My current idea is to expose one service to let client request a streaming, then server allocate unique tag id sets and prepare a set of server side objects, including context, stream object. When the client gets the tag id set, it can consequently using prepared streaming to communicate with server side where those asynchronous streaming facilities are ready to work. Does this make sense?
在 2017年9月10日星期日 UTC+10下午10:59:15,
jinyi.s...@gmail.com写道: