Читать книгу Fog Computing - Группа авторов - Страница 33
1.5.1.1 Server Heterogeneity
ОглавлениеDifferent to the common IaaS/PaaS-based cloud services, which are virtual resources, fog services are hosted on resource constrained physical equipment that has limited computational power and networking performance. Therefore, when tenants intend to deploy their applications, they need to consider the compatibility, connectivity, interoperability, and reliability of the fog servers. Specifically, we can classify the heterogeneity of the fog servers into two aspects: hardware type and software type.
Hardware type. Represents the hardware component specification and configuration. In detail, the provider should clearly provide information on the hardware in terms of the computational resource specifications, such as CPU model code and speed, RAM model code and speed, read/write speed of storage, independent or integrated GPU, vision processing unit (VPU), field-programmable gate array (FPGA), application-specific integrated circuit (ASIC), AI accelerator, etc.; the available networking resources specification, such as IEEE 802.11a/b/g/n/ac, Bluetooth LE, IEEE802.15.4, LoRa, NB-IoT, etc.; extra components such as inbuilt or connected sensors that are accessible via the API provided by the fog server. Furthermore, if the fog server is hosting on a mobile Fog node, the provider should also provide the corresponding mobility-related information, such as the route of its moving path, the moving speed, and so forth.
Software type. Denotes the software application deployment platform supported by the fog server. For example, the fog server may support VM-based service which allows a flexible configuration. Alternatively, the fog server may provide a containerization-based platform service, such as Docker (https://www.docker.com). Whereas, for the resource-constraint devices, which can serve only a limited number of requests, the provider may configure a FaaS-based platform that allows the tenant to deploy functions on the fog servers to provide microservice to tenant-side clients instead of the completed applications.