The combination of cloud services in IaaS environments integrated with edge computing facilitates the development of applications with very low latency that currently cannot be deployed in today’s centralized architectures of public cloud services.
What is edge computing?
This concept basically refers to an IT architecture model in which data storage and processing takes place in a location as close as possible to the user, device or service that will consume the data.
This architecture has three key benefits:
- It dramatically reduces application latency.
- It reduces the possibility of errors.
- It offers advantages in terms of security.
A good example of a use case is autonomous cars, which need constant data traffic to monitor the status of the vehicle, traffic management and communication between vehicles. However, edge services are mainly used by devices of the Internet of Things (Internet of Things, IoT) – one of the key components of smart cities – as well as by artificial intelligence applied to industry assembly lines.
5G and the cloud push the edge to new limits
The on-going deployment of 5G networks will increase the volume of data processed in edge environments and facilitate an explosive growth in the machine-to-machine communications of the Internet of Things.
According to Cisco data, by 2023 half of internet connections will be machine-to-machine. It is important to remember that the latency requirements of these types of connections, which cannot exceed 20 milliseconds, are much more demanding than those of the average user.
In certain use cases, such as industrial automation or autonomous transport, these requirements can be even more demanding, with latencies down to 1ms or less.
Hybrid cloud and edge architectures
A hybrid architecture of interconnected IaaS services with locations close to the user allows a company to combine both approaches, cloud and edge, so that users can take advantage of the low latency of edge computing and the high performance, security and flexibility of a private cloud.
The interconnection between edge devices and an IaaS platform enables easy management of multiple locations. Being able to manage and access data from any geographic location at any time also favours remote work policies.
It is therefore essential to partner with a firm capable of offering neutral connectivity, to ensure the most reliable, fast and appropriate connection for each location. An additional benefit of having such a partner is the ability to scale connectivity while maintaining the management simplicity of dealing with a single provider.
However, even this architecture model – characterised by the high performance of an IaaS environment with high-speed and reliable 5G connectivity – needs an edge layer capable of acting as autonomous distributed nodes that can analyse and act on real-time data. This is to ensure only the necessary data is sent to the main infrastructure in an IaaS instance.
An example of this type of architecture is the one proposed by Microsoft, which has a range of edge devices capable of executing intelligent applications and models that incorporate artificial intelligence, machine learning, and algorithms capable of analysing data and acting in real time even without connectivity.
The data is also transmitted to the cloud, where superior computing power can be harnessed to update and train algorithms before sending it back to edge devices.
The advantages of IoT devices include a higher quality of analysis (which improves decision-making), the ability to detect problems in a production chain before they occur, and an ability to optimize and improve the efficiency and profitability of operations. Such advantages are only possible through the analysis of the large volume of data generated by these devices.
In this respect, an IaaS infrastructure, on account of its superior storage and computing capacity, is essential for analysing large volumes of data in a cloud environment. In addition, the elasticity of an IaaS solution allows the infrastructure to be resized and adapted to the moments of greatest need, increasing resources to process the data and reducing them once the analysis has finished. This means the infrastructure can also sit well with budgetary concerns, since only the resources used are billed.
Working with an IaaS environment, with direct connectivity to edge devices, also significantly reduces the attack surface. Critical data, applications and devices, such as those of an automated assembly line, have a higher level of protection than those connected to a public cloud.
The future is in hybrid architectures
The future of IT infrastructure is not in a false dichotomy between edge, cloud or on premise, but in hybrid architectures, which take advantage of the best of each approach, adapting them to the specific needs of each company.
By combining the advantages of edge computing in data capture with the storage and processing capabilities of the cloud, companies can take full advantage of the benefits of data analytics to leverage their innovation and optimization capabilities, while simultaneously and effectively managing IoT devices in the edge.