This article is more than 1 year old

MongoDB launches Atlas to manage deployments: Taking the Ops out of DevOps

'The most interesting thing we've done as a company since day one,' says exec

MongoDB is launching Atlas, the company's first DBaaS, offering easy management of instances - initially on AWS, but soon to come to Azure and Google Cloud Platform.

The service intends to alleviate devs of tedious evenings wasted with hardware provisioning, covering failure recovery, software patching, upgrades, configuration and backups.

Kelly Stirman, the company's strategy veep, explained to The Register that the user experience was receiving a single bill from MongoDB “while we handle the infrastructure, the upgrades, the security. You get to decide what region, and how big a server you want – and if you need change in that size you just click a button and our engineers sort it out.”

Atlas is pitched as “unlimited, elastic scalability, either by scaling up on a range of instance sizes or scaling out with automatic sharing, all with no application downtime.”

The company believes that the majority of MongoDB's use is in the cloud, and the plan is that Atlas is the best way to manage that. “There are millions of Mongo instances,” said Stirman, “this is a huge opportunity for us.”

“This is a very big strategic product release for us,” Stirman continued, claiming it was “the most interesting thing we've done as a company since day one.”

Stirman accepted the concern that consumers might have with the product being newly brought to market, but said: “Step back and look at how we got to this point,” said Stirman, explaining that MongoDB launched its first cloud product four years ago, before releasing a backup service a year later. “Just over a year ago we released the software for automated upgrades. This builds on those core products and packages with a way of handling the infrastructure.”

While not a “brand new suit,” Kelly thought Atlas' infrastructure offering was the kicker. “We automatically deploy your cluster across three availability zones in an Amazon region, that means that if that availability zone fails or gets rebooted – as happens time to time in AWS – you are still running in at least two other availability zones.”

The product is not aimed at the enterprise market, however. Atlas will initially support MongoDB community server, with the hope at some point in the future on introducing the enterprise model, although Kelly was unable to tell The Register when this would be.

Community Server is MongoDB's open-source product, which is available for download from its website, while it also offers a proprietary version of the software creatively dubbed Enterprise Server. At the moment Atlas will only be available for Community Server users. Stirman explained that most Enterprise Server users didn't need it as they were already "running in the cloud or building their own infrastructure, to build out their own private DbaaS."

Eventually Enterprise Server will be available on Atlas, but that, according to Stirman, “is a function of time. We want to get to market with Community Server first.”

Atlas will be pay-by-hour, and there will be seven different instance sizes, from the low-end offering 1GB of RAM and 10GB storage, on a single core v-cpu, which will cost 2.6 US cents per hour, through to the high end of 160GB RAM, 1TB of storage, and 40-core v-cpu, at $3 per hour.

The first target market for Atlas use would be those already using MongoDB as a service. “There are a few around,” said Stirman, mentioning IBM.

The other primary case, “contingency”, was for former users of Parse, the Facebook-acquired and then shuttered development platform. “There are potentially north of a million deployments using a potentially similar service, and all of those applications need to find a home,” said Stirman.

Then there are those running MongoDB in the cloud already, “maybe managing it themselves, who want us to take care of it for the inconvenience,” added Stirman. Lastly, there are people running on-premises who are not-yet on cloud and “just want to use the best service already”.”

MongoDB has also announced a new Spark connector so devs and data scientists can bang out some neat tricks on live, operational, and streaming data, it said. ®

More about

More about

More about

TIP US OFF

Send us news


Other stories you might like