At the close of 2022, companies everywhere are embracing novel cloud native development methodologies as the shift to microservices picks up steam. In 2023, new approaches such as WebAssembly (Wasm) will take their place to help developers tackle the new challenges inherent in building and running distributed applications in multiple clouds, edges and devices, writes Liam Randall, co-founder, CNCF wasmCloud and CEO, Cosmonic.
It feels like we’re transitioning from the age of containerization to the era of low-boilerplate, agile application development models. New techniques abstract non-functional requirements (NFRs), allowing developers to go from sketch to scale in a few steps. As standards reach finalization, cloud providers add their own WebAssembly flavors to the mix and new development languages – such as Python, Java and .NET – are added to Rust and TinyGo, I suspect 2023 will be a milestone year for cloud native application development.
Here are my cloud native development predictions for the next 12 months.
Shifting cloud economics, as cloud giants unveil WebAssembly lambdas
The economics of the cloud are fundamentally changed as the three main cloud providers roll out their Wasm lambdas. Webassembly lambdas are less expensive to operate and don’t suffer from the cold-start problem. This ushers in a new age where the first choice for fresh application development moves from containers to native WebAssembly. Wasm’s superior security and portability propositions, and the associated cost economics, make it a no-brainer for enterprises and developers everywhere, which is why we’re already seeing the major players jump aboard.
Data locality balkanizes the cloud: multi-cloud and edge adoption grows everywhere
In 2023, we will see the rise and expansion of regulations around data and information handling. The bar is being raised and we will start to see the nationalization of data – this is already happening in some regions. We’ll see others start to enact data locality laws so that that data cannot leave, be processed or stored outside of their borders. Increasingly, countries don’t want to lose the governance of their data, they don’t want it governed and subject to the laws of other regions, hackers, corporate entities and, above all, they don’t want people’s rights to be compromised. Organizations are already struggling to successfully deploy this pattern and we’ll see more turn to projects like CNCF wasmCloud to build distributed systems.
What this means is organizations will have no choice but to move to multi-cloud and multi-edge strategies. Even those that have long held out, choosing to stay in one cloud, will have no choice but to adopt additional clouds if they want to retain control of their data and continue to operate in certain localities. This will drive a move to agile new application development models which allow rapid, hassle-free creation and the capability to easily run applications in any cloud and on any edge or device.
Platform engineering via vendors delivers a ‘No Ops’ experience
CIOs everywhere say that 80% of developer time is spent on application operations and maintenance, according to a Deloitte study. Companies have responded to this crisis by constructing internal platform engineering teams that are responsible for hiding the complexities inherent in modern cloud native development. The goal of platform engineering is to give modern developers a simpler way to increase the velocity of feature delivery while, at the same time, ensuring that every application meets the incredible array of complex legal, regulatory, and best practice requirements.
Many cloud native organizations, such as Adobe and BMW, realize internal bespoke platforms may not be the most economical way to increase developer velocity. Major players like Cloudflare and Fastly have been quick to deliver on this vision. Now, a cohort of new startups, such as Cosmonic and Fermyon, are building the new compute standards and platforms that will dramatically simplify the application development lifecycle. While approaches and value propositions may differ, the community shares the belief that developers should be able to easily submit business logic and have it scale automatically to meet demand whilst the surrounding platform provides common and composable components like key-value stores and web servers.
New revenue streams emerge at the edge
Ushered in with the rapid adoption of WebAssembly comes the emergence of new business models. Wasm changes the way developers can build, run and deploy software and allows them to push more advanced logic out towards the edges. Fundamentally, this will lead to the evolution of familiar industries via completely novel revenue streams. This will be driven by peer-to-peer services whose unicost economics are less dominated by the cost to deliver the service in the cloud and are, instead, driven by the fact that more and more logic can run at the edge – even on users’ own devices.
Finalization of WebAssembly standards will quicken adoption
In 2023, an important milestone will be the finalization of the Wasm component model. This will lead to greater interoperability between languages and vendors, enabling developers to pick and choose pieces of their application that are implemented in different languages as different value propositions.
More specifically, we will see github distribute libraries as Wasm components. Developers are then free to treat github as the world’s largest crate of ‘Lego’. They will start to realize that, instead of pulling large amounts of code into their applications that they then have to maintain, they can instead choose contracts – lining up with the industry’s overall trend towards Saasification and Paasification.
The industry will start to agree on common standards for key-value, blobstore, message queues and as WASI-Cloud. Wasm will have a similar effect to what the electrical outlet had on power generation. You don’t know what’s behind the power outlet – be it coal, solar, wind etc. You simply plug in your appliance and get what the provider says you’ll get — power on demand. Applying this logic to application development means greater portability and more rapid innovation.
Making software supply chain vulnerabilities a thing of the past
In 2023, Wasm will finally start to make a dent in the SBOM model (vulnerabilities present in the contents of applications), where inherited vulnerabilities can bring development to a standstill. As most developers will recognize, the ingredients in the can can go bad after you’ve built the application – when you made the app it was a good library, now it’s bad and the application is poisoned. People are starting to realize the only way to avoid SBOMs, like the Log4Shell vulnerability in Log4j for instance, is not to play at all. The adoption of the Component Model will remove tightly coupled non-functional requirements where vulnerabilities often reside.
In the same way containers and Kubernetes hastened the shift to the cloud, WebAssembly will help overcome the challenges that have resulted from that move. As AWS, Google and Microsoft introduce their Wasm lambdas, it will be exciting to see the use cases for WebAssembly proliferate, not only across multiple clouds but also at the outer reaches of the network edge.