Participants

@Tina Tsou@Jeff Brower Victor Lu @Wilson Wang @Moshe Shadmon @Jeff Brower Oneli AI @Caleb @Haruhisa Fukano 

https://bytedance.us.larkoffice.com/minutes/obusqmhssu862s7y6bt44m33?from=from_copylink

Meeting Summary

The meeting discussed the use of a communication tool called Lark to hold a meeting, and tested whether everyone could use it. Participants included Wilson Wang, who recommended Tina Tsou to use Lark, Victor Lu, who is working on edge CLA, Wan Li, who is interested in AI and HIOT, and Caleb, who graduated from the same university as Tina Tsou and is now working as a software engineer in an IoT company in Dallas, Texas. The meeting also discussed Edge Lake, a project that transforms edge nodes into a virtual data lake. Wilson Wang said Edge Lake is a platform that could add whatever services that you want. It has a hardware obstruction that allows you to look at wherever you deploy Edge Lake from the outside as if it is the same thing and is operated in the same way. It also has command line interfaces and API that you could talk directly to the node, enable services, run queries and so on. The conversation discussed how to manage data and hardware in edge computing environment. The CLI and APIs are on top of the abstraction layer, which sits on top of services. The abstraction layer maps physical components to logical names, regardless of the actual hardware configuration. A shared metadata layer implemented in a blockchain allows applications to understand what's available on each edge node and makes it appear as a single machine. Users push data to the nodes, and a query is sent to one of the nodes, which acts as a gateway to the data. The node looks up the metadata to determine where the relevant data resides and then processes it locally, returning the results to the application. This approach provides low latency, real-time data processing, and eliminates cloud costs. The demo showed how this works in practice. The summary is as follows: The presenter showed the slides of Edge Lake and explained why he was interested in the group and the project they could potentially work on together. He mentioned that federated learning is a very efficient approach for AI, and briefly introduced how the blockchain is leveraged in this setup. He also showed a demo of Edge Lake's functionalities, such as real-time monitoring of resources and rule engines. Finally, the presenter invited comments and questions and expressed his interest in working with the community to develop and integrate the technology presented.​

@Wilson Wang