R.I.P. MBaaS — Long Live BaaS
MBaaS technologies grew out of a very specific need, but one by one they've been shutting down. Where does the industry go from here?
Join the DZone community and get the full member experience.Join For Free
The Mobile Backend-as-a-Service field grew out of a very specific need: developers needed to build mobile applications fast. It was as simple as that. Desktop and web applications had been around for generations (in Internet time) and in the mid-2000s developers slowly started waking up to the fact that those applications needed companion mobile apps. However, while developing applications for the desktop and web environments was already a well-refined process, the mobile development that was about to take place was completely new and different. Gone were the days of the fractured device-oriented J2ME and BREW ecosystems. Now, companies like Google (Android), Apple (iOS), Blackberry (BlackberryOS) and Microsoft (PocketPC) were getting serious about releasing desktop-grade operating systems built specifically for mobile devices. Developers needed to quickly move from having either no mobile presence or (worse?) a WAP-only mobile site, to having a full-fledged application running on a mobile device.
And thus, MBaaS was born. At the time, we often considered the mobile, desktop and web environments three completely disparate systems. While they always talked to each other, REST APIs hadn’t completely taken over yet to break down the existing barriers between the three different platform environments. So MBaaS came about and solved a very particular problem; it allowed developers to focus on building applications without having to worry about dealing with push notification certificates or spinning up new databases (among many other things).
That time has now passed; the MBaaS of yesterday is dead and it is being replaced by its evolutionary successor: Backend-as-a-Service. Nothing puts that concept into more stark relief than the shutdown of Parse (one of the original heavyweights in the MBaaS space). There are a number of things about the MBaaS of yesterday that needs to disappear into technology extinction; the primary among them being the “M” in MBaaS. In today’s technology world, consumers increasingly move between desktop app, web app, tablet app, mobile app, smartwatch app etc. in the blink of an eye. Just ten minutes ago, I switched from using the Slack desktop app to their iPhone application and quickly followed that up by moving to their iPad application; and I didn’t even realize that what I was doing would have not long ago been difficult for a developer to code. With the “always available” use cases consumers demand of their applications now, there’s no longer a need for the mobile-focused backend infrastructure; developers need a more versatile backend infrastructure that can address all of their needs.
That brings us to the next requirement for the BaaS of today: flexibility. Developers have a choice in the tools they use. Do they use Cisco’s Tropo for SIP and SMS integration or do they choose Twilio instead? Do you use Dweet.io for IoT messaging or do you use PubNub? Any BaaS that doesn’t allow for a plug-in architecture in the near future is dead on arrival. As a developer, my choices are dictated by a multitude of factors; chief among them being business needs. If a BaaS provider can’t give me the flexibility to be nimble in my decision-making process, then it’s not worth my time.
Unfortunately, with the incredible amount of choices I have in platforms to integrate with, this translates to more work. A developer today is spending more time integrating new technology into their development process than they ever have. No one wants to reinvent the wheel; that’s why REST APIs are so popular. However, as the amount of REST APIs explodes, so does the amount of work I have to do familiarizing myself with those same platforms. That’s why we’re going to see a slow merger of BaaS platforms with iPaaS (Integration Platform-as-a-Service). For BaaS platforms to be truly useful, they need to start providing easier ways for developers to integrate the platforms and toolsets they use with the systems and a myriad of data sources that matter. While BaaS isn’t inherently set up to do that, iPaaS has been doing it for years. Again: why reinvent the wheel?
Finally, all of these changes are going to necessitate the inevitable divergence between enterprise BaaS solutions and consumer BaaS solutions. The two markets are driven by very different business needs, despite a similarity in the technology stack. How do we prevent another massive ecosystem upset as we saw with the death of Parse? We focus on serving specific market needs. That means that in the enterprise market, we focus on providing solutions that can survive the demise of any given BaaS provider; solutions that allow for dedicated and on-premise deployments, which significantly extends an organization’s runway, should a technology provider go out of business; solutions that provide a plug-and-play architecture where you can assemble best-of-breed microservices – if one vendor fails, simply pick another. Add in managed DevOps services and, all of a sudden, BaaS becomes a pretty powerful and future-proof tool, without introducing the dangerous dependencies of yesterday’s MBaaS.
MBaaS is dead. Out of its ashes will rise a Phoenix that allows for the growth of a more stable ecosystem in general. The needs of the developer are still the same: to build and deploy stable applications incredibly fast. For this familiar problem we now have a new solution: Backend-as-a-Service.
Opinions expressed by DZone contributors are their own.