Presentation is loading. Please wait.

Presentation is loading. Please wait.

UbiStream 2006.10.16. Motivation  Streaming data are abundant in our surroundings: Length of queue at cafeteria If the stadium is crowded or not Course.

Similar presentations


Presentation on theme: "UbiStream 2006.10.16. Motivation  Streaming data are abundant in our surroundings: Length of queue at cafeteria If the stadium is crowded or not Course."— Presentation transcript:

1 UbiStream 2006.10.16

2 Motivation  Streaming data are abundant in our surroundings: Length of queue at cafeteria If the stadium is crowded or not Course video/audio for e-learning Live streaming of concerts or games  Great demands to access these streaming data at any time, any place

3

4 Functional blocks  Turn information in the surroundings into streaming data Camera, sensor, counter  Process the streaming data for further use  The streaming sources and processing units are regarded as services  Look for published services, concatenate them if needed  Retrieve the (processed) streaming data

5 System components  In this architecture, some components are deployed in a static manner Indexing server (discovery mechanism) Cameras (streaming source)  Others are relatively dynamic. We can recruit available computing resources from UniGrid, and install processing programs on them Transcoding server (processing unit)

6 Scenario  User retrieve a group of 16 video screens from live cameras and VoD servers. A set of transcoding server will help aggregate those video sources

7 Sensor Indexing server User UniGrid Camera Video-on- demand Processing server Streaming server Decoder server 1 2 3

8 Workflows  3 main workflows involve: 1. Indirect query From user queries, to server replies an HTML page 2. Interpreting service logic User ’ s browser download decoders (ActiveX), and execute the logic (JavaScript) 3. Streaming data delivery User ’ s ActiveX control fetches media streams

9 Indexing server User Processing server Streaming server Query strings HTML Server translates metadata from XML to HTML

10 Decoder server User Processing server Streaming server URL Browser downloads decoders (ActiveX) ActiveX Browser (IE) executes service logic (JavaScript) Logic interacts with decoders

11 User Processing server Streaming server ActiveX UniGrid …… Transcoding server UniGrid portal Aggregation tree coordinator Request Reply Grid portal recruits machines

12 User Processing server Streaming server Customized query ActiveX UniGrid …… Transcoding server Media stream Decoder fetches streaming data Aggregation tree coordinator Tree root

13 To-do  Decoder Sensor ActiveX client (async. socket, UI) Camera ActiveX client (play/stop) Multiple screen ActiveX client  Transcoding server Receive MJPG, combine, transmit MJPG  Aggregation tree Ask for transcoding servers, launch transcoding programs, wait for feedback  Streaming server VOD server  Video content Prepare MJPG video files


Download ppt "UbiStream 2006.10.16. Motivation  Streaming data are abundant in our surroundings: Length of queue at cafeteria If the stadium is crowded or not Course."

Similar presentations


Ads by Google