Technologies involved in Edge Computing

Edge computing is an emerging technology that brings data processing and artificial intelligence as close as possible to where data is generated, such as sensors, IoT devices and cameras, rather than sending it to the cloud for processing. This reduces latency and increases the efficiency of data processing in real time. In this article, we will explore some of the main technologies that make it possible to implement Edge Computing, but if you would like to go deeper into this paradigm, I encourage you to read our previous post.

Connectivity

To talk about Edge Computing technologies, we must start with connectivity.

Edge Computing networks are used to connect edge devices with data centers and the cloud. Edge networks can use various types of connectivity:

⦁ Long-range connectivity where we encounter both licensed band networks such as the well-known 2G/3G/4G/5G, as well as unlicensed LPWANs such as Sigfox and LoRa primarily.

⦁ Local area connectivity where we can find different protocols such as Bluetooth, Wifi6, UWB, Zigbee, Thread , RFID/NFC, etc., as well as some more innovative communication models such as V2V (vehicle to vehicle communication) or V2I (vehicle to infrastructure communication).

Protocols in licensed bands are making it possible to move from simply sending voice over the 2G network to being able to include critical communications with 5G networks.

A few years ago, we could not imagine that an operator could control a machine remotely, mainly due to the slowness between the time the operator signalled that he wanted to carry out an action and the time when the order reached the machine and was executed. Thanks to communications with low latency, these business models are already a reality.

We have commented above on the number of existing technologies that enable connectivity. One of the fundamental steps in implementing an edge computing solution is to choose the right one that best fits our business model.

To this end, here is a summary of the characteristics that each of them can offer us. As we can see, each of the networks ranging from sigfox, lora to 5G have very different characteristics and which one is the best fit for our project will depend largely on the device we need to provide connectivity.

A clear example: If we need to have connectivity on a device located in a crop field, we may need to have a long battery life, so in this case we might choose to make use of a LoRa network, while if we need to have high speed, we might want to go for a 5G network.

Another factor to take into account when choosing the communication network is that it may be owned by a telecommunications operator. In this case, we must also take into account the services that we can offer. But what services do we mean? The 5G MEC (Multi-Access Edge Computing).

What is the 5G MEC?

Today, MEC is broadly defined as an evolution of cloud computing that uses mobile technologies, cloud services and edge computing to separate application hosts from the data centre where they are located and move them to the edge of the network.

And this is where operators have seen a new opportunity, being able to provide the design, configuration and operation of private 5G networks and low latency for companies, seeking to accelerate the deployment of MEC in their current network infrastructure

How does this work exactly? Operators provide storage and processing resources at the antenna by putting servers in the same booth. In this way, we do not need to go through the antenna, then through the internet to get to the cloud, but we will be able to upload our software on these servers. This way we can carry out ultra-low latency operations.

Artificial Intelligence and Machine Learning

Artificial intelligence and machine learning are used to make real-time decisions on edge devices without the need to send data to the cloud for processing. This reduces latency and increases the efficiency of data processing.

Can you imagine being in a lift and it stops working because it has lost connectivity? No, right? For that it is important that we equip our edge with the intelligence to be able to:

⦁ Generate new business models, improve operational effectiveness and productivity by giving value to all data generated at the edge
⦁ Detect anomalies based on edge data processing
⦁ Make better decisions
⦁ Predict customer or process behaviours by feeding Machine Learning systems.

Due to the vast amount of data generated in the Edge, we cannot expect to upload all this information to the cloud to carry out the appropriate analyses, and that is why it is essential to be able to provide artificial intelligence to the Edge based on the above aspects.

Electronics and Firmware

Edge Computing hardware is an important part of technology. High speed and low energy consumption processing devices are used, such as IoT gateways, edge servers, routers and switches, to process data as close as possible to the source of the data. Edge Computing devices can also include sensors and actuators that collect and process data.

Data analytics

One of the fundamental problems with the Edge is the sheer amount of data that is generated because it happens in real time.

Platforms such as PowerBI or Microsoft BI are not ready to take on such a volume of real-time data, nor are they ready to work on the Edge, they are ready to work in the cloud.

However, having all this information travel from the Edge to the cloud in real time to be processed by these data analysis tools will cause us to incur excessive costs, quickly drain the battery of the devices, have data privacy issues, etc.

That is why it is very important to define how we filter the data to be sent from the Edge to the cloud, how often we will send the information or with what accuracy.

Platforms and Ecosystem

Edge computing platforms are software that run on edge devices and provide an abstraction layer so that applications can run more efficiently. Edge platforms also provide development tools to create edge applications.

Below are some of the platforms that are used for edge computing projects, in terms of applications, cloud tools, connectivity and networks and endpoints.

Security

Security is a major concern in Edge Computing. Edge devices can be vulnerable to attacks and must be properly protected. Technologies such as end-to-end encryption, role-based security access and device authentication are used to protect border data and devices.

Some of the technologies that help us to secure our Edge projects are MDM/EMM solutions through vmware or ivanti, platforms that allow us to perform identity management such as Okta or Azure Active Directory, or MTP/EDR solutions such as Lookout, vmware Carbon Black.

Conclusions

In short, Edge Computing is based on hardware, networks, platforms, artificial intelligence, security and management tools to carry data processing and artificial intelligence as close as possible to the source of the data. By using these technologies, the efficiency of data processing can be improved, latency can be reduced and the ability to make real-time decisions can be enhanced.

At Seidor we have experience in implementing this type of solution and we will be delighted to help you improve your business processes through this new paradigm.