What is edge computing?
In IT, edge computing is the processing of data at the network edge. The technology is a distributed computing paradigm that takes a decentralized approach. Instead of sending data from devices to central systems or the cloud for processing, edge computing processes the information where it is needed – at the edge of the network. The benefits of such a distributed topology are low latency and bandwidth requirements.
Reading Time: .
A definition of edge computing
Unlike the ongoing cloud boom, edge computing involves aggregating and processing as much data as possible locally on devices at the edge of the network. Particularly in complex IoT networks and in Industry 4.0, this has the advantage that information can get quickly and reliably to where it is needed. In practice, networked production robots, wind turbines, and elevators can use predictive maintenance to respond directly to signs of wear and tear and request any necessary maintenance work. Compared to traditional maintenance processes based on operating hours, with a system like this, maintenance is only performed when it is actually required. This saves costs and extends maintenance intervals – and thanks to edge computing, a permanent connection to the internet is not needed. This is made possible by increasingly powerful microchips that can even handle intelligent workloads locally thanks to neural processor engines and lean AI (artificial intelligence) algorithms. This is the prerequisite for many modern solutions for augmented or virtual reality (AR & VR), smart cities, or autonomous driving.
What are the advantages of edge computing?
Compared to a centrally configured server and cloud topology, edge computing provides lower latency because the data does not have to travel long distances to the nearest data center. At the same time, edge computing reduces the required data bandwidth since data processing takes place locally on each device or in nearby edge gateways wherever possible. In practice, the reduced demand for bandwidth also results in lower operating costs. With these properties, edge computing lends itself to all applications that require short latency times and where the workloads can also be processed on each endpoint. AR and VR headsets are popular examples of edge computing. Fast response times are required here for an ideal user experience. For this reason, the image and position data determined by the sensor system is processed directly on the device.
How well does edge computing work in conjunction with the cloud, IoT, and 5G?
Edge computing is often misunderstood as a competitor or alternative to “traditional” cloud computing. In fact, both approaches work together in perfect harmony. The same can be seen with other popular technologies such as IoT (Internet of Things) or 5G mobile communications. In all of these areas, edge computing acts as a catalyst and enabler that enhances existing solutions or makes innovative approaches to solutions feasible in the first place.
Edge computing & cloud computing
Edge computing & IoT
Edge computing & 5G
What are the vulnerabilities of edge computing?
Compared to centralized topologies, where only a primary server system or a cloud data center needs to be secured, all endpoints involved in edge computing must be given equal attention in IT security. Depending on the type and scope of the edge network, this quickly results in a great deal of time and effort.
Secure hardware for the edge of the network
When selecting edge hardware, companies should limit themselves to the previously defined minimum requirements so as not to unnecessarily increase the network’s attack surface. For instance, is a USB slot on the device needed for operation, or would such an interface unnecessarily provide attackers access to the systems?
Using encryption to secure data
Simply due to local conditions, securing data at the network edge cannot hold a candle to the security standards of a data center. Data transmission between edge endpoints, gateways, and central control and cloud systems should be encrypted end-to-end. Even if attackers manage to infiltrate the system with a man-in-the-middle attack, data traffic is secure against eavesdropping and sabotage. Anyone who fails to use encryption protection risks expensive failures. Cybercriminals could, for example, tamper with the data sent by sensors in industrial plants to inflict serious damage on production equipment.
Protection from tampering is required
Edge systems must have a minimum level of protection against tampering to make it difficult for attackers to gain access to the network. This includes securely mounted housing covers and sensors that immediately report physical attempts of tampering. The issue of tamper-proofing is particularly important for devices that are installed in public spaces, where protection against access provided inside offices, production facilities, or factory halls is not available.
One-time setup and configuration of edge systems is not enough. With increasing requirements and new business processes, companies often have to adapt or expand their edge networks. In addition, the endpoints are not made to last forever and require maintenance and replacement when necessary. In order to avoid losing track of the network, detailed lifecycle management is required for deployment, decommissioning, onboarding to gateways and the cloud, and maintenance (software and hardware).
What are the threats posed by edge computing?
The use of a reliable security solution to defend against malware is also mandatory for edge projects. The challenges here are very similar to securing IoT networks. Attackers use specially designed Trojans that spread automatically in networks. One popular example of this kind of malicious software is the malware Mirai, used by cybercriminals to set up botnets.
In the fall of 2016, attackers succeeded in using a DDoS attack to disrupt the servers of DNS service provider Dyn. Consequently, the US provider’s customers were unavailable for hours, including major websites such as Twitter, CNN, the Guardian, and Netflix. The attack was carried out via a powerful botnet that the hackers had built out of vulnerable IoT devices such as IP cameras, printers, smart TVs, and the like using Mirai malware. This combination of networked devices was able to bring Dyn’s server systems to their knees by bombarding them with a huge number of requests.
Similar scenarios are also conceivable with endpoints for edge computing. Due to the generally higher computing power of edge systems, the risk of more severe attacks is even greater.
What you need to know about edge computing
With edge computing, workloads are moved to the edge of the network and, if possible, executed on the endpoint or nearby gateways where the data they need is also generated. This enables real-time applications with the lowest level of latency, which would not be possible with traditional IoT cloud architectures. In addition, edge computing gets by with less bandwidth because data is processed locally. This saves costs and frees up resources in the data center. In practice, edge computing is used primarily in the areas of Industry 4.0, AR, VR, smart cities, and autonomous driving. As with the IoT, securing networks for edge computing is difficult because all of the endpoints involved must be factored into IT security. If this is not done, companies run the risk of cybercriminals infecting the endpoints with malware and misappropriating their computing power to launch attacks, or stealing or tampering with the data generated there.
If you are interested in futher informations, we are willing to send you our product sheet for free
How to divide, classify, and analyze incoming website traffic:
- How is the web traffic structured and what bot types are there?
- What are the advantages of Myra Deep Bot Management?
- Which product features does the solution have?