Five Unexpected Trends that Are Driving the Adoption of Serverless Computing
Serverless computing has quickly become one of the foundational pieces of modern distributed systems. An important component of platform as a service(PaaS) stacks, serverless computing is a very popular architecture style within the generation of cloud-first developers. Despite its popularity, here are questions about the paths that could help drive the adoption and ultimately monetize servile computing in the near future.
The subject of serverless computing scenarios has been long debated across technical communities. Most people associate serverless architectures with scenarios such as IOT or gaming but I believe those examples are too generic to infer specific paths to commercial adoption. Sometimes, identifying the monetization paths of low-level infrastructure technologies such as serverless computing is far from trivial. From that perspective, I would like to explore a few technology trends that I believe can help drive the commercial adoption of serverless technologies in the short term.
1-Blockchain as a Service
As strange as it might sound, blockchain as a service(BaaS) technologies can become one of the big drivers for the adoption of serverless technologies. Specifically, serverless stacks are a great solution to enable the access to off-chain data without violating the integrity of the blockchain. Ethereum’s ORacles and Oriject Bletchey’s Cryplets are good examples of the integration between serverless capabilities and BaaS stacks.
Bots technology is another trend that can drive the immediate commercial adoption of serverless stacks. In that scenario, serverless frameworks can be used to implement the business logic actions for a specific bot. Alexa Skills Kit and Azure Bot Service are example of that model.
Another example that might come across as surprising, deep learning can be another accelerator for the adoption of serverless computing technologies. Deep learning problems are typically modeled as async graphs on which nodes perform individual computation. For large scale deep learning solutions, serverless stacks can be used to implement the specific computation tasks connected asynchronously as part of a deep learning algorithm. Leveraging serverless computing as part of deep learning models mostly makes sense on large scale problems like the ones tackled by technologies such as Google’s DeepMind or IBM’s Watson.
Recently, I’ve written about how serverless technologies can become the new from of lightweight enterprise middleware. Trends such as IOT are driving new models of integration in the enterprise and serverless computing is becoming the default way to address many of the new generation of integration scenarios. Technologies such as Node-red or AWS Step Functions are showing the possibilities of leveraging serverless patterns in integration scenarios.
Hardware solutions are notoriously difficult to version an upgrade. Serverless stacks can drive new architecture models that can be used to improve the capabilities of IOT devices without the need to deliver new versions of the specific hardware. This approach is particularly relevant on invisible interfaces driven by sensors or voice commands. Amazon Echo family of devices is a great example of how IOT devices can leverage serverless computing to power invisible interfaces.