Architect Your Edge Cloud Solution: Kubernetes Containerized or Windows Server Management

posted in: News | 0

Today, many varieties of intelligent applications are implemented on cloud platforms which help enterprisesimprove productivity and flexibility. Many Small and Mid-size Enterprises (SME) have chosen to use publicclouds out of cost considerations. But enterprises in certain industries like semiconductors and CNCmanufacturing have deeper concerns over data sensitivity, internet security, and site restrictions—in short, theywould prefer to have their own private cloud. However, most private cloud systems are designed for largebusinesses, requiring large capital investment and system infrastructure, making them unaffordable for mediumand small businesses. To address the needs of SME for private cloud services, Advantech recently rolled outedge cloud solution—an all-in-one solution incorporating Advantech hardware and WISE-DeviceOn IoT devicemanagement software to provide data collection, management, visualization, and other on-demandmicroservices. The system offers high scalability and SME customers can manage all their resources from oneconvenient terminal. They can also choose additional services to create and organize their own private cloudsystem that will fit their needs and their budget.

Kubernetes and Windows each offer unique advantages in their own ways

To boost flexibility and scalability, Advantech designed their edge cloud solution with a dual architecture. BothLinux-based Kubernetes and Windows Server environments are supported. Kubernetes, is currently the mostprevalent container management architecture for Linux. It was created to address the shortfalls of traditionalarchitectures for data centers in terms of system relocation flexibility. In contrast, traditional architectures requireresources to migrate data to other platforms, while Kubernetes is an open source platform for managingcontainerized workloads which can be moved anywhere quickly and conveniently. For example, if a manufacturer wants to move their company’s data from a public cloud platform to a private cloud each withdifferent architectures, containerized management will be the most efficient way.

Just like app developers traditionally have had to create different versions of programs for iOS and Androidmobile platforms, with a containerized management approach they can simply move their data containers toanother platform for reuse without the need to redevelop programs. Kubernetes helps a lot in shortening systemdevelopment time and increasing operational flexibility.

The Windows server architecture, on the other hand is simple and easy to use, on which users can access allthe services and advantages of WISE-DeviceOn directly without extra development effort. So, even SME whichdo not have strong in-house teams of developers can quickly build their own systems and integrate, manage,and visualize their data themselves. All this can be done via the simple and easy-to-use interface provided by Edge Cloud Solution along with Advantech’s technical support.

Looking at the two versions in practice, Edge Cloud Solution for Kubernetes runs on the Linux Ubuntu OSdistribution, while the Windows Server runs on Windows Server 2016. In terms of expansion, users ofKubernetes only have to increase hardware to extend the original system performance seamlessly; while theWindows server can run WISE-DeviceOn only on standalone system. In terms of integration, Kubernetes allowsthe same microservices to be implemented on different architectural platforms and therefore has greater systemflexibility.

For the application of AI, Windows server has fewer tools for machine learning, while the Kubernetes’s machinelearning platform Kuberflow supports various open source software—such as Tensor Flow and Caffe, which isthe main reason why most AI developers prefer Kubernetes for their AI development environment.

Though private clouds keep data within the enterprise and reduce the possibility of leaks, it is not practical formost SME to construct their own private cloud. In addition to the huge investment on hardware and equipment, italso necessary for the establishment of an IT team specialized in developing, maintaining, and supporting thesystem. Also, it’s difficult to precisely assess the total length of time needed for system development, and after the system goes online it will require another period of time for testing and debugging before it can truly fit theneeds of end user. For many SME sized manufacturers whose resources are restricted, the capital, manpower,and time required for developing their own private cloud would be too much.

For more details, please download whitepaper