Fog computing and edge computing have great potential to play major roles in the real-time processing of data. An astronomical volume of data is created by millions of networked devices around the world every year. This has prompted questions about the future of cloud computing: Is cloud computing on a path of decline? Can fog computing and edge computing be used to complement the cloud? To find answers, it is essential to understand the specifics of, and use cases for, each technology.
Edge Computing
Edge computing is a computing model that moves the data processing and storage capabilities to the edge of the network. Edge computing enables storage and processing of data on or closer to the machines that generate or consume the data instead of depending on cloud servers. Local data processing helps reduce the quantity of data that needs to be sent to the cloud for processing. In general, edge computing is useful in scenarios wherein instant data processing is critical, such as those involving smart cities, self-driving vehicles and robotics. Edge computing creates a distributed computing architecture that enables data processing and computing nearer to where data is generated and consumed, resulting in faster computing at the edge of the network.
For example, edge computing is critical for self-driving vehicle technology because it permits real-time decision making. The vehicles process the sensor data locally, allowing for instant responses and securing safety. Local data processing also allows enterprises working to create autonomous driving technology to easily comply with regulations that require local traffic and driving data to be contained within the jurisdictions in which it was collected.
Fog Computing
Fog computing is a decentralized computing model that stretches cloud computing power to the periphery of the computer network. Fog computing enables faster storage and processing of data by shifting the computing resources closer to the sources that generate the data. It serves as a convenient technology in scenarios such as those that involve the Internet of Things (IoT), wherein real-time decision making is needed. In general, fog computing creates a crossbreed architecture that relies on centralized and decentralized computing resources and enables data processing at intermediary nodes located near edge devices, resulting in low latency.
For example, fog computing supports edge computing in IoT environments. The hierarchical architecture enables edge devices to communicate with the nearest fog nodes for processing, storage and analysis of data. This facilitates smooth management of IoT deployment in cases where hundreds of devices are involved. The proximity also enables DevOps teams to manage asset inventory more easily and concentrate on the technicalities or issues involving the deployment of IoT devices. The cost of securing IoT devices is also greatly reduced due to the proximity of the fog nodes. From a compliance standpoint, auditors can easily observe the configurations and architecture of IoT devices or products and audit them at a quicker pace, reducing the walkthrough times spent with the DevOps teams.
Fog Computing vs Edge Computing
Fog computing and edge computing are closely linked in their objectives to move computation nearer to the source of data. However, there are certain significant distinctions between the two technologies.
Fog computing serves as an emissary between the cloud and the edge, offering data processing, storage and computational power locally. The scalability of fog computing is usually large because it has a hybrid architecture (centralized and decentralized) that extends across many regions. The intermediary nodes can handle many edge devices at the same time, offering increased storage and processing capabilities compared to individual edge devices. The ability of fog nodes to offer coordinated and optimized data processing, storage and computation makes them the ideal choice for big data management.
Edge computing focuses on the immediate data processing on the edge of the network. The scalability of edge computing is typically small and concentrates on individual machines and smaller regional networks. The availability of resources is usually constrained due to edge computing’s local nature, and, as a result, edge computing has lower latency, storing and processing capabilities. The pyramid model of edge computing enables devices to operate individually with minimal coordination, making this paradigm an ideal choice for the immediate processing of small amounts of data in a confined geography.
Limitations of Cloud Computing
Although cloud, fog and edge computing are interlinked computing models that form an extensive computing and processing ecosystem, when cloud computing is used on its own it has certain limitations: Storing critical and sensitive data in the cloud comes with the potential risk of data theft, data hacking or leakage. Although cloud service providers (CSPs) offer various security and privacy controls, cloud security breaches happen often, especially in shared environments and unfamiliar geographies.
- CSPs usually host applications and services across many data centers, especially for redundancy purposes, but outages and downtime can still occur. Outages may cause disruption to service availability and affect critical business operations.
- Uploading and downloading big data beyond the assigned bandwidth limit can be expensive, especially for small or emerging enterprises. Cloud services are sometimes hosted in data centers located thousands of miles from business operations.
- Data protection laws and residency requirements also pose challenges to the storage and processing of data in the cloud. The cost of limiting the storage and processing of data to certain geographies is high. Enterprises must also comply with regulatory requirements based on the type of data and their geographical location.
- Excessive reliance of cloud services on a stable Internet connection is seen as a limitation because the end user’s experience and accessibility to services hosted on the cloud can be impacted in cases of poor Internet connectivity.
Benefits of Fog Computing
The distributed computing paradigm of fog computing has many advantages such as:
- Fog computing can work effectively with low latency networks because it brings data nearer to the edge computing network and assists in improving the performance of applications and services on the network.
- Distributed architecture enables fog computing to reduce inefficiency by removing useless and irrelevant data to ensure that only efficient data is transferred to the cloud computing network.
- Fog computing also improves the security and privacy of data because any unnecessary and unhealthy data is eliminated from the system, thereby reducing potential security risk.
- Fog computing enables real-time processing of data because the data is supported by edge applications. Real-time processing is required for certain technologies (e.g., IoT).
- Fog computing’s reliance on Internet connectivity is limited. Devices can continue to compute and process data and services even if they are not connected to the central cloud. This makes fog computing an ideal choice in places with limited or unstable Internet connections.
Benefits of Edge Computing
The local data processing paradigm of edge computing has many advantages such as:
- The proximity of the data sources to the computational resources helps reduce the latency involved in sending and receiving data to and from a cloud server.
- Data processing and computation are performed on edge devices. So, edge computing optimizes bandwidth use by transmitting only required information to the central cloud and minimizes cloud expenses for small enterprises.
- Local processing of data helps avoid single points of failure common in the cloud environment, thereby improving fault tolerance. The load on cloud servers is significantly reduced by allocating the computational assignments to various edge systems.
- Local processing of data also aids in real-time decision making without the need to leverage the central cloud. Technologies that require immediate response and real-time decision making such as self-driving vehicles, IoT and event detection and response systems use edge computing technology to save time.
- Edge computing keeps data storage and processing near the source of the data, thereby minimizing the need to transmit to the cloud, and enhancing the security posture of the organization.
- Edge computing enables independent operations and computations to take place offline. So, edge devices can continue to compute and process data locally, ensuring uninterrupted functioning of critical services without Internet connectivity.
Conclusion
Cloud computing is an evolving technology that has played an important role in building modern computing systems and networks. Factors such as costs, availability, security and innovation will drive the future growth of cloud computing. Cloud computing, fog computing and edge computing can coexist. They are complementary technologies that can build a holistic cloud architecture so that the benefits of different computing services can be derived, leading to more efficiency in the system. Cloud computing is suitable for the permanent and detailed analysis of data and storage. Fog and edge computing are suitable for the immediate processing of data required for real-time response and instant decision making. The use of all three computing technologies can help develop an edge-to-cloud data channel that empowers enterprises to succeed in building sophisticated products and cutting-edge solutions. The volume of data generated by digital devices is only going to increase as more smart, autonomous and virtual reality devices hit the market. Edge and fog computing are at the nascent stage. Major cloud providers will see more requests from clients for edge computing and fog computing infrastructure support. Emerging technologies such as artificial intelligence and the metaverse must be regulated. When these technologies embrace edge computing and fog computing along with the cloud, regulating them will become the next necessary element for their success.
Author’s Note
The views expressed in this article are for information purposes only. The views are the author’s personal opinions and in no way a complete list or a one-stop checklist for edge computing, fog computing or cloud computing. Organizations should consult a licensed accounting firm for professional guidance.
KISHAN SATHYANARAYANAN | CISA, CCSFP
Is a Managing Director of third-party attestation practice at BDO USA. He has more than a decade of experience in the accounting industry, including with the Big Four accounting firms. He has worked with clients in the Asia-Pacific region, Europe and the United States. He specializes in information systems audit and has experience with clients in many areas including cloud, cybersecurity, finance, entertainment, healthcare, manufacturing and blockchain technology.
