Initiatives to Improve Quality of Service to Remote Centers

Currently IIT Bombay has an ongoing project in satellite based Distance Education [1] where the live lectures in post graduate level IT subjects are multicast to fifteen centers across six states. Participants at the Remote Centers (RCs) have the facility to ask questions, and clarify doubts by having answers from IIT faculty in real time. Moreover, the participant asking question is seen and heard by participants in other RCs as well as by IITB faculty. The classes are held in the mornings and evenings to enable IT teachers of engineering colleges and working IT professionals to pursue studies while continuing their jobs. Participants can accumulate credits toward a post-graduate diploma from IITB in the credit mode. They are also provided with an option of taking the course only for self improvement in the non-credit mode. Apart from differential fees for credit and non credit modes, there is an additional fees discount to bonafide students and teachers.

While the program has been functional for over couple of years, the feedback from participants indicated that measures to improve the Quality of Service (QoS), were needed to make the lectures more effective. Crisp images, clear sound and smooth movements typically measure the perceived QoS by the participants. Such measures were also expected to lead to greater acceptance of the program by industry and academic institutions.

Thus, the objective of the project can be summarized as:

Enhancing competency of IT teachers and industry professionals through distance education by ensuring better QoS through:

§ Providing reliability of mission critical components of the existing system

§ Employing better digital techniques for recording and storage

§ Improving offline transmission of lecture material by efficient File Transfer Protocols (FTP) or On-Demand Video Streaming

§ Deploying enhancement techniques for satellite transmission and streaming of lectures.

In order to ensure that the lectures are received with acceptable quality and without interruption, the following parameters have to be considered:

§ Bandwidth: a certain minimum bandwidth is required to ensure acceptable quality of reception at the RC.

§ Data loss: Loss of data packets has to be minimized to have a good quality of reception. In the case where the channel is loss prone, error correction techniques have to be applied to render the transmission without glitches.

§ Delay Jitter: When packets arrive at the RC with varying inter packet delay, it causes disturbances in the reception. To improve quality, it is important that inter packet arrival is kept constant at the RC so that the resultant play out is smooth. Generally, buffers are used to mitigate this problem where media packets are buffered and played out at a constant rate.

Given a bandwidth, data loss and delay jitter have to be minimized. Research in this area has been taken up as part of the project.

In this report, the QoS issues and the mechanisms implemented for improving the quality at the RCs for Distance education are discussed. Section 2 provides details of the financial aspects of the project. In Section 3, the satellite model along with the mechanisms implemented for QoS enhancement is discussed. Section 4 outlines the enhancement techniques explored as part of the research and development toward providing better QoS. Section 5 provides a discussion on other options and techniques considered for improving the QoS and satisfaction level at the RCs. Conclusions and future directions are provided at the end of the report in Section 6.

1.2. Current status

List of Fifteen Participating Remote Centers (RCs)

Indian Institute of Technology Bombay

Center for Development of Advanced Computing, CDAC Khargar

Mahindra British Telecom Ltd, Mumbai

Mahindra British Telecom Ltd, Pune

Sinhagad Technical Education Society, Vadgaon, Pune

Sinhagad Technical Education Society, Karve Road, Pune

Persistent Systems Pvt. Ltd, Pune

Visvesvaraya National Institute of Technology, Nagpur

MGM College of Engineering, Nanded

Virtual Institute of Training, Goa University, Goa

Nirma Institute of Technology, Ahmedabad

MJ College, Jalgaon

Shri G. S. Institute of Technology and Science, Indore

VJTI, Mumbai

Institute for Development & Research in Banking Technology, Hyderabad

Indian Institute of Information Technology and Management, Kerala

The courses taught are high end IT courses at Post Graduate level. Recently some related theoretical courses as well as courses in management area are added.

1.3. List of courses offered:

Code	Name
OOPJ	Object Oriented Programming using Java
MIS	Modern Information Systems
EE603	Digital Signal Processing, Wavelets and its Applications
IT 605	Distributed Systems
COOS	Computer organization and Operating Systems
DSAL	Data Structures and Algorithms
IT 621	Foundation Algorithms – Part 1
IT 602	Multimedia Systems
IT 606	Embedded Systems
IT 601	Mobile Computing
IT 603	Database Management Systems
IT 619	Foundation Lab
IT 605	Computer Networks
EE 659	A First Course in Optimization
MG 655	Corporate Strategies
MG 647	Innovation & Entrepreneurship
CS 317	Database and Information Systems
CL 692	Digital Controls
IT 628	IT Project Management
IT 653	Network Security
CSE	Advance Programming in C++

2. Section 2

This section gives details of the projects implemented during the year December 2004 till October 2005.

2.1. Financial Summary

	Consumables	Contingencies	Equipment	Pay/ Allow	TA/DA/ TTR	O/H	TOT
Total Funds Allotted	*1,820,000*	*300,000*	*4,000,000*	*400,000*	*300,000*	180,000	7,000,000
*Less Expenses as on 30.09.05 (A)*	*1,029,724*	*196,647.00*	*1,898,608*	*281,802*	*215,130*	*180,000*	3,798,880
Balance as on 30.09.05	790,276	103,353	2,101,392	118,198	84,870	-	3,195,058
Committed expenses details
COMMITTED EXPENSES (B)	606,388	13,120	844	5,000	6,084	-	631,436
*Total Expenses (A+B)*	1,636,112	209,767	1,899,452	286,802	221,214	180,000	4,436,378
NET BALANCE	183,888	90,233	2,100,548	113,198	78,786	-	2,563,622

Table – 2.1: Summary of Expenditure Status

2.2. Details of Second year activities envisaged and schedule for completion

To improve QoS at the RCs, several activities were identified. The schedule followed to implement these activities is given in Table 2.2. The activities are represented in the y-axis and the time period is represented in the x-axis. Table 2.3 lists the target activities and the status of these activities till the end of this report period, June 2004.

S.no.	Target	Status	Date of completion
1.	Testing of UPS units in DEP	Completed testing of the parameters	15 June 05
2.	Implementation of Reliance Telephones for quick reliable communication.	Completed.	15 July 05
3.	Development of Software for text interaction between remote centers and faculty	Completed development, testing and setup.	20 July 05
4.	Setup Audience Response Capture System	Completed the setup and testing of system.	30 Aug 05
5.	Survey of possibility for implementing DG Set along with UPS	Completed	10 Sept 05
6.	Survey for RF measurement equipments	In progress.	-

Table-2.3: Target activities and the status of these activities

2.3. Expenditure priorities and details.

The grant available for the first year was sought to be channeled into the following specific categories:

I. Implementation of Reliance Telephones

– For effective and quick communication between the Remote Centers in case of emergencies and also for daily correspondence.

II. Setup Audience Response Capture System

– To capture the response of individual participants at the Uplink Center.

Details of category wise expenditure are as given in Table-2.4 below.

S.No.	Category	Equipments	Qty	Cost (Rs)
I	Implementation of Reliance Telephones		10	32,161.00
II	Setup Audience Response Capture System	Microphones	30	8,548.80
		Mixer	10	14,096.25
		Cable	167m	2,254.50
		Holders	30	843.75
		Hardware (Clamps, Bolts, etc)	40	4,000.00
		Electrical		6,397.00
		Miscellaneous		**
III	Bandwidth + Consumables		1	6,06,388.00

	TOTAL			6,74,689.30

Table-2.4: Expenditure for Equipments (Dec 2004 – Oct 2005)

4. Section 4

4.1. Audience Response System to capture participant interactions

For effective participation at the Remote Centers, especially during the live classroom sessions, it is important that the interactions taking place between the faculty and the students at the Uplink Center is heard clearly at the Remote Centers as well.

The audience response capture solutions available in the market as of now are not capable of providing the right solution tailored for the requirements at DEP Uplink Center, catering to large number of participants (about 100). The solutions available are suitable for situations such as game shows. Moreover these solutions are very expensive which educational institutes may not be in a position to invest in.

We at DEP have come up with a solution which is very effective as well as very affordable.

4.2. The features of the setup are as follows

§ Uses corded microphones spread out throughout the classroom.

Cordless microphones need to be passed around from one participant to another which usually disturbs the attention of the participants. Most of the time it is also difficult to find the location of the cordless microphones, which disturbs the class. Moreover many cordless microphones are required to effectively capture the audience response making the setup very expensive.

Using corded microphones overcomes this disadvantage associated with cordless microphones.

§ One microphone for every three participants.

With an installation of a total of 28 microphones, this ensures that every participant in the classroom has an easy access to a microphone without having to disturb anyone during the lecture.

§ Helps Remote Centers to better understand the discussions at the Uplink centre.

Especially during the live classroom sessions, the interaction between the faculty and the students at the Uplink site is heard clearly at the Remote Centers. This helps the participants at the Remote Centers to understand the proceedings of the session more effectively.

§ Improved capture of audience response.

As compared to the cordless microphone setup and other setups available in the market, this solution is better able to capture the audience response.

§ Very cost effective setup.

This kind of setup is very affordable in terms of cost which can be adopted by any educational institution with such kind of a setup requirement.

4.3. ADVANTAGES

· This setup of fixed microphones has resulted in increased use of the microphones by the participants at the DEP Uplink site as compared to the previous setup wherein cordless microphones were used.

· It has proved to be very effective in lecture sessions where management related subjects are taught which requires extensive responses and interactions from the participants.

· As a result the participants at the Remote Centers attending the courses have a feeling of a lecture well attended and understood because of the responses heard by the participants at the Uplink site, which in turn satisfies most of the queries arising from the participants at the Remote Centers.

5. Section 5

5.1. Remote Center Messaging Software

DEP makes use of application software from VCON®. There are a few drawbacks found in the software system. The current software system does not have the facility to provide 2-ways communication system between Uplink Center and the Remote Centers by way of text messaging. Also this software does not provide any logging facility of any kind.

Another problem DEP was facing in the software domain was of communicating with the faculty’s desk. There was no facility available to send the text based queries asked by students from remote centers to the faculty. Thus it was decided to design and implement the software as an add-on and provide DEP with an efficient communication and logging system.

This Messaging System provides the user with a user-friendly environment for efficient text message communication and floor accessing module. It also maintains a log regarding the log in times of all the participants, quality of audio and video received, problems encountered if any and the time each participant had logged off, queries asked to the faculty, interaction with the console room operator, being some among the many fields in the log file.

Development of this application software was handed over to students in final year of graduation in engineering as their project work. The interface for all the applications was developed using C# which is done by Smitesh Hawaldar. All the three applications communicate with each other through socket programming, which is done by Shreyas Phadnis and the persistent data is provided by the database, which is done by Viswasundar Desaraju.

5.6. DATA FLOW DIAGRAM

Oval: 1
Check User & Password

Username

& Password

USER

LOGIN

Oval: 3
Process User request Oval: 4

Update Database

VCON LOG TABLE

Valid

Administrator Login

vcon log info

log info

Oval: 2
Load
Admin.
System

LOG

TABLE

Update Conversation

CONVERSATION

Fig-5.4: Context Level 1 DFD

5.7. RESULTS AND DISCUSSIONS

5.7.1. Targets achieved

The following main modules were implemented are

· Communication system between IIT, Mumbai and Remote Centers.

· Communication system between IIT, Mumbai, Remote Centers and Professors Desk

· Logging facility at IIT, Mumbai and its retrieval using a web-based system.

The development of this software was further divided in sub-modules:

· Developed IITB Centre (server) interface.

· Developed Remote Centre (client) interface.

· Developed Prof’s Desk (client) interface.

· Data handling in all the applications.

· File handling part.

· Socket programming for the entire communication system between IIT Bombay & Remote Centers.

· Socket Programming for the entire communication system between IIT Bombay & Faculty’s Desk.

· Developed web-forms for log checking system for IIT Bombay

· Creation of the database to maintain logs using MS-Access©.

· Database connectivity for the server to register the logs and updating of the database.

· Database connectivity to the web form developed for querying purposes.

5.7.2. Positive Features

· This communication system developed uses the C# socket class and extends various other C# core library classes to provide a customized functionality.

· The interface is designed more like a chat system considering user familiarity with such chat systems.

· The settings option on the remote centre side has been kept minimum so as to make the application more user friendly.

· Security has been stressed greatly while developing this software. Any PC which wants to connect to the server needs to be registered with Uplink Center first. If a PC is not registered then one cannot connect.

· Another unique feature of this software and a strong requirement of this software was maintenance of log both in flat-files as well as database.

· An advantage of this feature is that in case the web-interface fails, users can view the database using MS-Access. In an event even if the database fails the data can be retrieved from the text files.

· Another feature is that the faculty's desk has been designed in such a way so as to minimize faculty interaction.

· The web-based log checking system is also password protected. Also the interface is very user-friendly and the query can be done based on RC NAME and/or Date of Lecture Transmission

· This software also scan and maintains log from a log file created by VCON. So this feature gives IIT, Mumbai the exact login and logout time of remote centres.

6. Section 6

6.1. CONFORMANCE TESTING OF UPS UNITS.

The purpose of conducting these tests is to provide a benchmark for the Remote Centers in case they want to procure a UPS for backing up their systems.

DEP has procured 3 UPS units of rated at 1.5KVA, 3KVA and 5KVA. The 1.5KVA UPS is being used for backup power supply for the RF transmission equipments at the Uplink Centre. The 3KVA is used for the setup of the experimental Remote Center at IIT Bombay. The third UPS of 5KVA is being used to power the equipments in the lecture hall at the Uplink Center.

6.2. EXPERIMENTAL RESULTS

Tests were performed on three Aplab UPS. These UPS are used to support the transmission equipment used for the DEP Project. The results are as follows:

6.2.1. Backup Time Test – Under Normal Load:

Switched on the entire equipment that is on the UPS. Then switched off the mains supply and noted the time for the UPS to give the first warning beep, and time to shift to the auxiliary supply.

*UPS*	*VA Rating*	*Normal Load*	*Time for first Warning Beep*	*Time after warning for shifting to Auxiliary Supply*	*Total backup Time*
UPS for Uplink RF Equipment	1.5kVA	13.1%	24mins	8mins	32mins
UPS for DEP RC	3kVA	33%	2hrs 20mins	20mins	2hrs 40mins
UPS for Lecture Hall	5kVA	37.2%	65mins	15mins	1hr 20mins

Table-6.1: Backup Time Test

Analysis:

1. The backup time of a UPS decreases as the load on the UPS is increased.

2. There is some error in the performance of the UPS in DEP Studio since it is giving such less backup time at 13% load.

3. Checked if the reading on the LCD of the front panel was correct. The actual load was about 60% while the UPS showed 13%load.

6.2.2. Load Regulation Test – Under Full Load

The load test was performed on UPS in Room No. 201 by connecting different loads on the UPS and noting the output voltage.

*Load*	*Output Voltage*
4% (approx. no load)	229.7V
26.1%	228.7V
48.4%	228.8V
71.3%	228.5V
93.6% (approx. full load)	228.2V

Table-6.2: Load Regulation Test

Analysis:

% Load Regulation: ((VNL-VFL) / VNL) *100 = ((229.7-228.2)/229.7) *100 = 0.65%

As per the specifications of the UPS (see the appendix), the regulation must be +/- 1%.

Hence, the UPS complies with the specifications.

6.2.3. Input Voltage

Minimum value of input voltage : 217.8 V

Maximum value of input voltage : 230.0 V

As per the specifications of the UPS (see the appendix ), the input voltage should have a range of 160V-265V AC one phase.

Hence, the UPS complies with the specifications.

6.3. TEST RESULTS

UPS Rating: 3.0kVA Normal Mains AC: 230V AC

Date: 21/06/2005 DC Voltage:

*S.No.*	*Test Parameter*	*Test Specification*	*Observation*
1.	Make	Aplab	Aplab
2.	Model	Rhino Series	OK
3.	Capacity	3000VA	2400W at UPS output on Lamp load
4.	Technology	IGBT Based Pulse Width Modulation (P.W.M.)	IGBT Technology
5.	Input Voltage (Nominal)	160V-265V AC one phase	160V-265V
6.	Input Frequency (Nominal)	50Hz	Range 47-53Hz
7.	Input Power Factor	>0.96	>0.96
8.	Protection	Over Current and Short Circuit protected	Provided
9.	Voltage Regulation	±1% typical	0.65%
10.	Output Frequency	50Hz±0.5% and ±3% (sync mode)	50Hz
11.	Waveform Type	Pure Sine Wave	Complies
12.	Transfer Time	£ 4 msec	Complies
13.	Load Power Factor	0.6 to unity	0.8
14.	Efficiency	90%	89%
15.	Cold Start	Should allow to start the computer and use the battery power when the mains is absent	Complies
16.	Communication Port	RS232 for UPS monitoring software/ DSS/ Smart Slot	Provided
17.	Battery Charging Time	4 hours for 90%	--
18.	Battery Support	Native as well as std. Car/bus battery, user replaceable- subject to training to user	--
19.	Protections	1.Input Over/ Under Voltage 2.Overload at the Output 3.Soar Protection for IGBT devices 4.Output AC Over/Under Voltage 5.Battery Over/Under Voltage 6.Snubber Circuits for Power Semiconductor Devices	Provided
20.	LCD Display		Provided
22.	Alarm indication	Battery backup, Low battery, Over load	Exists
23.	Audible Noise	< 45 dBA	Site was not suitable

Table-6.3: Test Results

7. Section 7

7.1. TRIALS AND DEMO WITH MPEG-4 STREAMING OPTIONS

Venera Technologies is an Indian software development company specializing in multimedia software. The company’s product line includes software encoders, which can at present encode video signals in composite, and S-video to MPEG-4 compression format, which were mainly developed for Internet based streaming. Trials were conducted by Venera Technologies at IIT Bombay’s DEP cell.

7.2. DEMONSTRATION WITH VIDEO CONFERENCING SYSTEMS

Other means of transmitting the video content to participants at remote locations were tried out. This included trials and demonstrations with Tandberg Video Conferencing system.

The aim of the experiment was to demonstrate the capability of the equipment to stream the video content to receive only terminals within a closed user group connected in the local network or in any other network. The receive only terminals would be then connected to the central site by means of an ISDN or Leased line.

Conclusion of the demonstration

It was found that, though the system was capable of achieving the above mentioned requirement, there were other short comings in the system.

1. The unit was able to stream the video content only in MPEG-2 format. It was desired that the system would use MPEG-4 compression standards since it has much better qualities as compared to MPEG-2. Moreover MPEG-2 is also being used by the current system deployed for transmission through satellite.

2. The system also did not have any logging facility, which would have been of importance considering the nature of the application for distance education.

3. The cost of the system was quite high in comparison to the satellite setup.

Considering these drawbacks and the possibility of achieving similar results by other means, it was decided not to pursue the purchase of video conferencing equipment.