Best Practice of Baidu AI Cloud Native Edge Computing
With the development of 5G technology, edge computing has emerged as a new technology with the advantages of low latency, low cost, and scalability. However, from the perspective of developing operati …
Talk Title | Best Practice of Baidu AI Cloud Native Edge Computing |
Speakers | Ding Han (Architect, Baidu AI Cloud) |
Conference | KubeCon + CloudNativeCon |
Conf Tag | |
Location | Shanghai, China |
Date | Jun 23-26, 2019 |
URL | Talk Page |
Slides | Talk Slides |
Video | |
With the development of 5G technology, edge computing has emerged as a new technology with the advantages of low latency, low cost, and scalability. However, from the perspective of developing operation and maintenance, the traditional Internet architecture is less supportive for edge computing.Therefore, we adopt the cloud native technologies. However, the diversity of business scenarios determines that the cloud native architecture is not a cure-all, and network, storage, and security all need to be upgraded and transformed.Compared to the cloud native architecture, we have added multiple NIC support, high defense and flow control, scalable edge storage, and edge multi-container runtime support. Through the above series of transformations, the business can carry 100% up traffic limit, container security is also upgraded from kernel sharing to kernel isolation, combined with high defense and flow control strategies, security is further guaranteed.