Edge Computing Vs Cloud Computing
Published on : Wednesday 03-11-2021
Edge computing is playing an important role but Cloud computing will remain as an essential part of the IoT ecosystem.
Cloud computing facilitated data storage over the internet, easy to secure and control while still allowing for reliable remote access from any device. Cloud computing also enabled edge computing, which while using the cloud concept, brings the data closer to the end user, decreasing latency, improving speed, and creating wholly new opportunities for digital transformation across industries. So both these methods of data storage, analytics and usage have distinct advantages as well as some drawbacks. So what explains the rising popularity of edge computing?
“There are multiple reasons behind the growth in popularity of edge computing,” says Sudhanshu Mittal, Head – CoE Gurugram & Director – Technical Solutions, NASSCOM Centre of Excellence – IoT & AI. According to him, the three most important reasons are:
1. For automation where low latency response is needed, it is not feasible to rely on the cloud, it is imperative to use Edge computing where latency and responsiveness can be controlled. In similar manner, when response can be critical to operation, it becomes difficult to rely on internet services where you can face the packet loss.
2. For cost controls – cloud service can become expensive very quickly with the growth in storage requirement. Corporates may want to move large data management services to the edge.
3. Security – sometimes corporates have security requirements whereby their data should not leave their premises. In those cases they will go for edge processing.
“Communicating the data to the cloud and bringing it to and fro to the edge device, makes the system ‘near’ real time,” says Pradyumna Narayan Tiwari, Senior Consultant – Edge Consultant. “There is a cost attached to cloud based storage, analytics and device onboarding solutions, wherein edge computing only connects back to cloud for billing requirements, so it can be said Edge is a one-time capital investment, for running various cost effective edge solutions,” he adds.
Rob van den Dam, Thought Leader, Lateral Thinker and Strategic Visionary, maintains that the rising popularity of edge computing has all to do with its ability to analyse data at its source. This is important as the increasing number of connected devices and sensors, the growth in industrial automation, and the demand for immersive experiences will require more decision-making at the edge of networks. “Artificial Intelligence plays a central role in the success of edge computing. In fact, AI has become the key driver for the adoption of this technology. AI can process large amounts of data in short periods and provide the insights to drive quick, local, data-informed decision making,” says Rob.
“Edge computing is the computer processing and storage power available at locations where data generation and usage happen; fog computing provides data, compute, storage, and application services to end-users by extending cloud computing and services to the edge of the network,” says Rajabahadur V Arcot, Independent Industry Analyst/Columnist and Business Consultant. “Cloud computing refers to internet-accessed computing resources such as data storage, processing, applications, development tools, software, servers – both physical and virtual servers – and others. These three computing resources represent different layers with each layer leveraging the capabilities of the other,” he adds, making the distinctions clearer.
So what are the unique advantages of Edge Computing vis-à-vis cloud models? Is cloud computing more of a status symbol? “Cloud computing cannot be called a status symbol by any stretch of imagination. For small businesses, it provides an invaluable service where they can start the digital journey and scale as they grow, instead of putting upfront CAPEX investment and also investing in in-house resources to manage the deployments,” asserts Sudhanshu Mittal. “The key advantages of edge computing are predictability of response, low latency and data security. However it is unlikely that anybody would go for purely Edge computing – mostly corporates would go for hybrid models where sensitive services would be within their own network and others would go to the cloud,” he elaborates.
Continuing in a similar vein, Pradyumna Narayan Tiwari, says Edge computing is all about running models and algorithms at the ‘edge’ i.e., near the machines, which increases the decision taking capability of the system. Many industries still rely more on ‘in-house’ or ‘on-prim’ solutions, in such cases edge computing works as a gateway to Industry 4.0 with taking care of privacy issues or customer’s priorities. “Cloud rather than just being a status symbol becomes a must in the present day scenario, it helps organisations to abide the industry 4.0 solutions with other corporate networks,” he emphasises.
“Improving operational responsiveness is the key reason why organisations are increasingly counting on edge computing. They believe edge-induced responsiveness will lead to significant business benefits, including reduced operational costs, more automated workflows and increased productivity. But it is not only about improving responsiveness. Edge computing will also ease the growing pressure on the core network. Instead of overwhelming the network with a flood of relatively insignificant raw data to centralised cloud resources, edge applications can analyse, filter, and compress data locally. This will also improve energy efficiency as more data is processed at the edge and less moves to and from the cloud, thus reducing power consumption,” explains,” Rob van den Dam.
Does using cloud models of business eventually play into the hands of the giant corporations? How can companies protect against this monopoly? “A large corporation, which has invested in building service capabilities, will always have an advantage and will try to extract maximum value from that. However once other corporates see the benefit, they will also jump into that segment. Same is true with cloud services – while Amazon, Google have first mover's advantage, others like IBM, Microsoft and Indian players like Zoho and others are entering the business and provide extremely good service at attractive price points,” says Sudhanshu Mittal. He also opines that ultimately it will be upon the users to identify the right service, which meets their requirements instead of blindly running after one service. “One of the challenges is that since users are comfortable with one major player, they are not willing to move to another one even if the other player is providing services at substantial discount. It is this reluctance which builds and sustains the monopolies.
Nowadays many other players are coming in the market with cloud infrastructures but you will find that most of the Fortune 500 companies are kind of ‘shop’ for few cloud providers. Other reasons for monopoly can be seasoned, secured and compliant solutions provided by these cloud infra providers so that corporations find them easy integrations to the present IT infra,” reasons Pradyumna Narayan Tiwari.
Rob van den Dam concurs with this. “No doubt, the giant tech companies have dominated the cloud space since the beginning. In Q1 2021, the top 5 providers – Amazon AWS, Microsoft, Google, Alibaba and IBM comprised 72% of the $129 billion market for infrastructure as a service (IaaS), platform as a service (PaaS), and hosted private cloud service, with Amazon taking nearly a third. Although the market for cloud infrastructure comprises more than just these 5 companies, these players stand alone as hyperscalers with the technology, resources, capital spending budget and customer momentum that make them unique,” he states matter of factly.
There are many other questions in this Edge Vs Cloud debate. At what size or maturity level of an organisation do the disadvantages of cloud computing out-weigh advantages? What factors play a role when a startup or SME does an evaluation of going in for cloud computing? How do the issues of data localisation get resolved by providers of cloud services?
According to Sudhanshu Mittal, there is no hard-and-fast rule here. “Even large organisations use cloud services and move to hybrid models where more sensitive data is managed in-house while other services are deployed on cloud. For any startup or SME, the key criteria is what kind of services they can deploy and how much they can just outsource to the cloud provider as they are much more focused on their core business,” he says. “As for data localisation, the Government of India has come up with rules requiring cloud service providers to store users' data in India. Users normally won't have the required clout to get a service provider to change its policy,” he adds.
There were other terms in play like fog, which should be between the edge and cloud. Are these technologies still relevant? “Fog computer gathers data from edge computers or IIoT devices, processes them, and generates output. It is typically located close to a cluster of edge computers/IIoT devices. With the use of fog computing, the efficiency, security, and privacy of the overall system are improved,” explains Rajabahadur V Arcot. Working in conjunction with edge computers, fog computers improve the response time and reduce bandwidth usage as only the need-based minimal amount of data is exchanged both with the edge computer and cloud. “Edge computer force multiplies the power of edge computing and is often used in cases where it is technically and economically more efficient to gather, process and act on chunks of data that are interrelated and emanate from a cluster of edge computing devices or sensors and actuators embedded to plant equipment,” he elaborates.
“Fog computing is a hybrid model, where you use edge devices to handle major parts of your computation and cloud for storage purposes. It is still relevant but ultimately it depends upon local requirements. As stated above, for large enterprises it is a hybrid model which is likely to work, not pure in-house or cloud,” states Sudhanshu Mittal.
“Fog computing is all about running small analytics at the edge. People in industry use Fog computing and Edge computing interchangeably as both means bringing analytics closer to edge. Cloud is meant to run bigger analytics models for a longer term, fog in turn means running small models. Many edges can use a single fog computing model,” says Pradyumna Narayan Tiwari.
How would rollout of 5G change the equations for Cloud? “5G is designed to support applications that depend on increased connection speeds, improved traffic capacity, very low latency, high reliability/security and support for high density of devices. But it is especially the blending of 5G, edge computing and AI that represents a unique foundation for enabling new ground-breaking use cases simply not possible today. Without AI-based edge computing, 5G applications and services will rely upon connecting through the core network to centralised cloud resources for storage and computing, losing much of the positive impact of the latency reduction enabled by 5G,” says Rob van den Dam.
“5G would enable many more services than it is currently feasible to do with 4G/LTE. It is also expected to provide low latency and guaranteed data communication, which may cause users to move from in-house deployments to cloud. Both of these – new services and low latency capabilities will increase the demand for cloud services,” opines Sudhanshu Mittal.
“5G interestingly is not only meant for the internet of humans, it is a great boost to ‘massive internet of things’ as well. The telemetry from small devices would travel in guard bands to cloud or other servers. Any small device such as car keys, spoons, mouse and aquarium water temperature would be easily accessible that too in an isolated manner. Massive IoT does not target on running machine learning, AI models on edge, rather it targets at setting up massive amounts of small telemetry devices that can pump the data to cloud, which is going to increase the need of cloud computing, cloud engineers several times. Cloud has a brighter future ahead,” concludes Pradyumna Narayan Tiwari.
(Note: The responses of various experts featured in this story are their personal views and not necessarily of the companies or organisations they represent. The full interviews are hosted online at https://www.iedcommunications.com/interviews)