At the edge, through the fog, looking at the scattered clouds in the sky
Cloud, sky, and fog once symbolized aspects in the atmosphere; today these words resonate with companies that are planning their IT infrastructure. Though working on pooled compute resources is a dated concept, which was furthered by the advent of virtualization technologies, yet, it picked up steam in recent years. Today it touts to be a choice that cannot be ignored by any business that has innovation, responsiveness, expansion, transformation, and digitization at its core.
Cloud computing is the ability for users to get on-demand access to compute, storage, and network resources and to pay for only the resources they use, for the duration they use, thus accelerating innovation for businesses that employ technology and applications to reach their customers and potential buyers.
Gartner says: Global End-User spending on Public Cloud Services is expected to exceed $480 Billion in 2022.
Sky computing is the next level of cloud computing. This is a concept that was discussed by professors at the University of California Berkeley Ion Stoica and Scott Shenker. Sky computing attempts to erase the boundaries between multiple clouds and treat them as a one compute platform that users can employ to deploy their workloads. This is today gaining acceptability in the industry under the hood of multi-cloud architectures. For sky computing to become a reality there has to be a synergistic collaboration between the multiple key cloud computing providers of today to create the three tenets of sky computing as defined by Stoica and Shenker.
Sky computing is made up of three layers: compatibility, intercloud, and peering.
The compatibility layer attempts to abstract the low-level functions of the cloud computing providers. Intercloud capabilities require the cloud computing providers to establish protocols for seamless application and data exchange between their clouds, it also should enable the use of compute resources across the different clouds based on the capacity and the economics of use for compute-intensive workloads such as for AI training and hyperparameter tuning use cases. Peering is the layer establishing the agreements between the cloud providers on the terms and conditions and rules of application and data exchange.
Edge computing is the use of compute resources on devices that are at the edge of a network to store, analyze and work on the data collected, closest to the point of collection of the data. This primarily helps by preventing the moving of data to other compute facilities (datacenters or a commercial cloud), thereby reducing bandwidth costs and improving latency that is normally incurred when offloading data processing outside of the network. Edge computing today drives most of the businesses that use devices and sensors to capture, analyze and act on huge amounts of data.
Edge computing is a nascent, confusing trend that will complement cloud computing for nearly every enterprise by 2025. — Gartner
Fog computing is more like extending edge computing. The term Fog computing was first introduced by a product line manager at Cisco, Ginny Nichols. Fog computing is providing the compute, storage, and network resources for analytics workloads, closer to the place where data resides or is collected, which is just above the edge where the devices and sensors are situated. Fog extends the compute capabilities of edge for those workloads that require more than the resources available at the edge. Fog computing applications include smart grids, smart cities, smart buildings, vehicle networks, and software-defined networks.
Hyperscale computing, as the name suggests emphasizes on a computing architecture that scales out or scales in based on the workload or on demand. The compute, storage, and networks used to set up a hyperscale data center are different compared to data centers that are set up to perform compute-intensive or machine learning workloads. Hyperscale architectures are built for economies of computing with scalability at the heart. Hyperscalers are companies that offer hyperscale compute platforms, this requires hyperscalers to heavily invest in their infrastructure. Domination in the cloud computing space gives companies an edge in the race to be among the top hypescalers. Hyperscalers are also looking to expand into industry verticals and regulated businesses to establish themselves as a leader among hyperscalers. Today, most of the top hyperscalers are the same top players in the public cloud computing space. There could be more players entering the list of hyperscalers by bringing in services, partnerships and ecosystems. Hyperscaling of resources is one of the key aspects considered by Gartner when evaluating the top players among hyperscalers (or public cloud providers).
Gartner’s magic quadrant for 2021 depicts the market players in the space of CIPS (Cloud Infrastructure and Platform services), which follows close on the heels of the market for Hyperscalers of today (though there are other notable players like SAP, Accenture etc to watch out for).
“Change is the only constant in life” — Heraclitus.
What Heraclitus said holds true in the technology landscape too. Transformation, evolution, and change have been rapid in technology like never before. What better way to conclude than to hope for and welcome a future that brings together a convergence of technologies pivoting to solve some of the hardest problems that humanity and earth are facing or are yet to face.
