MongoDB forges ahead with multi-cloud agenda - announces multi-cloud clusters for Atlas

Leading noSQL database provider MongoDB today announced the general availability of multi-cloud clusters for its cloud database, Atlas. MongoDB claims that for the first time, customers will be able to deploy a fully managed, distributed database across AWS, Google Cloud and Azure simultaneously - which will allow for running applications across multiple cloud providers, without having to manage data replication and migration across platforms. 

MongoDB has made it clear over the past few years that it sees multi-cloud as the future of modern digital architectures for buyers, with its cloud platform Atlas already offering features such as integrated billing and customer support across GCP, Azure and AWS. 

The announcement today claims that multi-cloud clusters will allow organisations to extend the geographic reach of their applications by allowing them to replicate data across all three major cloud providers, to any of the 79 supported cloud regions worldwide. In addition, it will allow buyers to migrate their data from one cloud provider to another, to meet any changing requirements of their application or business. 

On paper, this should allow developers using MongoDB Atlas multi-cloud clusters to make use of the offerings of their choosing from each cloud provider. For instance, an organisation running its applications on AWS should be able to make use of Google's AI/ML services on the operational data, without the complexity of moving the data itself. 

In discussion with MongoDB

We got the chance to speak to Mat Keep, product director at MongoDB, about the latest Atlas release, where he explained that the demand for multi-cloud flexibility is being driven by senior buyers. He said: 

We see MongoDB customers embracing a multi-cloud strategy, for a variety of reasons. Fear of lock in is one. And when you get to the CTO level, a lot of them have been spending the past 10 years trying to unlock themselves from proprietary on premise software. And you know, all the pernicious licensing and audits. But what happens is that they modernise and move to the cloud, exchanging one level of lock-in for another. And so having an optionality and a portability is important to them. 

For example, Keep said that he has been working with a major automotive manufacturer that decided as a corporate mandate to go all in on GCP. However, developers being developers, there was a lot of continued use of AWS too. In addition to this, the company has a lot of product development in China, which is done on-premise in a private cloud. Multi-cloud clusters will give such customers increased flexibility, he said: 

So, having that optionality to be able to pick up an application and its database and move it from on-prem to different clouds and back again, without changing a single line of code that talks to the database, was huge for them. 

It also gives developers the opportunity to use whichever services makes the most sense for that particular part of their application. I think being able to have your e-commerce application on top of AWS infrastructure, but then seamlessly piping data into Google to do analysis of customer browsing habits, was something that you could do before, but it was a lot of additional work. We didn't natively support the nodes in your clusters spread across multiple different cloud providers. Now you can, so it gives you that kind of freedom to use the best service in any cloud. 

In addition to increased flexibility and having the option to more easily take advantage of multiple services across different cloud platforms, Keep argues that multi-cloud clusters will also be a boost for cloud resiliency - particularly for those customers that are focused on data sovereignty. He explained: 

The other big driver is in-region resilience, while maintaining data sovereignty. If you look at AWS, for example, it has one region in the UK, which is London. Now if you have to keep data in the UK based on privacy laws, there's nowhere else to go, especially with us leaving the EU. So if AWS goes down you've lost everything. 

Whereas, Google has a region in the UK, Azure has a region in the UK, so you can start to create that cross-cloud resilience that you didn't have. So being able to span clouds within a region gives you the ability to maintain data sovereignty, whilst also ensuring that should that region suffer some sort of issue, which does happen, is a big factor. 

In terms of incoming demand for the new Atlas capabilities, Keep said that there has been a high expression of interest from C-level executives across a number of digital boardroom events that MongoDB has been running. He said: 

I would say probably 80% expressed interest in having this multi-cloud freedom. Does that mean that they're going to deploy that way on day one? Probably not. But having that capability, having that freedom, I think is incredibly important for a lot of customers. Do I think 80% of Atlas users will take advantage of this? No. But I think for a certain class it's going to be very useful for them, I think it's going to be pretty sizable. 

My take

This announcement marks the culmination of a number of years of work for MongoDB, towards its ambition of being a true general purpose database for multi-cloud environments. Anything that offers customers increased flexibility and choice, is a good thing. This too has been a mantra of MongoDB over the past few years - attract customers by making its database as easy to use as possible. This of course is just the announcement of the general availability of multi-cloud clusters, so we will reserve judgement as to its effectiveness until we can speak to some customers making use of the new features. However, what is noteworthy is that it moves the market in the right direction and will force others to avoid the lock-in mistakes of the past.