Redis Cluster API, so that users get an endpoint and can send the data to that endpoint. It is frictionless for the customer,” he says.

Redis Labs recently announced that its customer base extends beyond 3,000 businesses of all sizes. Redis Labs provides the power of Redis open source technology with scalability, availability and fully managed service agreements.

Mark Boyd is a ProgrammableWeb writer covering breaking news, API business strategies and models, open data, and smart cities.