Event-driven applications drive next wave of IaaS evolution
There’s a new paradigm for the public cloud that combines the best of infrastructure as a service, platform as a service and software as a service to improve the cloud business case for both vendors and users. This new paradigm — which revolves around cloud web services and, especially, event-driven applications — will shape the future of the public cloud market.
Infrastructure as a service (IaaS) offers virtual machines on demand, meaning vendors can sell IaaS to pretty much anyone, no matter which applications they run. But, for the user, IaaS impacts only server costs — organizations still have to sustain their software platforms and applications. This means that public IaaS clouds can’t significantly reduce the operations burden for IT. It also meant, until recently, there wasn’t much IaaS providers could compete on other than price.
Amazon Web Services (AWS), Microsoft Azure and Google Cloud Platform have all recognized that they can add features to IaaS for differentiation, and to help users build applications that exploit the unique features of public cloud. These features — sometimes called platform services or cloud web services — are a set of application program interfaces (APIs) to cloud-hosted functionality that simplifies the development and deployment of complex applications built specifically for cloud. Like platform as a service (PaaS), these cloud web services are written into the software itself but, unlike most PaaS offerings, they don’t impose a specific OS or middleware. Some of them are nearly complete application middleware platforms, approaching SaaS functionality.
The evolution of cloud web services in IaaS
Amazon was the first cloud provider to see the potential in supporting new application development in the cloud.
Cloud web services aren’t exactly new. All public cloud providers offer database services, image management and basic deployment tools via APIs. What is new, however, is the range of features cloud providers now offer as APIs — from expanded relational database management services to the internet of things (IoT) and mobile device support, event handling and even virtual middleware platforms for hybrid cloud. These new features share a common goal of bootstrapping development teams to build cloud-specific versions of new applications.
Microsoft really started this off with its Azure PaaS model. Azure didn’t propose to shift everything to the cloud, but rather to build a platform — OS and middleware — that extends from the cloud to the data center and allows applications and components to migrate between the two. This made Azure the logical choice for hybrid cloud, which initially threatened Amazon’s IaaS dominance. AWS was Amazon’s counterpoint to the Azure model, and Amazon was the first cloud provider to see the potential in supporting new application development in the cloud.
Event-driven applications rise in the cloud
Amazon’s special insight into the cloud was that many new, cloud-specific applications were probably related to event processing — which is the next generation of computing. It’s essential for things like IoT, but valuable even for the mobile empowerment of workers. AWS Greengrass, a platform to host event-driven applications outside Amazon Elastic Compute Cloud, is Amazon’s counter to Azure’s on-premises Windows Server capability that allows users to place some application features close to the point of event generation to reduce latency.
Amazon and Microsoft built their event processing on something called functional programming or lambda functions. While Lambda is the name of an AWS service, it’s also a general programming term for pieces of logic that perform a complete, atomic functional step in processing, storing nothing between uses. That means you can spin up one piece of functionality as needed, scale the processing capability by using as many copies as you need and pay for things only when you use them. It introduces not only a new cloud pricing paradigm of true pay as you go, but also a new development paradigm.
Google watched this play out, and its response is Google Cloud Functions. Google Cloud Functions’ focus is on microservices, which were already popular with enterprises, rather than on lambda functions that are only emerging. Microservices are small, reusable, components designed to modernize applications’ use of the service-based features that service-oriented architecture first defined. Because they’re small and lightweight, microservices fit well into containers and allow enterprises to more easily deploy containers in private clouds.
This could give Google an advantage, but the vendor still has to catch up on private and hybrid cloud integration to counter AWS Greengrass and Microsoft’s common platforms across the cloud and data center.
These event-driven computing models will drive the future of the cloud. It can be hard to justify moving something from a data center to the cloud; it’s already running and you have to be able to prove cost savings. But it’s another matter to enable the cloud to drive the next wave of productivity enhancement — a wave that could add significantly to IT spending.
If new productivity gains are the engine of new cloud applications and growth, then providers have to offer functional and microservice features, support contextual and event-driven applications and facilitate hybrid IT. The major cloud providers have all these pieces and, in the end, the one who integrates them best may be IaaS market winner.