FoundationDB ACID-lovers price up NoSQL database
Unorthodox tech gets free and paid-for versions
NoSQL database startup FoundationDB has made its ACID-compliant tech generally available, after an extended beta that has seen over 2,000 people try out the company's unorthodox database.
The FoundationDB database is a key-value store that also allows for different data models – such as JSON documents, graphs, and SQL (via technology from recently acquired Akiban) – to live on top of the NoSQL system.
It claims to be both ACID-compliant – a database term meaning it supports transactions that have Atomicity, Consistency, Isolation, and Durability – while having the inherent scale-out capabilities of a NoSQL system. This claim has been met with skepticism and curiosity by the database community, as attaining both capabilities is assumed to be very, very tricky.
On Tuesday, the company announced its tech was generally available, priced up a commercial version, and also launched a free offer as it sought to generate enthusiasm and cash for its novel approach to data management and manipulation.
The free "Community Edition" offers the performance and capabilities of the paid-for version but limits deployment to clusters of up to 6 processes per month.
"A FoundationDB process is a single worker process you run on a computer," Rosenthal says. "It utilizes a single core of processing power but can use all the network capacity of the card, all the IO capacity of the hard disk."
The community version "will be available to use without giving us an email address," Rosenthal said, which may relieve devs deluged by pitch emails.
For the commercial version, the starting price is $99 per process, per month with 8-5 support, going up to $199 per process, per month for the platinum offering, which offers "concierge" support and a guaranteed response time of 2 hours or less.
Known limitations of the system include a transaction limit of 10,000,000 bytes, key size limit of 10,000 bytes, and value limit of 100,000 bytes. For the time being, the tech also has trouble dealing with clusters of more than 96 cores, and the company is not quite sure what will happen when the database goes over 100TB in size. ®