One of the big business benefits of cloud is to enable scale. With scale also comes economy. The growth of cloud platforms such as Amazon’s AWS, Microsoft Azure, and the Google Cloud Platform have driven down the cost of hosting computing and storage. The larger these platforms get, the more they can drive down cost.
Another economic benefit of the cloud is that it has created demand for the concept of disaggregation of the cloud architecture components, enabling software to be installed on commercial off-the-shelf hardware (COTS). In this model, cloud software such as CMPs, virtual machines, SDN, container-based applications, and even network functions virtualization (NFV) are installed on top of standardized, COTS hardware. The growth of cloud and disaggregation means that these standardized components can be produced in scale at lower cost than customized components. Much of the customization can be done in software, rather than hardware.
With scale comes management challenges, both for deploying and managing private clouds of medium to large scale and managing IT workloads that may also be running in public clouds. As cloud computing continues to evolve there are two fundamentally different classes of legacy and cloud native applications that need to be managed. How to approach that challenge is now a subject of fierce debate. Gartner has now long advocated a bi-modal approach to managing enterprise IT that draws a line in the sand between legacy applications and “cloud native” applications. Gartner essentially contends that legacy applications that largely run on premise don’t need to be updated as frequently as cloud native applications. As result, Gartner says it makes sense to manage these diverse application environments at two different speeds using tools and frameworks optimized for each.