Private:progress-choudhary

Summer 2012 (RA)

11 May

• Read up about DASH and browsed through the standard ISO IEC 23009-1.
• Set up two virtual machines with Lubuntu 12.04 on them (one to work as client and another as server).
• Compiled Nginx web server with H264 streaming module and necessary patches. Steps required to do so and errors encountered with their fixes are documented here. The server and client information are also available in the document.
• Tested H264 streaming module with an mp4 file. Works fine on our server setup.
• Built vlc player source on client with DASH support. Haven't tested it yet.

• Virtual server IP: 142.58.185.225 VNC server: 142.58.185.225:1
• Virtual client IP: 142.58.185.224 VNC server: 142.58.185.224:1
• For both server and client- Username: nsl Password: nsladmin

Spring 2012 (RA)

Feb 27

• Have figured out a way to demo the system. Prepared scripts and have tested them once. I can present a demo of the system on or after March 5.
• Have started collecting material and writing. I have written a draft of chapter 1 (Introduction) and collected material for chapter 2(Background and related work). I will use a bulk of the material for chapter 3 (Design and Implementation of ESG Server) from my previous report.
• I am a little bit puzzled over chapter 4 (Evaluation). Since this is an implementation project, the only measurement metric I have been able to think of is SNR since it depends on the implementation of PSI/SI. I haven't been able to think of another metric.
• I am behind the original discussed plan. At this point, I think I will be able to send out my first draft by March 11.

Jan 9 - Jan 27

• Finished integration of the ESG server code base into the mtv. Upon conducting overall system tests, I found that the system was generating a TS packet that was not read correctly by the receiver. Analyzing the packets on a DVB Inspector (dvbSAM is no more available for trial downloads), the packets showed continuity counter errors. I had initially suspected that the errors were introduced in the implementation of PSI/SI however, that does not seem to be the case. Taking out the PSI/SI code from the original code base and introducing the FEC code seems to cause problems. I don't know if it has to do anything with the MPE-FEC puncturing feature that was not implemented (I saw this feature mentioned in an old report). I did not look into the MPE-FEC code.
• Changed the configuration manager code to represent programs in the XML format described for ESG to correctly transmit the programs on the FLUTE/IP stack.
• Investigating the design changes I had earlier proposed in the scheduler to migrate to a window based scheduling policy revealed that there will be major code changes required in the testbed to accommodate such migration. I therefore decided to abandon that idea due to shortage of time.
• I am now spending time on the contingency plan as indicated in the project plan, and trying to put together a system that can be used for a demo. I am looking into the scripts that Cheng had used in his demo. I am planning to have this system ready before Feb 10.
• I have not been able to spend much time on the report yet but I will send an updated version by the end of next week on Feb 3 detailing the evaluation I performed. Also, according to the project plan I should have finished the introduction chapter by now but I haven't been able to do it. I will try and get it done along with the above section on testing but I am not sure if I will be able to complete the introduction section by Feb 3.

• Courses: CMPT 894: Directed reading
• Course report here.

Fall 2011 (TA)

• Courses:
  • CMPT 880: Programming Parallel and Distributed Systems

September 28

• Continued to work on issues highlighted in the last report.
• In order to test the ESG implementation, we need to revisit the design of scheduling the bursts. The earlier design did not consider scheduling the PSI/SI packets while creating the actual program bursts. Also the two EPG channels(bootstrap-EPG and actual-EPG) needs to be scheduled at the same time as the media programs. According to the current design the video stream data is first scheduled into bursts and then converted into group of TS packets. The Transmitter generates the PSISI TS packets and inserts them into the outgoing TS buffer as required. This design may lead to situations where the PSISI TS packets may delay the video packets. I am looking at a way to avoid doing this and test the packets directly.
• As for the analysis, I couldn't find any better tool than dvbSAM and the analyzer software we have.

September 13

• Draft of the report with details of design, completed work and challenges ahead can be found here.
• TODO:
  • Convert the report to the thesis format.

Summer 2011 (TA)

• Courses:
  • CMPT 826: Automated Learning Reasoning.

August 12

• Completed ESG transport. Integrating with Mtv.
• Looking for a method for evaluation.
• Could not work on the report much as I was busy completing a course project. Will work on report draft in the coming week.

May

• Summary of what has been done:
  • Finished implementing classes under ESG bootstrap complete with ESG Access Descriptor and ESG Provider Discovery. ** Also implemented ESG representation and encapsulation modules.
  • In representation, I only implemented the TextualDecoderInit to support XML fragments in only Gzip format which seems to be the standard.
  • ESG data repository containing that carries the different ESG fragments has also been implemented.
  • Currently, I am working on:
    • ESG transport which involves using the PSI/SI table fields to announce and transmit the ESG fragments.
    • The final report in the suggested format.

Spring 2011 (TA)

• Courses:
  • CMPT 885: Special Topics in Computer Architecture
  • CMPT 771: Internet Architecture and Protocols.

April 8

• I have been busy with my course project for CMPT 885 and I am currently working on the report. I haven't been able to allot time for the ESG project in the last two weeks.

Mar 28

• Resumed work on ESG representation. Finished implementation of all four classes.
• Started implementing ESG Encapsulation and so far I have finished the ESG Data Repository implementation.

Mar 14

• Work on course projects.
• Implemented the TextualDecoderInit and ESGInitMessage classes but have not been able to test it yet.

Mar 7

• Continuing implementation of ESG representation. Could not make a lot of progress this week due to course projects.

Feb 28

• Continuing implementation. Working on ESG representation.
• Updated design document with changes in ESG representation and ESG encapsulation. The document is here.

Feb 21

• Started implementation. Implemented classes for ESG Bootstrap.
• Downloaded sample ESG files to analyze fragment information. This will help in deciding what fragments to send. (Incomplete)

Feb 14

• Was busy with course project.

Feb 7

• Further updated design document after understanding the transport structure of ESG.
• Could not spend a lot of time on the project this week.
• Will start implementation this week.

Jan 31

• The updated design document may be found here: /nsl/students/choudhary/projects/PSISIESG/documents/reports.
• Could not do much this week.

Jan 24

• Prepared a class diagram for ESG server on paper. Currently digitizing it.

Jan 17

• Analyzed nokia.ts using SAMalyzer and SAMCorder in dvbSAM.
• Skimmed through ETSI TS 102 469, ETSI TS 102 472, ETSI TS 102 471, ETSI TS 102 591 and ETSI TS 102 592.
• Planned a possible design for ESG server in testbed.
  • Open areas:
    • Determine how to test if ESG is sent correctly. One possible way is using SAMCorder's Navigation option.
    • Finding a standard compliant receiver for testing ESG. Nokia implementations have proprietary components.

• • TODO:
    • Try receiving WingTV TS files on Nokia phones. If this experiment works, we could model our TS files on WingTV TS files as they are standard compliant.
    • Fixing MPE-FEC and continuity counter errors discovered using SAMCorder in current implementation.

• Read a paper on delivery of 3D video over DVB-H.

Jan 10

• Investigated requirements for ESG server implementations referring ETSI TS 102 471.

Fall 2010 (TA)

• Courses:
  • CMPT 705: Design/Analysis Algorithms
  • CMPT 820: Multimedia Systems

Summer 2010 (RA)

• Courses: None.
• Worked on developing signalling mechanism (PSI/SI tables) for DVB-H mobile TV testbed.

Spring 2010 (TA)

• Courses:
  • CMPT 886: Special Topics in Operating Systems
    • Worked on project to propose a dynamic prefetching mechanism for multicore processors.
  • CMPT 822: Computational Vision