Difference between revisions of "BindSight"

From Hack Manhattan Wiki
Jump to: navigation, search
(Pipelines)
Line 13: Line 13:
 
==== Pipelines ====
 
==== Pipelines ====
  
===== SnapSource => Broadcast => unit tests =====
+
===== SnapSource Stage =====
* snapsource ought to ignore demand and just cycle over camera snaps
+
* tests ought to max_demand equal to test sizes
+
* clear up extraneous poor folx's debugging
+
* consumer-producer to handle broadcast
+
  
===== Flesh out Pipelines Further =====
+
* revert changes
* auto-flush final buffer (small size) when backed up
+
* ignore backpressure, just dispatch events from task loop
 +
* test grab a batch
 
* catch dead cameras rather than erroring out repeatedly
 
* catch dead cameras rather than erroring out repeatedly
 +
 +
===== Broadcast Stage =====
 +
 +
* test grab a batch
 +
 +
===== Further Stages =====
 
* consumer-producer to check for bad frames
 
* consumer-producer to check for bad frames
 +
* consumer-producer to queue multiple events into single batch event (handle_demand)
 +
* retool bad frames checker to max_demand: 1, task_async over batch event, dispatch as events (ignore backpressure)
 
* consumer to flush and track quality
 
* consumer to flush and track quality
 
* web consumer as Task
 
* web consumer as Task

Revision as of 14:46, 4 August 2019

Github icon.jpg Github Repository | Bricodash screenshot.png A Bricodash Sub-Project | Beads Land-Trujillo.jpg A Beadsland Creation

Concurrent frame-scrubbing multiplexing webcam gateway daemon.

Devoted service to stream doorcam and spacecam to Bricodash and public gateway, respectively, while tracking activity and performance of these and other webcams at the space. Will be more efficient and reliable than spawning PHP and Python processes on an as-they-come basis.

Written in Elixir, will be taking advantage of various new features of the language, building on the strengths of Erlang/OTP, including Mint (web client), GenStage (backpressure event pipelines) and ultimately mix release (build-time deployment packaging).

Current project plan:

Pipelines

SnapSource Stage
  • revert changes
  • ignore backpressure, just dispatch events from task loop
  • test grab a batch
  • catch dead cameras rather than erroring out repeatedly
Broadcast Stage
  • test grab a batch
Further Stages
  • consumer-producer to check for bad frames
  • consumer-producer to queue multiple events into single batch event (handle_demand)
  • retool bad frames checker to max_demand: 1, task_async over batch event, dispatch as events (ignore backpressure)
  • consumer to flush and track quality
  • web consumer as Task
  • web consumer poolboy

Streaming

  • implement stream as snapshot flipshow
  • ensure recovery when flipshow fails
  • reimplement stream to consume source stream
  • ensure recovery when source stream fails

Advanced

  • provide for throwing timeouts
  • implement greytoss checking
  • configure to launch as daemon
  • bootstrap to obtain dependencies and compile cold
  • incorporate sous veil / swap out PHP touch points
  • incorporate upt chk / swap out Python CGI touch point

Performance

  • investigate periodic multi-hit events: server severing connections prematurely? badly behaved client device?
  • investigate specific issue of pishop launch causing multi-hit failures on all other clients