The Future of the Serverless Market
According to a recent report, the serverless market is projected to grow at 32.7% up to 2021. Serverless computing, which is sometimes also referred to as “Function-as-a-Service” (FaaS), is the concept of executing your application code as a series of tasks in an environment completely managed for you. Serverless has gone from being a futuristic technology trend to a widely used option for running applications with ease.
Even if your organization is not yet taking a full plunge into serverless computing, many organizations are dipping their toes with serverless and switching some of their workloads from traditional infrastructure to serverless platforms. Serverless computing is becoming a mainstay of DevOps teams.
And that begs the question: What will the future of serverless look like? How will serverless computing evolve going forward?
Here are some trends that will influence the future of serverless computing.
1. Less vendor lock-in, more integrations
Serverless is about offloading the underlying infrastructure to a vendor. With the number of moving parts and complexity of application infrastructure, vendors have to build very opinionated systems that proactively make decision on load balancing, storage, failover and more. This leads to very closed platforms, and the dreaded spectre of vendor lock-in.
This was clear in the serverless computing market early on, when AWS Lambda was the only major serverless platform available. A number of competing options have now appeared on other public clouds, including Azure, Google Cloud and IBM Bluemix.
But there is currently little interoperability between the serverless services on each of these clouds. You can’t easily migrate functions from one cloud to another.
This is a major concern for many prospective users of serverless, especially those who build systems with portability in mind from the start. In future, serverless computing will mature to give developers more control over any part of the stack they really want control over.
One of the ways this will be made possible is by integrating APIs. Serverless platforms will expose APIs to integrate them with other platforms and cloud vendors. Google recently acquired Apigee, an API management service. Apigee is a great option for managing microservices. Autodesk uses a combination of AWS Lambda and Apigee to manage its microservice apps.
2. Serverless will be used for MVPs
Because of how quickly you can go from code to app, serverless is ideally suited to build minimum viable products (MVPs). In fact, for startups, serverless is a great accelerator. You don’t need any IT resources until your app crosses the initial validation phase, and you have enough funding to hire IT professionals. Because of how easy it is to change functions with serverless, you can make numerous changes to the MVP easily and bring it closer to production.
3. Open source frameworks will emerge
Another antidote to vendor lock-in is open source serverless platforms. This may seem like an oxymoron considering how closed typical serverless platforms are today. However, this trend is on the rise. There are already existing open source options for serverless computing, like Apache OpenWhisk, Fission, and Iron.io. Many of these solutions use containers as the basic unit for computing.
4. Builds on Kubernetes to manage microservices
One open source serverless vendor, Fission, integrates with Kubernetes to manage microservices applications the serverless way. This means you don’t need to create and manage Docker containers, as it’s automatically done for you by Fission. This gives you a simpler backed, and importantly, the agility of using containers to package and execute your functions.
5. Enable distributed IoT apps
No post about future predictions would be complete today without a mention of IoT. In this case, serverless computing is ideally suited for IoT jobs which are numerous, on task-specific, and short lived. With the scale of IoT computing, you need a distributed system. Maintaining this system can be resource- and time-consuming. This is where a serverless platform comes in to take the heavy lifting off you, and let you focus on how you want to architect your IoT system.
6. Bring high-performance computing to everyone
Enterprises are starting to look to serverless computing to provide a refreshing new way to manage high-performance computing workloads that require intense IT operational skill. For example, Nordstrom, the fashion store, uses Lambda to power a recommendation engine on their website by using live user requests as input. PhotoVogue, an app from Vogue, leverages Lambda to do real-time image conversion in the cloud. When a user uploads an image to PhotoVogue, a trigger initiates the image conversion into many formats like GIF, JPEG, PNG, and TIFF.
These tasks would typically require a highly skilled IT team and some heavy-duty infrastructure to pull off smoothly. But with serverless environments, these organisations can bring advanced functionality into their apps with little effort.
The market for serverless computing is in hyper-growth right now. AWS Lambda is one of the fastest-growing AWS products as of 2016, and AWS has plans to invest in it heavily in the coming years. Azure Functions, Microsoft’s take on the serverless revolution, has some unique features like support for additional programming languages, and an open-sourced runtime. Serverless is the future of cloud computing, and its projected CAGR of 32.7% is no myth.
Want to dive deeper? Check out our introduction to serverless with AWS Lambda.