 Hello, I am Fabrizio Rovati, I am a software platform and cloud director in ST, and today I want to talk to you about PDT maintenance. PDT maintenance is the technique by which you are able to actively detect when your machinery is breaking down, so you save cost by reducing the maintenance and doing exactly the maintenance when you need it. For PDT maintenance being effective, you need to have something deployed at the node, so near to the machine, at the gateway hedge and at the cloud. So we have done just that. On the node we have here some devices that are using our sensor, our industrial sensor with better accuracy and 10-year availability, our microcontroller SM32 and our connectivity, and also our motor control actuator. And we have made them smart in order to be able to preprocess the data that needs to be gathered, so that you don't really have to send to the cloud all the data, but just the information that you need. For example, here we do FFT of the vibrations so that we don't send all the raw data, but just the frequency that matters. Then we connect these nodes with the gateway that is deploying edge computing frameworks, which means that you can put on the gateway the same things that you would run on the cloud. The advantage is that you are able to close the loop much faster. So, for example, if a machine, if a motor is getting unbalanced, you are able to interact with it much better than you save cost of connectivity and you save cost of storage on the cloud. In some cases, the device will be directly connected to the cloud. For example, here we have another device that is directly connected via cellular to the cloud. The cloud part, here we have demonstrated that we are built as well. What we are able to do is that we are able to enroll the device in the platform and this is done with the state of the art security, so we are using certificates, we are using encryption and secure connection. And what we are able to do is that we are able to track what we want to track and we are able to set thresholds of events that are happening. For example, here we are tracking all these devices, so we are able to see all the data and we are able also to understand if some predictive maintenance condition is happening. For example, this device that we have unbalanced on purpose is clearly needing maintenance for this device. We have set up the range so that we know that here there is nothing to be done. What we are also able to do from the cloud is to gather all the different devices that are at different locations and even connect with different connectivity so that on the cloud side you are able to run all your inference algorithm, for example, of artificial intelligence, gather all the data in the single place in the cloud, be able to refine your methodology and then send it back to the node or to the cloud. So in summary, this is an end-to-end, kind of end-to-end SDK that a customer can take to develop an end-to-end predictive maintenance and not worry about having to set up all the different pieces together. We have done that for our customer and our customer can focus on adding the added value. For more information visit ST.com and thank you for watching.