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

Hybrid multiaccess computing ensures network scalability by introducing guidelines that constrain distributed applications to exhibit low latency and consume less power. This can be best achieved by locating and operating workloads at the mobile or network edge, which means maximizing communication performance between the local network or device endpoint onto the Internet. Products surrounding the IoT and computer infrastructures that rely on minimizing latency or collecting real-time data would benefit greatly from this. Furthermore, edge computing embedded in hybrid and multicloud models helps with encrypting data before sending them onto the cloud network, minimizing risks of data breach and complying with privacy regulations.

Telecom services providers are slowly becoming cloud service providers in an effort to make cloud resources more readily available on demand at the edge of the network. To do so, they are implementing hybrid multiaccess computing solutions and will eventually compete against or coordinate with the big cloud vendors. Businesses and individuals will leverage telecom service providers’ edge cloud resources to develop “bring your own cloud” solutions. A bring your own cloud (BYOC) solution involves an organization allowing employees or users to freely decide on the cloud vendor that best suits their tasks rather than standardizing a single specific provider. Services such as AWS and Azure offer their edge-specific tools and frameworks to process data locally and deploy and manage edge devices for real-time analytics and machine learning. In the same way as Starbucks (as an example) plans to provide charging station services for electric cars shortly, it will also go beyond providing simple wireless access to its customers by being able to connect them to the cloud resources needed to operate cloud applications (e.g., extended reality applications that the store anchors to objects in the cloud.)

Looking at the near future, it is likely that the next generations of cloud-based solutions will evolve toward multiaccess networking architecture. Note that this will not change the need for hybrid or multicloud workload deployment or the use of cloud infrastructure or platform services. What will change is which organizations will make these services available closer to the edge of the network to maximize performance and enhance users’ experience. In fact, machine learning and large language models such as GPT and Bard may still need to be trained in remote big clouds depending on their resource requirements, but the resulting models will be moved and operated at the edge.

This discussion highlights the need to understand how to leverage cloud infrastructure and platform services to support the development and operation of modern applications. In general, cloud services are available today via the big cloud portals (e.g., portal.azure.com) and via software development kits (SDKs) or APIs readily accessible from pretty much any programming language used to develop cloud applications. In the next two sections, we will delve into the details of how to use these services practically to implement cloud applications.