第13章ハイブリッドマルチクラウドのデジタルソリューション開発

By the end of this section, you will be able to:

• Learn how to use Internet of Things cloud PaaS services

• Learn how to use shallow and deep machine learning cloud PaaS services

• Learn how to use blockchain cloud PaaS services

• Understand PaaS services support for extended reality applications

• Understand PaaS services support for 3-D/4-D printing services

• Relate to PaaS services for cloud application development

For businesses that look for all existing features in IaaS, as well as a service that provides a complete platform to develop, test, and launch their applications, platform as a service (PaaS) has become one of the top options. On top of servers, network, and security that IaaS provides, PaaS also includes middleware such as operating systems and development tools that enable quick application development and market launch. As the cloud provider manages more layers, PaaS will require less time and skill from the engineering team to manage their infrastructure. It will provide freedom for the company to focus on developing applications and providing services. However, because of those abstracted services, PaaS may have higher costs with dependency on the particular platform or vendor. The higher cost may come later when the organization explores different technical decisions and cloud vendors.

In this section, we will explore various PaaS services that enable organizations to effectively launch and maintain large-scale applications. This includes AI incorporations or XR platforms that allow for operations of scripting and modeling. Through real-life examples, we can analyze how companies slash costs while also accelerating time to market and application development.

A 5G network enables mobile computing at the edge of modern telecommunication networks with support for a variety of IoT devices, including laptops, smartphones, and smartwatches. 5G has higher radio frequencies, which transfer considerably more data over the air at faster speeds while reducing congestion and lowering latency. Thanks to 5G, more IoT devices can be used simultaneously within the same geographic area. As a result, today’s dynamic information networks consist of interconnected sensors, actuators, mobile phones, robotics, and smart devices.

IoT network traffic falls broadly into two categories: telemetry and telecommand. One category, telemetry, aggregates data generated by sensors and devices and sends them to a server. The other category, telecommand, sends commands across a network to control devices or sensors. Figure 13.9 illustrates the typical flow of IoT data generated by mobile edge devices and data processing and storage via cloud PaaS services.

To serve the purpose of IoT, several application-layer protocols have been developed, such as Message Queuing Telemetry Transport (MQTT), Advanced Message Queuing Protocol (AMQP), Constrained Application Protocol (CoAP), Extensible Messaging and Presence Protocol (XMPP), and Simple Text Orientated Messaging Protocol (STOMP). This was necessary because application layer protocols, such as HTTP, are not suitable for IoT telemetry and telecommand applications. HTTP is designed for one-to-one communication rather than one-to-many communication between many sensors and one server and is a reliable protocol for web applications. However, HTTP supports unidirectional synchronous request-response communication and cannot send data in both directions simultaneously. IoT sensors cannot work efficiently in a synchronous manner. HTTP also is not designed for event-based communication. Generally, HTTP’s scalability is achieved by loading the server, which puts a heavy load on sensor devices connected to multiple other devices. In addition, HTTP uses high power consumption, making it unsuitable for advanced wireless sensor networks.

To understand how IoT application-layer protocols work, consider MQTT, which is an example of a lightweight application-layer messaging protocol. It is based on the publish/subscribe (pub/sub) model typically used for message queuing in telemetry applications. Multiple clients, or sensors, can connect to a central server, known as a broker, and subscribe to topics that interest them. Through the broker, which is a common interface (router) for sensor devices to connect to and exchange data, these clients also have the option of publishing messages regarding their topics of interest. To make the communication reliable, MQTT uses a TCP connection on the transport layer for connections between sensors and the broker.

Big cloud vendors use IoT protocols such as MQTT to provide IoT PaaS services and related frameworks that facilitate the interactions of IoT devices with the cloud. The use of these services makes it possible to build and deploy innovative IoT solutions without developing and managing IoT frameworks on local computers. In particular, Microsoft Azure IoT manages cloud services that can interconnect, monitor, and control billions of IoT assets. An IoT solution is typically made up of one or more IoT devices that communicate with one or more back-end services hosted in the cloud. IoT devices can be constructed as circuit boards with sensors attached that use Wi-Fi to connect to the Internet (e.g., presence sensors in a room). Devices may be prototyped using a Raspberry Pi or the Microsoft MXChip IoT DevKit, which has built-in sensors for temperature, pressure, and humidity, as well as a gyroscope, accelerometer, and magnetometer. Microsoft also provides an open-source IoT device SDK to help build apps on devices. In addition, Microsoft’s IoT Edge and IoT Hub frameworks can be used to facilitate the operation of IoT applications at the edge and the collection and transfer of data to the cloud via common communication protocols such as MQTT and AMQP.

The technologies, PaaS services and solutions provided by Azure IoT are summarized in Table 13.1. AWS, GCP, and IBM Cloud also provide equivalent IoT PaaS services and related capabilities.