100 likes | 219 Vues
This document outlines a robust architecture for streaming video using Redux, focusing on Quality of Service (QoS) management. The receiver subscribes to a video stream and must maintain a balanced load from the video server, avoiding both starvation and flooding. Key components include distributed QoS code, TCP communication for connection management, and sender-receiver feedback channels. The sender application pulls images, and the receiver processes them, ensuring adherence to Service Level Agreements (SLA) while monitoring CPU usage and managing timing effectively.
E N D
Image Application Redux • Still: receiver subscribes to stream from video source • And: Receiver must not be swamped or starved by video server • Things to note: • Code for QoS gets distributed • QoS code interleaved
Base Code Just TCP communication Binding, connecting, marshalling, etc. No application code No QoS
Sender Control • Sender-side Feedback channel • MUXs control messages • Binds, connects • TCP transport
Sender SLA Implements SLA pieces for sender Param’ed by the SLA Can send messages over the control channel Responds to receiver feedback
Sender Application • Pulls image from source • Uses Infopipe to send image
Receiver Control • Receiver-side feedback channel • DeMUXs control messages • Binds, connects • TCP transport • Mirrors sender’s control code
Timing Code • Inserts variables for time • Inserts system timing calls
CPU Monitor • Calls system for resource usage • Uses timing from timing aspect • Calculates CPU use %
Receiver SLA • Implements Receiver end of SLA • Returns feedback via control channel • Acts based on CPU % usage data
Receiver Application • Processes images received off network by Infopipe