Amazon’s new managed non-relational database service raises questions about the future of open source in the cloud era
Amazon Web Services (AWS) has launched a custom-built non-relational database designed to run MongoDB workloads in the cloud, in a move that raises fresh questions about the future of open source technologies in a world increasingly dominated by the cloud.
The launch of AWS’ DocumentDB comes three months after MongoDB shifted its software to a new licence designed to prevent cloud providers such as AWS from running managed versions of the technology without either making the source code of their service publicly available or purchasing a commercial licence .
DocumentDB uses version 3.6 of the MongoDB application programming interface (API) to interact with MongoDB clients.
That version, dating back to 2017, is covered by the open source Apache licence, a move intended to circumvent MongoDB’s new licensing structure, based on the specially created Server Side Public License (SSPL).
Open source ‘capture’
At the time of MongoDB’s licence change, MongoDB co-founder and chief technology officer Eliot Horowitz said that once an open source project became popular, it was “too easy for cloud vendors who have not developed the software to capture all of the value while contributing little back to the community”.
AWS launched a managed version of open source data streaming tool Apache Kafka about six weeks ago.
As with the Kafka offering, AWS said the new DocumentDB database is designed to allow organisations to run MongoDB workloads without the challenges of maintaining the database themselves — something also offered by MongoDB’s managed Atlas service.
Atlas accounted for 22 percent of MongoDB’s $65 million (£51m) in revenues during its third fiscal quarter, the company said in December.
AWS said DocumentDB would also give users access to the company’s popular back-end infrastructure.
“While other companies have taken the same emulation approach we have to providing a MongoDB compatible service, nobody has built the unique, distributed, fault-tolerant, highly scalable, self-healing storage system that AWS has to work seamlessly with MongoDB,” AWS vice president of non-relational databases Shawn Bice said in a statement.
Bice said AWS had built DocumentDB from scratch over the course of two years.
It wasn’t clear how AWS’ MongoDB emulation would work in practice, with MongoDB suggesting that as it was based on an older interface, the tool wouldn’t support all of the database’s present features.
It said DocumentDB lacks MongoDB features such as multi-document ACID transactions, change streams and global clusters.
“In order to give developers what they want, AWS has been pushed to offer an imitation MongoDB service that is based on the MongoDB code from two years ago,” Horowitz said in a statement.
AWS outlined the MongoDB data types and operations supported by DocumentDB in an online developer guide.