Azure Percept Pricing And Features: Edge AI Computing Solution
The contemporary lifestyle has increasingly become simpler and more organized due to the development and adaptation of artificial intelligence in the various fields of daily life. Apart from this, the business arena has also benefited enormously from the new age products and devices that employ artificial intelligence and computing on the edge. A range of instances where Edge AI models are utilized in regular life includes the elevators with speech response which implies that they are designed to respond to voice commands, cameras that are structured to notify the store managers when the shelves need to be restocked, video streams that monitor cash registers, available parking areas and others. In this blog, we will read about one such service Azure Percept from Microsoft.
In this context, Azure Percept is a noteworthy edge AI platform that has been developed by Microsoft as a part of its cloud services to simplify how customers utilize artificial intelligence technology.
Azure Percept Pricing And Features
Meaning and its use cases
The Azure Percept is a comprehensive platform that has the capability of helping user organizations to create intelligent edge devices with the use of the service. This platform consists of responsible AI principles which provide the user with the base sensor access control to the database services. In turn, the use companies then can make informed actions in the process of carrying out the required business actions and in building smart industrial infrastructure. The use cases of Azure Percept can be listed in the following manner:-
- Smart cities can be built by leveraging edge AI to improve efficiency and safety while using the features of digital worlds.
- The scope of making homes, offices and public spaces more safe, secure and smart are limitless with this platform.
- If the retailers can cloud with guidance, the possibilities of reshaping retail with Azure Percept are innumerable.
- The manufacturing processes can be made faster, safer and more productive through this product.
The pricing structure of Azure Percept:
The pricing structure of Azure Percept consists of intelligent edge devices that are a part of a comprehensive platform consisting of added security that facilitates the creation of edge AI-based solutions. The customers can start the proof of concept within minutes with hardware accelerators that are designed for smooth integration with Azure AI Cognitive Services and Azure IoT Hub. In addition to this, the platform works beyond these limitations with Azure Cognitive Services and Azure Machine Learning, and other Azure cloud services to provide vision and audio insights in real-time.
|Azure Percept DK preview
|A pilot-ready development kit for creating edge AI solutions with Azure Percept Vision, a camera-enabled system on the module or SoM for rapid scenario-based customization.
|Azure Percept Audio preview
|A voice-enabled SoM that is extended to Azure Percept DK vision capabilities to audio capabilities like custom commands, far-field speech, and anomaly detection.
|Azure Percept Studio preview
|A collection of Azure products, services, and tools used for device provisioning, deployments, and AI model lifecycle management at the edge. This is free during preview and is fully integrated with Azure Percept hardware. Only pay for the services used
All the parts of the Azure Percept other than Azure Percept Studio are currently available to be bought even in the preview. On the other hand, the Azure Percept Studio is free during the preview period and the user is to pay for the services used. The range of services includes among others, Azure IoT Hub, Azure IoT Edge, Azure Container Registry.
Essential Features of Azure Percept:
The unique features of the Azure Percept can be widely enumerated in the following manner:-
- The service provides an end-to-end edge AI platform through a great presence of hardware accelerators that are integrated with Azure AI and Azure IoT Services.
- There are AI models already built and solution management already provided to help the user to start the proof of concept within minutes.
- The security measures that are built-in to the Edge AI solution of the user can help in protecting the most sensitive and high-value assets.
Hardware accelerators: An Explanation
The term Hardware Accelerators refers to products that help in the process where certain applications offload some tasks to the hardware in the system to accelerate the task. This gives the user access to more performance capacity and efficiency in comparison to the efficiency that could be achieved by using only the general-purpose CPU. As these are in a way specialized forms of hardware, they are better suited to perform certain activities.
Meaning of TPM and Attestation Service: An Azure Percept perspective
The term TPM is the abbreviated form of the phrase Trusted Platform Module version 2.0. This platform can be used to connect the device to the Azure Device Provisioning Service with additional security systems. The TPM is best described as an industry-wide ISO standard provided by the Trusted Computing Group. Being a hardware security module, there exists an attestation procedure while using the Trusted Platform Module.
Software emulated TPMs are best suited for prototyping or testing but they do not provide the same level of security as given by integrated TPMs. The TPMs use a device called the Endorsement Key as the secure root of trust which allows communication with security. This is known to be unique to the TPM and changing it implies that the device is changed into a new one. The process of attestation is quite detailed and can be categorized into the following steps:-
- The Device requests an IoT Hub Assignment – In the beginning, the device firstly connects to the Device Provisioning Service and requests to provision.
- Nonce challenge – The Device then uses the nonce and goes on to use the private parts of the endorsement key and storage root key to decrypt the nonce into TPM.
- Validate the nonce and receive credentials – Finally, the device can sign a SAS token using the decrypted nonce and reestablish a connection to the Device Provisioning Service using the signed SAS token. With the nonce challenge completed, the service allows the device to then provision.
Discussion on the various products included under Azure Percept:
The various parts or products available during the preview can be categorized in the following manner:-
- Azure Percept DK preview
- Azure Percept Audio preview and
- Azure Percept Studio preview
These products can be enumerated as follows:-
Azure Percept DK preview –
This product can be best described as a pilot-ready development kit that consists of a carrier board, mounting tools, and camera-enabled system on module or SoM known as Azure Percept Vision. The SoM is used for rapid scenario-based customization and model development of custom AI-model. The kit is designed to be deeply integrated with the Azure Portal, Azure Machine Learning, and Azure IoT Services. The components of the development kit can be stated in the following way:-
- Azure Percept DK carrier board: i) NXP iMX8m processor, ii) Trusted Platforms Module (TPM) version 2.0, iii) Wi-Fi and Bluetooth connectivity
- Azure Percept Vision system-on-module (SoM): i) Intel Movidius Myriad X (MA2085) vision Processing Unit( VPU) and ii) RGB Camera sensor.
Azure Percept Audio preview –
This device can be described as an accessory that adds speech AI capabilities to Azure Percept DK. The device consists of an audio processor and a four-microphone linear array that is pre-configured. This enables the user to use voice commands, keyword spotting, and far-field speech through the Azure Cognitive Services. It is integrated with Azure Percept Studio and other Azure edge management services. The product can be bought from the online store of Microsoft. The components of the product can be stated in the following manner:-
- Production-ready Azure Percept Audio Device (SoM) with a four-microphone linear array and audio processing via XMOS Codec
- Developer (Interposer) Board: 2x buttons, 3x LEDs, Micro USB, and 3.5 mm audio jack
- Required cables: FPC cable, USB Micro Type-B to USB-A
- Welcome Card
- Mechanical mounting g plate and integrated 80/20 1010 series mount.
Azure Percept Studio preview –
This is a suite of tools equipped with Azure AI edge lifecycle management designed for developers belonging to all skill levels. The Percept Studio is an integration of several services including the following:-
- Azure IoT Hub
- Azure IoT Edge
- Device Update
- Device Provisioning Service
- Azure AI
- Azure Machine Learning
- Azure Cognitive Services and other Azure Services
This suite of services is provided free of cost during the preview period as an additional product with a purchase of Azure Percept DK. The user is required to pay only for the Azure Services put to use.
Features that bring AI to the Edge:
The essential features that come together to bring AI to the Edge for the user organization can be aggregated as follows:-
- The built-in hardware accelerators and the development kit is capable of running AI models without a connection to the cloud. This in turn facilitates the process of performing hybrid cloud workloads.
- The hardware root of trust is the built-in security feature.
- The service has seamless integration with Azure Percept Studio and other Azure services such as the Azure Cognitive Services and Lice Video Analytics.
- The platform is also compatible with Azure Percept Audio which is an optional accessory for building AI audio solutions.
- It also provides support to third-party AI tools like ONNX and TensorFlow.
- The integration with the 80/20 railing system makes way for the Device mounting configurations.
This platform consists of a development kit along with an intelligent camera known as Azure Percept Vision. In addition to this, the platform also provides a “getting started” experience which is called the Azure Percept Studio which allows customers with little or no coding experience to train and deploy proof of concept ideas.
Sas Chatterjee is a Senior Architect with EPC Group. His focus lies in making sure that the execution of each engagement is delivered in a forward compatible, best practices manner. Sas is an extremely devoted professional and takes each project he is assigned very seriously. During the project execution phase, Sas invests the time needed with his clients to gain a full understanding of their requirements and develops a roadmap for achieving their desired end goal.