MongoDB finally takes cloud into its own hands with launch of Atlas

Derek du Preez Profile picture for user ddpreez June 27, 2016
NoSQL giant MongoDB has finally committed to a database-as-a-service offering of its own - Atlas. It made the announcement at it’s annual user conference in New York this week.

mongo db
MongoDB, one of the leading players in the NoSQL space, has announced the launch of its own cloud database-as-a-service offering at its annual user conference in New York this week.

The company hopes that its new Atlas offering will be able to woo customers away from other third party cloud providers that already offer MongoDB as-a-service. The assumption being that users would prefer to buy a service from the company that engineers it, rather than an alternative source.

We got to speak to MongoDB’s VP of Strategy, Kelly Stirman, ahead of the announcement to get his take on Atlas. He explained that the full database-as-a-service offering extends MongoDB’s existing cloud offering, which includes a cloud management tool that it offers as SaaS. Stirman said:

This is a further embracing and deepening of our cloud strategy. We believe that certainly more than half of MongoDB use is in the cloud. And if you think about that, what could we be doing to strengthen MongoDB’s position in the cloud and make it easier still and even more productive for users of MongoDB when they’re in that environment.

There is always going to be a contingency of folks that aren’t ready for the cloud or feel that their particular application isn’t right for the cloud, but t I think increasingly people will be in the cloud first and foremost.

What we did almost four years ago is that we started with a service to monitor to our customer’s infrastructure and we provided that in a SaaS model. Most people using that service were in the cloud. We have continued to build out that cloud family of products with a back-up service, people didn’t have to worry about their own back-ups and they could just pay by the gigabyte for a fully managed service that we provide. Then we added capabilities to install, upgrade and make configuration changes so they wouldn’t have to do it manually. That continues to build up and we call that product today Cloud Manager.

Having dabbled in the cloud with the Cloud Manager offering, Mongo is now ready to go full throttle with Atlas.

Stirman compared the database-as-a-service offering to Uber, where he said that customers are now willing to give up some control over the product in exchange for the convenience. But he recognises that some customers will still want to keep their use on-premise, whilst others will adopt a hybrid approach. But the key is that MongoBD now has the capability to offer all of this itself - rather than customers having to go to IBM, Rackspace or mLab. He said:

We think that our customers, some of them will just want to completely outsource the management of their database to MongoDB. Some are going to want to have total control over things themselves and use our cloud manager product. And some are going to want to take advantage of both.

The details

Kelly Stirman, MongoDB

MongoDB is making Atlas available via partnerships with the three largest cloud providers on the market - AWS, Google and Microsoft - as opposed to building out infrastructure on its own. The sell is that MongoDB Atlas is “operated by the experts who design and engineer the database” meaning that “developers no longer need to worry about operational tasks such as hardware provisioning, failure recovery, software patching, upgrades, configuration or backups”.

MongoDB adds that the service also offers unlimited elastic scalability, either by scaling up on a range of instance sizes or scaling out with automatic sharding, without any application downtime.

In terms of the detail, Atlas comes packaged as any other strong cloud offering - in that Mongo says it is designed for high availability, will have high performance for demanding workloads (claiming an 80 percent cost reduction compared to traditional enterprise database software), is secure and has a flexible pricing model.

But where does Mongo expect its customers to come from? If over half are already in the cloud, which ones will pick and switch to Atlas? Stirman envisages three different scenarios. He said:

We believe that there’s a first group of people that are already using MongoDB as-a-service (via a third party provider) and they will want to migrate to Atlas because it gives them advantages in terms of functionality and cost and they are going to want to use the service provided by the company that designs and engineers the database.

The second group are people that are running MongoDB in the cloud and they’re sort of doing it on their own, and they’d be really happy to have somebody do all the heavy lifting for them.

Then there is the third group of people that aren’t using MongoDB yet, and they’re thinking about where am I going to run this deployment and it may be their first step into the cloud. We have optimised so many different variables that they’re going to say ‘yeah, that’s the best way to run this thing’.

Key to all of this has been MongoDB’s partnerships with the largest cloud providers in the market - something that Stirman was keen to emphasise. Equally, the company is putting emphasis on its ability lift and shift from one cloud provider to another, something that can be technically difficult. He said:

We are not building data centres all over the world. We have strong relationships with Amazon, Google and Microsoft. We will launch day one with support for AWS regions in North America, Europe and Asia. In the coming months we will add support for additional regions on Amazon, as well as Azure and Google.

And part of what’s core to our strategy is that customers will be able to choose which of these underlying cloud providers they want to run on - because MongoDB is part of the full stack and they have to think about their application servers and the other services that they’re using from those cloud providers. We will give them a certain amount of independence and flexibility over the long haul.

Let’s face it, choosing a cloud provider is entering into a long-term relationship. Each of these vendors has a highly proprietary offering in the market, so you can’t jump from one to the other in the course of a weekend. The hardest thing to migrate is the database. So part of our strategy is giving you the flexibility to easily move from one cloud provider to another, as well as giving you an easy way to move to your own on-premise infrastructure, or your own private cloud.

The features we provide, the tooling we provide and our commercial model supports that flexibility and does not lock you into one or the other.

The market

I asked Stirman whether or not MongoDB has seen any evidence of database-as-a-service being more popular amongst certain industries or customer verticals, so that the company has a clear idea of where demand is likely to come from.

However, Stirman said that he was “surprised by the diversity”, as there has been demand from all over the place - including manufacturing, retail, healthcare and banking. However, he said that he believes that industries that have a demand for protecting personal data are actually more likely to end up in the cloud, despite popular conceptions that suggest otherwise. He said:

You think about healthcare and that data is so sensitive, and the same is true of retail - you’d think that people would be reluctant to go in the cloud. I firmly believe that most cloud environments are more secure than what people would be able to develop themselves, the features and capabilities are all there, people just need to take advantage to them appropriately.

It makes perfect sense to me that sensitive data like healthcare and retail would find its way into the cloud.

However, Stirman did say that the common use case is mobile development, as this naturally lends itself to cloud environments. He said:

The one thing you could say is that the place that a lot of people get started with cloud is in mobile applications. When you deploy a mobile app you’re probably creating something new that is mobile enabling a workflow that you already have in your business. It’s doing something that isn’t replacing something wholesale in many cases and the volume of users is probably going to be significantly greater than what you’re used to delivering out of your existing apps.

The need to deliver to a global audience - you don’t exactly know where your users are going to be - and I think for those reasons people look at mobile apps as a way to try the cloud.

Finally, I wanted to get a sense from Stirman of what he believes the main challenges for MongoDB will be

Freedom to communicate and work with cloud technology © Romolo Tavani -
(© Romolo Tavani -
in relation to Atlas. He said that as Atlas scales up, the company will have to keep providing the technology in regions that suit the customer - but that this is likely to be more a case of timing and growing pains. Stirman said:

I think the challenge initially for some folks initially is that we don’t have the service available in a region that we want. For example, at the time of launch, in EMEA we will be in Dublin. And clearly for folks in Central Europe, the preferred region would be Frankfurt. For them the latencies associated with using a Dublin data centre versus a closer data centre in Frankfurt are going to be limiting for them.

I think the main challenge for us initially is just going to be the availability of the service to some users. But that is something we will be rolling out this year, broader coverage.

Frankly there are plenty of companies that have been successful in the market just with the data centres we are supporting at launch, so we have plenty of market opportunity to get this thing off the ground and be hugely successful over the summer and in the fall. But these other data centres are the next key thing we need to start delivering on.

My take

This makes sense for Mongo - it’s clear that there is demand for its product in the cloud (which is only going to grow), so why would it let third party providers control all that customer growth? It makes far more sense for Mongo to provide its services directly, where it not only has easier control over the revenue prospects, but it also can get insight into how its customers are using the product for future development needs.

Equally, as we know, Mongo has launched a pretty serious consulting business, which can now push the cloud agenda.

Challenges will come in attracting customers away from those already established offerings, but as Mongo highlights, there are advantages in being close to the source provider.

It also needs to realise that as a cloud offering, Mongo as a database is becoming increasingly commoditised. Where to next? If it wants to keep delivering business value and not become the next infrastructure commodity, Mongo needs to be thinking about the business outcomes further up the stack - not just integrations with other open source tech.

More news to come this week from MongoDB world, so keep tuned for further updates.

A grey colored placeholder image