We live in the age of “connected intelligence.” Computing technology has had a significant impact on global organisations and how people communicate and interact with information.

Edge computing has evolved as a computational model that uses the resources available at the edge of connected environments. Edge computing does not replace cloud computing; it augments it by distributing workloads in scenarios where traditional cloud architecture is ineffective. Multi-player gaming, augmented reality/virtual reality, autonomous vehicles, connected manufacturing floors, robots, and video processing are emerging use cases of edge computing. Central to the solutions chosen by organisations is the manner in which the data is stored, moved and accessed.

The proliferation of IoT has resulted in an explosion of data, forcing organisations to rely more heavily on computing and storage solutions. However, when migrating the entire IT infrastructure to the cloud, concerns about latency and cost feasibility arise. Businesses that use IoT sensors, actuators, and other IoT devices are increasingly seeking edge computing solutions like edge nodes, devices, and hyper-localised data centres. Edge computing augments the existing cloud archetype in IoT applications by bringing data processing closer to the data source, allowing enterprises to make better decisions faster.

The anatomy of IoT applications and challenges at the edge

Cloud-first IoT architecture

In a cloud-centered IoT design, devices connect to the cloud directly and exchange data with a remote data centre. This well-known design serves as the foundation for numerous web applications. The main advantages of cloud-based architecture include implementation robustness, easy management and control of IoT devices through centralised mechanisms, and device implementation simplicity. Data centralisation enables the processing of larger volumes of data and more accurate conclusions using machine learning technology.

Cloud-first applications do, however, come with limitations which should be considered. The cost of cloud design is considerable, and latency is unacceptable for many applications that require rapid answers from massive amounts of data.

From a security perspective, cloud app privacy concerns can be expected when all data is received and stored by third-party cloud operators. Centralised systems can be vulnerable to attack, potentially jeopardising the entire network of devices.

Peer-to-peer architecture

Direct, peer-to-peer communication is another method for Internet of Things (IoT) devices and apps to communicate with one another. Data is shared directly between peers in a peer-to-peer (P2P) network, with the server acting only as a connection broker. Each peer establishes a connection with a central server, mediating a direct end-to-end encrypted connection between them. Once the direct connection is established, the server is eliminated from the picture. This means that all data is stored on the devices and may be retrieved directly from them.

The benefit of this technique is privacy, as data is only transferred between devices with direct access. The dispersed design of the system strengthens security by forcing attackers to compromise individual devices to get access to more data. Because direct connections have a reduced latency, the flow of information can be substantially faster without the involvement of an intermediary. P2P systems are scalable, as an increase in the number of devices does not always indicate the need for additional processing capabilities.

Because P2P architecture is decentralised, it can prove an administrative challenge to conduct regular data backups. P2P raises the issue of trust and the provisioning of valuable services on top of IoT devices.

Hub-and-spoke

The hub-and-spoke design lies between cloud-first and peer-to-peer architectures. This strategy is used for IoT devices with limited resources and must connect to a hub before connecting to the Internet. The hub serves as a device gateway and implements network security and wide-area network (WAN) protocols. Hubs perform data pre-processing, use network controls and local data to configure and administer various devices in their local network.

The benefits of P2P networks are preserved in hub-and-spoke designs. Hub-and-spoke networks are easy to manage, and the method is well-known. However, hubs also pose potential risks. They may have a single point of failure. There is a high risk of vendor lock-in with equipment more difficult to replace than a cloud service and potential issues with serviceability.

Challenges at the edge

Network security has emerged as a critical factor in edge computing applications. As a result of decentralised apps, attacks have become more complicated and persistent. Traditional security measures focused on the network’s perimeter have proven ineffectual as standalone security strategies. Modern network security must deal with an ever-changing and diverse set of users and devices and significantly more pervasive attacks targeting formerly “trusted” portions of network infrastructure.

Edge apps generate massive volumes of unstructured data daily. This data provides real-time insights that can increase corporate efficiency, improve consumer experiences, and new revenue opportunities. Turning real-time insights into actionable insights requires analysing and processing data at the source – the edge – where people, devices, and things connect to the digital world. Using unstructured data at the edge necessitates a network that uses artificial intelligence to process data at a rate and volume well beyond what is humanly achievable. It requires infrastructure with an AI-powered “sixth sense” that identifies possible problems ahead of time, proposes a course of action, and uses automation to transform those suggestions into logical actions, all without manual intervention.

Aruba ESP

Aruba ESP (Edge Services Platform) assists organisations in managing their edge installations and accelerating transformation by continuously analysing network, user, and device information. Aruba ESP transforms data into knowledge to ensure business continuity with a single, cloud-native platform that can be deployed on-premise or in the cloud. This protects and integrates corporate infrastructure.

Aruba ESP is founded on the following fundamental principles:

AIOps is a critical component of Aruba ESP that leverages AI and analytics to precisely identify root causes of network issues and automatically remediate them. It then monitors user experience in real time, tunes the network to prevent problems before they occur, and continuously optimises and secures the network through peer benchmarking and prescriptive recommendations. AIOps is especially successful in actual customer deployments, as it dramatically increases throughput capacity, reduces issue resolution time, and improves end-user and IT experience.

With Unified Infrastructure, Aruba Central unifies all network operations for switching, Wi-Fi, and SD-WAN across campus, data centre, branch, and remote worker settings. Aruba’s unified infrastructure strategy gives clients the option of running controller services on-premise or in the cloud, allowing maximum flexibility at enterprise scale.

Zero Trust Network Security integrates role-based access technology, dynamic segmentation, and identity-based intrusion detection. This enables it to authenticate, authorise, and control every user and device connecting to the network and detect, prevent, isolate, and stop attacks before they disrupt business.

Aruba ESP enables the intelligent edge and empowers organisations seeking to accelerate transformation and ensure business continuity.

Conclusion

Given the importance of capitalising on opportunities at the edge, organisations should consider prioritising a solid network foundation. In the design of your IT infrastructure, it’s important to meet today’s expectations while preparing for the next significant technological advancement.

Aruba ESP is the industry’s first platform with AI-powered modules capable of meeting the needs of intelligent edges. Aruba ESP uses AIOps, Zero Trust security, and unified infrastructure principles to help IT and the network handle the velocity and volume of data generated and processed at the edge. It mitigates advanced threats from a vanishing security perimeter, and operational challenges posed by increasingly complex network architecture. As your IT partner, we can help you set up the best architecture with the relevant elements of Aruba ESP for your business needs.