By Dr. Abhinanda Sarkar, Director, Academics at Great Learning
The US National Institute of Standards and Technology (NIST) defined cloud computing in 2011 as composed of four deployment models, three service models, and five essential characteristics. These serve as good templates to see where cloud computing came from, where it is today, and where it might be headed.
Historically, the earliest deployment models were private clouds, which were a natural extension of client-server infrastructures that companies used. Community clouds, such as for companies with many locations, came next. The true breakthrough came with public clouds (for example, AWS in 2006) where an external company provided cloud services. Today, hybrid clouds which have both private and public features are increasingly common as a deployment model.
Three standard service models of cloud computing are currently useful across industries.
- Software-as-a-Service (SaaS) is useful for companies that need specialized software that may not be their core competency. For example, PepsiCo uses Microsoft Azure Machine Learning to identify customer shopping trends and produce store-level actionable insights.
- Platform-as-a-Service (PaaS) is useful for companies that need multiple technologies to build applications. For example, companies like The New York Times and Accenture use Google App Engine in order to leverage a ready development environment that includes Java and Ruby.
- Infrastructure-as-a-Service (IaaS) is useful – indeed critical – for large companies that need to integrate cross-functional digitization efforts. For example, Etihad Airways recently used IBM Cloud to retool their customer service and check-in processes with a new DevOps toolchain.
Cloud computing is characterized by a few defining features. These characteristics are also enabling cloud providers and users to respond to upcoming trends in the technology and broader industry.
First, on-demand self-service. Cloud computing is arguably at its most impactful in enabling access to digital technology whenever needed. Companies and professionals use the cloud to do tasks that require specific software available at specific times. Consider an engineer on a factory floor using 3D CAD to quickly design or analyze a machine part. She can log on to, for example, Siemens Solid Edge which is a cloud-enabled design platform. The trend of managers, technologists, doctors, teachers, and other professionals accessing cloud services to support their competencies in real-time will only increase.
Second, broad network access. Access to clouds is through computer networks. As new kinds of devices get connected, cloud usage expands. But there are also devices that are at the edge of the network and computing at the same time. McKinsey reports speak of the device edge (e.g. smartphones), the remote edge (e.g. electric vehicles), the branch edge (e.g. retail outlets), and the enterprise edge (e.g. factories). The trend of more connected devices on an ever-expanding edge will continue and possibly accelerate as more devices are added to the ‘Internet of Things.’
Third, resource pooling. The economics of cloud computing is based on the provider being able to process requests using their own computing resources efficiently. An emerging challenge here is to manage computing resources sustainably as well. Like for many things today, the environmental impact of the high-performance facilities that cloud providers use is being closely looked at. Overall, it is argued that resource pooling is precisely the reason using clouds are more environment-friendly than, say, on-premise computers that may remain on and idle. The trend of society and industry steadily reducing carbon footprint actually might work in cloud computing’s favor. AWS, for example, has announced a long-term commitment to achieve 100% renewable energy usage.
Fourth, is rapid elasticity. The users of a cloud service typically have variable demand and the cloud adapts to both high and low usage – ideally quickly and seamlessly. Think of CoWIN, the Web and cloud-based platform that supported India’s Covid-19 response and vaccination drive. The trend of rapidly adaptive IT services will accelerate, if anything, in the complex times we seem to be in.
Fifth, measured service. What to pay for and how much to pay is naturally a central concern for cloud users. Related to this is the technology that measures and tracks usage. Think of ‘smart meters’ which are digital devices that record electricity use. Smart meters connected to the cloud are representative of the trend of thinking of computing as a consumed resource, much like energy.
Cloud computing is embedding itself into company operations and the way software is developed. For instance, the idea of a ‘microservice’ that breaks up a development task into independent micro-tasks, can be thought of as a contribution of cloud programming to software engineering. In fact, cloud computing models and technologies are becoming so ubiquitous that we may stop thinking of them as being innovative. Like many disruptive innovations of the past, we have made the cloud normal and a part of all our technological and social journeys.