In recent years, a new type of information processing called “edge computer” has attracted attention. “Edge computing,” means that the Internet of Things (IoT), where everything connected to the Internet and can use many data, is becoming a reality. Edge centralized processing is also the decentralized processing of information, which enables fast and efficient processing of information and greatly increases expectations.
Background to the rise of “advanced computing“
In the early 2000s, the idea of ”cloud computing” emerged, with servers, networks, and storage located in the data center efficiently managing and utilizing these resources on the network.
The cloud started with a public cloud on the Internet and then created a drive to upgrade devices in corporate data centers as private clouds. Since then, several clouds have appeared, but now they are for Amazon Web Services, Microsoft Azure, and Google Cloud Platform for Public Clouds, VMware, and Private Clouds. Mainly used OpenStack and so on. In the past, public and private clouds drawn with a paradoxical combination, but today it is common to use a “hybrid cloud” that uses public and private clouds depending on the features of the program. In addition, we have changed from a public / private cloud selection and usage method to a “multi-cloud” using multiple public / private clouds depending on the situation. Here, various devices and information processors that cannot added to the cloud remain, but the idea of ”edge computing” has emerged that works and manages as efficiently as the cloud. In the future, it thought that we will enter the “multi-cloud + edge” era, which in addition to a few clouds, also controls the edges.
Define the edge
The edge is something outside the cloud, but it is a small device that has a wide range and is located near a device such as a sensor. In addition, operators sometimes refer to a group of servers that perform the closest processing to users on a mobile or landline network as an edge, sometimes known as an MEC (Multiple Edge Calculation).
Reasons for adopting “Advanced Computing”
In the age of multiple clouds, there are three main reasons to put your program on the edge.
- Low latency and high reliability
The first reason is that your application needs low latency and high reliability. Low latency is a program that requires a few milliseconds of judgment based on data.
For example, programs that control tools and robots to warn in factories and programs that control the car. This can be a serious delay problem and should dealt with on the edge.
- General processing optimization
The second is that when viewing the entire system, it is more efficient to work at the edge.
For example, by loading data into the cloud, you may be able to reduce line costs by compiling or compressing data close to each device. In addition, for applications such as AI that do relatively heavy processing, it may be better to process the components at the edge. For example, when identifying a person with a surveillance camera, if you just cut the face image from the entire border image and send it to the center to identify the person, you can effectively identify the person throughout the system.
The third is that data cannot sent to the cloud and so on. For security reasons.
For example, you can anonymize image data related to personal information, such as a face on the edge, or just upload processed cloud data instead of images. It is also possible to process sensitive data related to factory production at the edge and use them for judgment.