Edge computing is a decentralized computing model that brings computational resources closer to the data source or endpoint. It involves processing data locally on devices or at edge servers, reducing latency and improving response times. Edge computing is often used for real-time data processing and applications that require low latency, such as IoT devices and autonomous vehicles.
Cloud computing, on the other hand, is a centralized model that relies on remote data centers to process and store data. It offers scalability, accessibility, and cost-efficiency but can introduce latency due to data transfer to and from the cloud. Cloud computing is widely used for web services, data storage, and enterprise applications.
Mark Swinson, enterprise IT automation specialist at Red Hat tells EdgeIR that edge computing is an ecosystem play, bringing different parts together to create solutions with flexibility.
“With the development of AI and machine learning, it’s increasingly likely that an application’s lifecycle will circulate through both ends of the spectrum, from the data center to the edge – therefore new solutions must equally support both,” he adds.
“Developments in cloud computing, such as the way Kubernetes implements a ‘desired state’ system where outcomes are specified, make managing complex topologies at scale more feasible. Applying these same approaches brings benefits to edge solutions too. As computing resources at the edge continue to grow in capability and capacity, we’re set to see more sophisticated workloads to be run close to where data is produced. This necessitates not only common standards and approaches but an intent to collaborate and, to some extent, a willingness to experiment.
“There’s no getting away from the dispersed nature of edge computing, or the need for remote installation. Once a device has been plugged in and powered on, automating the setup reduces the scope for errors and mitigates the need for highly skilled technicians. In summary, edge computing will become less distinct from data center-based computing and rather a continuum with consistent architectures, tools, processes and security. This will lead to greater flexibility, agility and confidence, opening up more edge computing opportunities.”
Location of processing and latency
With edge computing, processing occurs at or near the data source or endpoint, making it ideal for applications that demand immediate response times. Examples include industrial automation, autonomous vehicles, and augmented reality. In comparison, with cloud computing, processing occurs in remote data centers, which can introduce latency. Cloud computing is suitable for applications where latency is less critical, such as email services, document storage, and data analysis.
Speaking with Rajasekar Sukumar, SVP and head of Europe at Persistent Systems, says: “For some businesses, the remote infrastructure of the cloud simply can’t provide the ultra-low latency needed to transfer data swiftly from point A to point B. This is where edge computing comes into play – bringing data processing physically closer to the source.
“Though it sounds like another buzzword, ‘edge’ is very helpful in understanding this technology, as it refers literally to geographic proximity – it happens at the ‘edge’ or periphery of the network. A helpful analogy is comparing the cloud to a restaurant’s kitchen. Edge computing positions computation at the ‘chef’s table’ – not in the thick of the action but much closer than a typical dining table.
“This proximity enables real-time insights from vast datasets that would otherwise suffer lag moving to and from the cloud. Like having your meal prepared right in front of you rather than waiting for a waiter, edge computing eliminates unnecessary distance between data and processing. For businesses looking to become data-driven, edge computing unlocks key capabilities like instant analytics and rapid response times. It represents a paradigm shift in how computation can enhance data-intensive workflows without the cloud’s inherent latency. The edge’s speed and localization are leading to transformative new use cases across various industries. It’s a disruptive new computing model unlocking innovation through proximity.”
Edge computing offers ultra-low latency since data is processed locally or near the data source. This makes it crucial for applications that require real-time decision-making, like self-driving cars or telemedicine.
Cloud computing can introduce higher latency due to data transfer between the edge device and the remote data center. It may not be suitable for applications that demand near-instantaneous responses.
Scalability, security and cost
When looking at edge computing, scalability can be limited by the physical infrastructure of edge devices. Adding more edge servers may require additional hardware and resources.
Cloud computing, on the other hand, offers high scalability due to the vast resources available in data centers. Users can easily scale up or down according to their requirements.
In regards to security, with edge computing, data remains closer to the source, potentially reducing the risk of data breaches during transit. However, edge devices may be more vulnerable to physical tampering. With cloud computing, data is stored remotely, which can be advantageous for data security. Cloud providers often invest heavily in cybersecurity measures. However, data transmission to the cloud can pose security risks.
When comparing cost, the initial setup costs for edge infrastructure can be high. Maintenance and upgrades may also incur ongoing expenses.
Nevzat Ertan, chief architect and global manager for digital machining architecture at Sandvik Coromant, says: “Edge computing and edge analytics describe data capture, processing and analysis that take place on a device — on the edge of the process — in real-time. Unlike traditional methods, which typically collate data from several machines at a centralized store, edge computing is a distributed computing that brings a single, or a group of machines computation and data storage closer to the sources of data. This can improve response times and save bandwidth.
“Conducting analytics at an individual device can provide significant cost and resource savings compared to data processing using a purely cloud-based method. For clarity, this cloud-based, method refers to streaming data from multiple devices to one centralized store and conducting data analysis there.
“Using the centralized method, huge volumes of data must be collected and transferred to one place before they can be analyzed. This method can also create a massive glut of operational data — and weeding out insightful knowledge from the monotonous can be a painstaking task. With edge computing, operators can instead set parameters to decide which data is worth storing — either in the cloud or in an on-site server — and which isn’t.”
Ertan adds that edge computing is not an alternative cloud-based methods, and highlights that these technologies are not competing against each other.
“In fact, each is making the other’s job easier. The benefit of this combined model is that it allows enterprises to have the best of both worlds: reducing latency by making decisions based on edge analytics for some devices, while also collating the data in a centralized source,” he continues.
Ideal for real-time applications like autonomous vehicles, smart cities, and remote monitoring in industries, edge computing can also enhance privacy by processing data locally.
While cloud computing is typically suited for web hosting, big data analysis, content delivery, and enterprise applications that do not require real-time processing, both edge computing and cloud computing can work hand in hand.
“Edge computing and cloud computing are two distinct computing paradigms that serve different purposes but can also complement each other in certain scenarios. Both edge computing and cloud computing can offload data processing and storage from local devices, reducing the burden on end-user devices and enabling more efficient resource utilization,” says Sundaram Lakshmanan, CTO at Lookout.
“Edge computing and cloud computing can work together in a complementary manner. Edge computing can handle real-time processing and immediate decision-making, while cloud computing can handle more resource-intensive tasks, long-term storage, and complex analytics.
“Edge computing focuses on processing data locally and reducing latency, while cloud computing offers scalability, extensive storage, and centralized processing. Both paradigms have their unique strengths and can be used together to create a hybrid computing environment that optimizes performance and efficiency based on specific use cases and requirements.”
In practice, a combination of both edge and cloud computing, known as hybrid computing, is often used to leverage the advantages of each model.
Edge computing and cloud computing are distinct approaches to handling the growing demands of modern computing, and as technology continues to evolve, the boundary between edge and cloud computing could potentially blur, creating more opportunities for innovation and efficiency.
cloud computing | edge computing | latency | security