Digital content of display distribution system optimization for technology management applications

To meet these requirements, DCDDS looked for its digital content solution with Apache 2.4 Web Server based on the Linux platform which the Event MPM would dedicate the threads for the ne

DIGITAL CONTENT ON THE DISPLAY DISTRIBUTION SYSTEM:

OPTIMIZATION FOR TECHNOLOGY, MANAGEMENT, APPLICATIONS

In Partial Fulfillment of the Requirements of the Degree of

MASTER OF INFORMATION TECHNOLOGY MANAGEMENT

In Computer Science and Engineer

DIGITAL CONTENT ON THE DISPLAY DISTRIBUTION SYSTEM: OPTIMIZATION FOR TECHNOLOGY, MANAGEMENT, APPLICATIONS

In Partial Fulfillment of the Requirements of the Degree of

MASTER OF INFORMATION TECHNOLOGY MANAGEMENT

By Nguyen Phuong Nam ID: MITM05014 International University - Vietnam National University HCMC

Third 2015 Under the guidance and approval of the committee, and approved by all its members, this thesis has been accepted in partial fulfillment of the requirements for the degree

Acknowledgements

First and foremost, I would like to express my sincere gratitude to my lecturer Dr Technology Nguyen Minh Son for the continuous support of my thesis study and research, for his patience, motivation, enthusiasm, and immense knowledge His guidance helped me in all the time of research and writing of this thesis

My sincere thanks also goes to the engineers of HTV, HDLC offering me the resources in their groups and leading me working on this exciting projects

Last but not the least, I would like to thank my family: my parents and my wife for supporting me spiritually throughout my thesis study

Ho Chi Minh City

Copyright Statement

This copy of the thesis has been supplied on condition that anyone who consults it

is understood to recognize that its copyright rests with its author and that no quotation from the thesis and no information derived from it may be published without the author’s prior consent

© Nguyen Nguyen Phuong Nam/MITM05014/2012-2015

ABSTRACT

The main purpose of he Digital Content On The Display Distribution System (DCDDS) is to wherever and whenever you need to catch up on something information, DCDDS would provide an eye catching, cost effective, high impact mobility or static displaying for its target audiences All proposed current multimedia display systems seem to meet this target basically In fact, there are many firms or organizations have only limited to the management of website content, news or local news by text simply These systems are often applied in the field of e-commerce In the huge buildings such as shopping malls, supermarkets, cinemas, amusement parks, high-class apartments, museums, libraries, the multimedia display systems are deployed with the high cost due to long signal cable and multiple controllers, the digital contents stored in the memory of display devices would not changed rapidly and securely, inflexible to integrate with the running IT systems Because the solution for these multimedia display systems do not apply the wireless network communication Therefore, the efficient technologies are the essential need for these multimedia systems The applications with the trend

of wireless connectivity would demand the lower cost for the performance and the security This thesis will go through the operation of the particular DCDDS, as well

as the performance and the security for DCDDS Then the operations and the quantitative assessments for DCDDS are also demonstrated in the open source operating system based on the VMware virtualization platform

TABLE OF CONTENT

ABSTRACT vi

TABLE OF CONTENT vii

LIST OF TABLES xiv

CHAPTER 1 INTRODUCTION 1

1.1 Overview 1

1.2 Problem Statement 2

1.3 Methodology 2

1.4 Thesis Scope and Limitation 5

1.4.1 Thesis Scope 5

1.4.2 Thesis Limitation 7

1.5 Structure of Thesis 7

CHAPTER 2 LITERATURE REVIEW 9

2.1 Introduction of Multimedia Network 9

2.2 The proposed Issues for Future Multimedia Networks 9

2.3 Wireless Multimedia Communication 10

2.4 Security for Wireless Multimedia Network 11

2.4.1 For Network Infrastructure 11

2.4.1.1 The Preshared Key Mode 11

2.4.1.2 The Enterprise Mode 11

2.4.2 Introduction of Video Compression Technologies 12

2.4.3 For The Video Encryption Algorithms 14

2.4.3.1 Fully Layered Encryption 14

2.4.3.2 Permutation based Encryption 15

2.4.3.3 Selective Encryption 16

2.4.3.4 Perceptual Encryption 16

CHAPTER 3 METHODOLOGY 18

3.1 Comparison of Microcontroller Unit (MCU) Technology 18

3.1.1 Tiny210V2 18

3.1.1.1 What Is Tiny210V2 ? 18

3.1.1.2 The Structure of Tiny210V2 19

3.1.1.3 Assessment 19

3.1.2 Raspberry Pi 19

3.1.2.1 What Is Raspberry Pi? 19

3.1.2.2 The Structure of Raspberry Pi 20

3.1.2.3 Assessment 20

3.1.3 Difference between ARM and x86 21

3.1.4 Apache Web Server 2.0 and The Earlier Versions 22

3.1.4.1 Apache MPM Prefork 22

3.1.4.2 Apache MPM Worker 23

3.1.5 Apache Web Server 2.4 with Event MPM 24

3.2 Comparison of Transcoding Technology 25

3.2.1 Grass Vally with Canopus Procoder 25

3.2.1.1 Assessment 26

3.2.2 Harmonic with Promedia Carbon 26

3.2.2.1 Assessment 26

3.2.3 Telestream with Flipfactory 27

3.2.3.1 The Typical Product of FlipFactory 27

3.2.3.2 Assessment 27

3.3 The Operation of DCDDS 28

3.3.1 The Operation of Server Zone 29

3.3.1.1 The Operation of Web Server 30

3.3.1.2 The Operation of H264-Storage 31

3.3.1.3 The Operation of FlipFactory Farm 31

3.3.2 The Operation of Terminal Zone 33

3.3.2.1 The Operation of Terminal 34

3.3.2.2 The Operation of Raspberry Pi model B+ 35

CHAPTER 4 IMPLEMENTATION 36

4.1 Introduction 36

4.2 Virtualization Limitation 36

4.3 Virtualization Purposes 36

4.4 Testing Scenarios 37

4.4.1 Web Server 37

4.4.2 H264-Storage 38

4.4.3 FlipFactory Farm 38

4.4.4 Raspberry Pi 39

4.5 Testing Evaluation and Discussion 39

4.5.1 Web Server 40

4.5.1.1 The Simple Content Test 40

4.5.1.2 Complex Dynamic Content Test 42

4.5.2 H264-STORAGE 43

4.5.2.1 Performance Test for mp4 files 44

4.5.2.2 Performance Test for HD, SD files .47

4.5.3 FLIPFACTORY FARM 50

4.5.4 RASPBERRY PI MODEL B+ 53

4.5.4.1 Consumption of unrefreshed Memory From 10MB To 50MB 54

4.5.4.2 Consumption of Refreshed Memory From 10MB To 50MB 55

4.6 Virtualization Conclusion 57

CHAPTER 5 CONCLUSION AND FUTURE WORK 59

5.1 Conclusion 59

5.2 Future Work 60

REFERENCES 61

APPENDIX 63

Appendix A: Install VMWare Workstation 10.0.0 63

Appendix B: Configure Apache 2.2.15 Web Server and Apache 2.4.2 Web Server 63

Appendix B1: Install Apache 2.2.15 Web Server 63

Appendix B2: Install Apache 2.4.2 Web Server 65

1 Compile and configure httpd-2.4.2.tar.gz with the Event MPM 65

2 Compile and configure php-5.4.4.tar.bz2 69

3 Compile and configure mysql-5.5.16.Tar.Gz 70

4 Compile and configure phpMyadmin-4.2.10 for mysql-5.5.16.tar.gz 71

Appendix B3: Issues of Security, Monitoring and Recording System Logs 72

1 Performing SSH without entering password on OpenSSH 72

2 Monitoring and Recording System Logs 72

2.1 System Status 72

2.2 Rsync Logs 72

2.3 APACHE 2.4.2 Logs 72

2.4 MYSQL Logs 72

Appendix C: Configure H264-Storage 73

Appendix C1: Install H264-Storage 73

Appendix C2: Issues of Security, Monitoring and Recording System Logs 73

1 Performing SSH without entering password on OpenSSH 73

2 Monitoring and Recording System Logs 74

2.1 System Status 74

2.2 Rsync Logs 74

2.3 Monitoring the failure of Raid device 74

2.4 Samba Logs 74

2.5 vsFTP Logs 75

Appendix D: Configure FF1 and FF2 75

Appendix D1: Install and Configure FF1, FF2 76

Appendix D2: Issues of Security, Monitoring and Recording System Logs 77

1 System Status 77

2 FlipFactory Service 78

3 Job Status 78

Appendix E: Configure Terminal 78

Appendix E1: Install Terminal 78

Appendix E2: Issues of security 79

Appendix F: Install Raspberry Pi Model B 81

Appendix F1: Issues of Performance, Security and Quality of Service 81

1 Performance 81

2 Security and Quality of Service (QoS) 82

LIST OF FIGURES

Figure 1 Multimedia information display system of SUN-TECH IDS 3

Figure 2 The proposed DCDDS in Viet Nam 4

Figure 3 Technical Card in Infosys+ 6

Figure 4.The advertising spots of HTV7, HTV9 and THVL1 6

Figure 5 IEEE 802.11 wireless LAN in OSI reference model 10

Figure 6 The model of IEEE 802.11 at physical layer 11

Figure 7 Selective Encryption System 16

Figure 8 Tiny210V2 18

Figure 9 The structure of Tiny210V2 19

Figure 10 Raspberry Pi model B+ 20

Figure 11 The structure of Raspberry Pi model B+ 20

Figure 12 Market share of the top million busiest sites 22

Figure 13 The Prefork Multi Processing Module 23

Figure 14 The Worker Multi Processing Module 24

Figure 15 The proposed model of DCDDS 29

Figure 16 The operation of DCDDS in server zone 30

Figure 17 Factory Farm in DCDDS 33

Figure 18 The operation of DCDDS in terminal zone 34

Figure 19 The testing scenarios in the server zone 37

Figure 20 The testing scenarios in terminal zone 39

Figure 21 Mean transaction time for simple content test 41

Figure 22 Mean transaction time for complex dynamic content test 43

Figure 23 Performance for file size from 10MB to 50MB – IPOPs 45

Figure 24 Performance for file size from 10MB to 50MB - Throughput 46

Figure 25 Performance for file size from 150MB to 400MB – IPOPs 48

Figure 26 Performance for file size from 150MB to 400MB - Throughput 49

Figure 27 The TimeDuration of the FlipFactory server history at HTV 52

Figure 28 The capacity of file on FTP Sever at HTV 53

Figure 29 Consumption of unrefreshed memory for file size 10MB to 50MB 54Figure 30 Consumption of refreshed memory for file size 10MB to 50MB 56

LIST OF TABLES

Table 1 MPEG and VCEG 13

Table 2 Comparison of Video Encryption Algorithms 17

Table 3 Simulation setup parameters for Apache 2.4 Web Server 38

Table 4 Simulation setup parameters for H264-Storage 38

Table 5 Simulation setup parameters for file size from 10 MB to 50MB 39

Table 6 ab parameters 40

Table 7 The data of mean transaction time for simple content test 41

Table 8 The data of mean transaction time for complex dynamic content test 42

Table 9 fio parameters 44

Table 10 H264-Storage performance from 10MB to 50MB for mp4 files 44

Table 11 The default block size in Linux system 46

Table 12 The write penalty of RAID 47

Table 13 H264-Storage performance from 150MB to 400MB for HD, SD files 47

Table 14 IOPs for RAID 5 and minNumDisks 49

Table 15 Performance of the process of transcoding and delivering files 50

Table 16 The data for I/O Write Bytes on the FlipFactory server at HTV 51

Table 17 Parameters for the theory formula 53

Table 18 Consumption of unrefreshed memory from 10MB to 50MB 54

Table 19 Parameters for consumption of unrefreshed memory 55

Table 20 Consumption of refreshed memory from 10MB to 50 MB 56

Table 21 Parameters for consumption of refreshed memory 57

Table 22 mdadm parameters for indicating RAID device failure 74

Table 23 The profile of MPG to MP4HD 76

Table 24 The profile of MPG to MP4SD 76

Table 25 The profile of MPG4 TO MP4TRIM 77

CHAPTER 1 INTRODUCTION

1.1 Overview

Recent years, it has been marked that the rapid, powerful development of information technology and microprocessor technology, that leads many kinds of IT equipment become compact, many functions, and portability Many manufacturers have indeed concentrated on producing the credit card sized single-board computers (SBC) with full of computer ports These single-board computers embedding the open-source OS and support IP based applications Seamless connectivity becomes more popular and preferable for these applications, it means that applications achieve the diversity of using purposes and ready in connection with existing multimedia display devices such as the high definition TV (HDTV) (1920×1080 pixels about 2.07 megapixels per frame) Along with the popularity of IEEE 802.11n standard (frequency 5 GHz, maximum speed 600 Mbps, activities range 50 -125m), the ability of communication and integration into these embedded systems

to distribute and show the digital contents richer and more flexible Which is the purpose of activities from DCDDS (Digital Content on the Distribution Display System) To meet these requirements, DCDDS looked for its digital content solution with Apache 2.4 Web Server based on the Linux platform which the Event MPM would dedicate the threads for the new requests, RAID-based Storage System based on the command line Linux platform which ensures the security and the data integrity as well as the performance in any given situation, Telestream FlipFactory which is a standard for transcoding, delivering the digital contents automatically and securely, Terminal based on the graphic Linux platform which is shell script programmed for distributing the digital contents into Raspbery Pi model B+ securely with the lower cost, Raspberry Pi model B+is a microcontroller which has the capable of a computer for receiving the encoded contents automatically and securely, then decoding and displaying the digital HD signals on HDTV with standard HDMI All of them incorporate together tightly

Therefore DCDDS would be the better candidate for the current multimedia display systems with the ability for the performance and the security Although, this thesis would not cover all technical aspects of multimedia display system DCDDS is nice

to support a diversity of eye-catching contents, including video, static displaying and resolves the problems from the management to the effectiveness of the mode of information transmission Thus DCDDS would be one of the models would applied

in Viet Nam for near future

1.2 Problem Statement

Looking at the history of the multimedia display system development, every multimedia display system has its own display purposes, so they have their own advantages and disadvantages The newer system seem to have the better features as they improve their previous disadvantages, and especially they always developed based on their own hardware technology In these multimedia display systems, they aim to the HD quality of the digital contents in wired network where is much more stable and reliable than wireless network Furthermore, the wired network have hardware resources stronger than a wireless network Actually, there are quite many paper and thesis about the wired digital multimedia system This thesis target is to approach with a wireless display distribution system of the digital multimedia achieving HD quality which based Linux platform

This thesis also mentions the performance and the security of DCDDS model during the displaying and distribution process of its digital contents In the other words, DCDDS is almost based on the Linux platform, so potentially leading to the diversity in its own display purposes This point is considered as a little contribution

of this thesis

1.3 Methodology

Sun-Tech IDS is a leading service audiovisual provider in Hong Kong which develops the high quality products for worldwide customers SUN-TECH IDS technologies based on the hardware are the multimedia display systems which support the diversity for the video formats in the wired network environment The

central management system supports to deliver the digital contents for audiences SUN-TECH IDS is used for displaying the digital contents in hotels, commercial buildings, shopping malls, stadiums, exhibition halls, etc

ASF/WMV/HTM Broadcasting Control Station

Audio Line

VGA/DVI Twisted Pair/Optical Fiber

VGA/DVI Audio Line

Twisted Pair/Optical Fiber

LCD

LED Screen

WAN/LAN

Figure 1 Multimedia information display system of SUN-TECH IDS

SUN-TECH IDS really achieves stable and reliable with the display systems in the wired network environment, and supports multiple formats However, the SUN-TECH IDS system also have the disadvantages Those disadvantages such as high cost for the new system deployment and difficulty for the integration into the existing IT system

This thesis target is to approach with a wireless display distribution system of the digital contents achieving the high definition (HD) quality which almost based Linux platform As Figure 2 shown, the proposed DCDDS model would be the better candidate than SUN-TECH IDS, because of the wireless communication between Terminal and Raspberry Pi for the lower cost with the new system deployment, or easier integration for the existing IT system The Linux platform for the performance and security of system in server zone and terminal zone, also supports multiple formats

HDMI WIFI

Figure 2 The proposed DCDDS in Viet Nam This thesis is also approached with the studying of performance of DCDDS in two main parts: server zone and terminal zone Apache 2.4 Web Server, RAID-based Storage System and Telestream FlipFactory in server zone, Terminal and Raspberry

Pi in terminal zone A qualitative assessment is performed among these nodes which would present pros and cons of each nodes And the next part is quantitative comparison via VMware virtualization platform the where open source system is used primarily in this thesis, it supports for setting up the virtualization scenarios, running the virtualization and generating result data The result contains raw data in the text files or the figures that should be processed to filter the necessary information which reflecting thesis’s targets Finally, the results help thesis answer the research question of itself

1.4 Thesis Scope and Limitation

1.4.1 Thesis Scope

Thesis defines two scopes: technology and management, and application

In technology and management, DCDDS concentrates on the operation of DCDDS

in the two main parts: server zone and terminal zone In the server zone, the server plays role as Apache 2.4 Web Server based on the Linux platform which the Event MPM would dedicate the threads for the new requests, RAID-based Storage System based on the command line Linux platform which ensures the security and data integrity as well as the performance in any given situation, Telestream FlipFactory is a standard for transcoding, delivering the digital contents automatically In the terminal zone, Terminal based on the graphic Linux platform which is shell script programmed for distributing the digital contents automatically and securely The Raspberry Pi is a microcontroller which has capable of a computer for receiving the encoded contents securely, then decoding and displaying the digital HD signals on HDTV with standard HDMI The evaluation provides the quantitative results for the performance in every zone

In application, Infosys + software licensed by Kantar Media company, is used to rank the spots on channel HTV7, channel HTV9 of HTV The sample includes

1575 local users in HCM City, and 6017 global users on hour frame from 06:00

to 22:30 hours, dated 06.07.2014 This rating result compares with channel HTV7, channel HTV9 and channel THVL1 (Vinh Long) According to the spot rating points, it is indicated that the attention to the content spots on the major channels and the entertainment channels of users is very high Besides that, Infosys+ is used to filter the kind of spot which have high points, then classify, identify and evaluate the potent of companies sponsoring the advertising spots These help writer identify the trends of users The profit problem is born in here

Figure 3 Technical Card in Infosys+

Figure 4.The advertising spots of HTV7, HTV9 and THVL1

This tool is used to identify, evaluate, and classify the potential companies which sponsor the advertising spots of DCDDS The key contribution of this thesis is the studying and the developing a system which applys technologies and brings the economic value for a new model of the wireless multimedia system in Viet Nam

1.4.2 Thesis Limitation

In Vietnam, the DCDDS could be considered as a new applied research topic, so it

is not feasible for a thesis to cover all technical aspects Moreover, this thesis is a partial fulfillment of the requirements of MITM program, and it requires to be completed within four months Due to the time limitation, this thesis is also limited

in scope and intensive study

As discussion in the introduction session, although the DCDDS is almost based on the Linux platform for the television and advertising field, potentially leading to the diversity in its own display purposes But DCDDS totally dependents on the technology software of FlipFactory achieving load balancing and enhancement of speeds job execution, and the technology hardware of Raspberry Pi with VideoCore

IV GPU which is great to decode and display the digital contents with the HD quality

Virtualization Limitation

In this virtualization scope, Telestream FlipFactory with FactoryArray achieving load balancing and enhancement of speeds job execution, which only reflects the minimum configuration to enable a FactoryArray operates, whereas how powerful FactoryArray is in fact RAID-based Storage System achieving the security and data integrity as well as the performance is exploited in any given situation, which not reflect network bandwidth and the performance when a huge of big files is transferred really

1.5 Structure of Thesis

This thesis includes five chapters and appendix

Chapter I: this chapter introduces overview of thesis, problem statements and how

to deal with these problems This chapter also covers thesis scope and limitation, and thesis structure

Chapter II: this chapter covers the introduction of multimedia distribution

networks and the proposed issues for the multimedia networks

Chapter III: the first part of this chapter compares advantages and disadvantages

between the existing techniques and DCDDS with technology of Apache 2.4 Web Server, RAID-based storage technology, Telestream FlipFactory technology, Raspberry Pi technology This part elaborates the major drawbacks of each technology, and why DCDDS chooses these technologies Then the last part covers the proposed DCDDS operation in details

Chapter IV: this chapter is for the virtualization implementing of Apache 2.4 Web

Server, RAID-based Storage System, Telestream FlipFactory, Terminal based on the VMware virtualization platform, and deploying Raspberry Pi The next part of this chapter elaborates on the quantitative comparison and discussion of the virtualization results

Chapter V: conclusion and future works of this thesis

Appendix: this part consists of the installation and the issues of QoS and the

security for DCDDS based on Linux platform

CHAPTER 2 LITERATURE REVIEW

This chapter covers the introduction of multimedia network and the proposed issues for the multimedia network These issues are also the related issues for the proposed DCDDS model in server zone with video compression technologies for the supported mp4 format, and for the proposed DCDDS model in terminal zone with aspects of wireless multimedia networks

2.1 Introduction of Multimedia Network

There's not a common multimedia network, but based on its purposes, there would have three basic kinds such as those networks deliver multimedia contents including video delivery, those networks deliver videos on demand, pre-recorded videos or pre-encoded videos, and those networks deliver live videos So the multimedia networks are used to fastly and effectively deliver a diversity of contents to various terminals, whether terminals are web browsers, smart devices, set-top boxes, HDTV, etc The targets of multimedia network are to supply contents for users with high availability and high performance

2.2 The proposed Issues for Future Multimedia Networks

In the third international workshop on Future Multimedia Networking (FMN) 2010 [1] with the committee members are professors from the leading universities in the world, the researchers and the practitioners have an opportunity to share the experiences and discuss the research results in the area of wireless multimedia network

The multimedia traffic and services would raise, and there would have important contributors to Internet traffic in the future The multimedia distribution services and the delivery of multimedia services would be diverse and efficient Therefore, this part would address the aspects of wireless multimedia network including

wireless multimedia communication, video compression technologies and the

security for wireless multimedia network

2.3 Wireless Multimedia Communication

For audiovisual communication requirements [2], IEEE 802.11 provides a standard which is more popular for communication in homes, offices, and public areas such

as airports, museums, shopping malls, etc IEEE 802.11 specifications are focused

on the Medium Access Control (MAC) layer and Physical Layer (PHY) of the ISO protocol stack They operate at 2.4 Mbps or 5 Mbps The current widely deployed 802.11n is a revision which is improved by adding Multiple Input and Multiple Output Antennas (MIMO) The IEEE 802.11n operates on 5 GHz bands MIMO antenna system permits up to 4 MIMO antenna pairings, with 1 MIMO is one pairing of transmit and receive The IEEE 802.11n OFDM has a maximum data rate about 65 Mbps, multiplied by 4 MIMO antennas So the data rate is about 260 Mbps This is a big improvement as it is compared to 54 Mbps for the previous IEEE 802.11

802.4 Token Bus

802.5 Token Ring

802.11 Wireless LAN

2.4 GHz

FHSS

2.4 GHz DSSS

2.4 GHz Infrared

2.4 GHz DSSS

2.4 GHz OFDM

5 GHz OFDM

Data rates: 1.2 Mbps 5.5, 11

Mbps

6-54 Mbps

6-54 Mbps

802.11n

Figure 6 The model of IEEE 802.11 at physical layer

2.4 Security for Wireless Multimedia Network

2.4.1 For Network Infrastructure

Wired Equivalent Privacy (WEP) is not safe so far This is the first security standard for wireless LANs by IEEE, has vulnerability which attackers would exploit In

2003, the Wifi Association has created a security standard named Wifi Protected Access (WAP) The first version of WPA with the better encryption mechanism has caused trouble for attackers, also has vulnerability which attackers would exploit In

2004, the second version of WPA (WPA2) was permitted, the security has developed with the best IEEE 802.11i security standard There are two modes for WPA/WPA2 consist of the preshared key mode and the enterprise mode

2.4.1.1 The Preshared Key Mode

This mode is only suitable for individual networks The password would created and saved on the wireless network mediums such as routers and access points for the first time Then the password would be entered by the user as they connect to these networks The network security would be not guaranteed The password would stored on the personal computers for all users The password would be changed on all network mediums manually The users are able to connect to these networks, they would be able to recover the encrypted passwords on these personal computers

2.4.1.2 The Enterprise Mode

This mode provides the highest security with wireless networks for enterprise environment Firstly, the users would assigned their credentials which the secret

information would be entered as the users connect to these networks These information would be changed by network administrator at any time Then the passwords would created safely, and they would assigned per one user session as the users enter their login credentials This would prevent someone tries to recover the encrypted passwords from the personal computers in these wireless networks

2.4.2 Introduction of Video Compression Technologies

The video applications become an important requirement for wireless communication However, the multimedia communication over wireless network is

a big problem, primarily because of elements such as limitations of bandwidth, characteristics of the transmission method, etc

The increase numbers of users and services would be handled and extended for wired networks, but the system capacity of wireless networks would not be extended easily Because of the big amount of data for video communication, the compression technologies would preferable for the multimedia transmission in wireless networks

As Table 1 shown, this part introduces about MPEG-1, MPEG-2, MPEG-4, H.264, and H.265 [3] The H.120 standard was created firstly After that, the VCEG (Video Coding Expert Group) has created the H.120 standard MPEG (Moving Pictures Experts Group) has created codecs for broadcasting field Nowadays, MPEG for broadcasting field, and VCEG for telecommunication field In 1988, VCEG has made H.261 standard with bit rate of 2 Mbps for ISDN and video conferencing MPEG has also permitted the H.261 standard with the first name is MPEG-1 limited 1.5 Mbps MPEG has defined sub-divisions, named parts Part 1 is used for the file format Part 2 is used for video, and Part 3 is used for audio So MPEG-1 Part 3 Layer III is called MP3, for example

In 1999, MPEG-2 was created Part 1 for program stream and transport stream MPEG-2 has been accepted by most broadcasting television, and its codec is used widely for broadcasting field so far

Table 1 MPEG and VCEG

Year MPEG Part Layer/Profile/Type Usage VCEG

1999 MPEG-2 Broadcast, Distribution, DVD

Program Stream Transport Stream

Layer 1 Layer II

H.264

MPEG-4 is the fullest standard From MPEG-4 Part 2, is also called H.263, the concept of profiles was defined MPEG-4 was divided into 30 parts Part 10 and Part 14 are discussed most H.264 or MPEG-4 Part 10/AVC defines the AVC

(Advanced Video Coding) format H.264 codec is widely used, because of the small file size after compressed MP4 or MPEG-4 Part 14 is a container format for MPEG-4 codecs (in *.mp4 format) H.264 supports up to 60 frame per second (fps)

In 2013, HEVC (High Efficiency Video Coding) or H.265 was developed, also called MPEG-H The H.265 supports up to 300 frame per second (fps), and the file size is the half of H.264 for the better audiovisual quality

2.4.3 For The Video Encryption Algorithms

The important information would not be accessed or would only be somewhat visible for the normal users That is the reason the security for video data becomes essential in the context of wireless multimedia communication The encryption algorithms (DES, RSA, AES, etc.) are used for text data and binary data would not meet the huge size of video data (MPEG 1, MPEG 2, MPEG 4, H.263/ H.264/AVC) In this part, the video encryption algorithms [4] are classified and showed The security is the most important condition, this is the first condition The encryption algorithms would not affect compression ratio for the video data, this is the second condition The cost for breaking the encryption algorithms is always higher than the cost for the videos’s copyrights The delay time is acceptable for the real time multimedia applications

Video data is usually transmitted in heterogeneous network The different encryption algorithms are used and classfied in four categories

2.4.3.1 Fully Layered Encryption

In this category, firsly the whole video data is compressed Secondly, the bitstream

is would be encrypted by a standard cipher DES or AES

2.4.3.1.1 Nạve Technique

This is the typical encryption algorithm of Fully Layered Encryption, this algorithm would use the encryption algorithm of AES or DES to encrypt every byte for MPEG stream There is no any decryption algorithms can break DES or AES This algorithm would encrypt video data after compression, the time delay and overhead would not be acceptable for the real time multimedia applications

2.4.3.2 Permutation based Encryption

In this category, the encryption algorithms would use the permutations to create the audiovisual degradation These algorithms achieve the faster speed due to the support of hardware

2.4.3.2.1 Pure Permutation

The pure permutation algorithm would use the permutations to scramble bytes for a frame of MPEG stream The permutation table is showed out, all frames would be decrypted

2.4.3.2.2 Zig-Zag Permutation

This algorithm would use a permutation table to map the specific 8x8 block to 1x64 vector There are many methods to calculate a permutation table This algorithm consists of three steps: creating a table of 64 permutations, splitting procedure and appling the random permutation to the split block

The compression ratio decreases is the main disadvantage for this algorithm Because the random permutation would affect the coefficients of Discrete Cosine Transform (DCT) Non zero AC coefficients would gather in the upper left corner

of the block The secret key is used for this corner

2.4.3.2.3 Huffman Codeword Permutation

This is a video encryption algorithm which would combine with the video compression in one stage The Huffman codeword table would be the secret key And Huffman codeworks have the same length for the stable compression rate

2.4.3.2.4 Compression Logic based Random Permutation

This encryption algorithm would be used for an amount of permutation groups The random permutations would created for each permutation group to encrypt a video frame This algorithm is safer than Zig-Zag Permutation Algorithm for the DCT vulnerability

2.4.3.2.5 Correlation Preserving Permutation

This algorithm would encrypt the video data before video coding This algorithm depends on the sorting permutation of previous frame, then it determines a sorting

permutation of the next frame There would have no the secret key is generated The attackers would calculate the sorting permutation for a certain frame, but they would not know about that frame

2.4.3.3 Selective Encryption

In this category, the whole video data would not be compressed This encryption algorithm would encrypt a certain part of video data This encryption algorithm is to reduce overhead and ensure the necessary security level for video data There are many proposed encryption algorithms based on MPEG structure (I-frame, P-frame, B-frame) for this category would not be mentioned within this thesis

As Figure 7 shown, an example of general system for selective encryption system

Compressed/Encrypted

Bitsteam

Encrypted/Scrabmled Coefficients

Secret Key

Coefficients

Motion Vectors

Reconstructed Video Frames

As Table 2 shown, there is a comparision of video encryption algorithms for security level, speed, encryption ratio and video size

Table 2 Comparison of Video Encryption Algorithms Algorithm

Class Algorithm Name

Security Level Speed

Encryption Ratio Video Size Fully Layerd

Encryption Nạve Technique High Slow 100% No change

Permutation High Fast <100% No change Compression Logic

based Random Permutation

High Fast <100% No change

Correlation Preserving Permutation High Fast <100% No change

Selective

Encryption

Many proposed encryption algorithms based MPEG structure (I-frame, P-frame, B-frame)

High Fast <100% No change

Perceptual

Encryption

Many proposed encryption algorithms based on data elements (RGB color space, fix-length code (FLC)

High Fast <100% No change

CHAPTER 3 METHODOLOGY

As discussion in Chapter 1, 1.3 Methodology, this thesis target is to approach with

a wireless display distribution system of the digital multimedia achieving HD quality which almost based Linux platform The proposed DCDDS would be the better candidate, because of the wireless communication between Terminal and Raspberry Pi for lower cost with the new system deployment and easier integration for the existing IT system The Linux platform for the performance and the security

of system in server zone and terminal zone, also supports multiple formats The first part of this chapter compares advantages and disadvantages between the existing technologies and DCDDS with the technologies of Apache 2.4 web server, RAID-based storage technology, Telestream FlipFactory transcoding technology, Raspberry Pi technology The next part elaborates the major drawbacks of each technology, and answers why DCDDS chooses these technologies The last part is the proposed model and the operation for DCDDS in details

3.1 Comparison of Microcontroller Unit (MCU) Technology

3.1.1 Tiny210V2

3.1.1.1 What Is Tiny210V2 ?

Tiny210V2 [5] is a typical high performance controller board The board length is only 74mm with width is only 55mm and height is only 11mm (74mm x 55mm x 11mm It is designed based on the Samsung S5PV210 microcontroller

74 mm

Figure 8 Tiny210V2

3.1.1.2 The Structure of Tiny210V2

Tiny210V2 board is a high performance Cortex A8 core board It uses Samsung S5PV210 processor as the main processor which runs at 1GHz The Power VR SGX540 graphics core is integrated to support for 3D graphics, and the full HD video standards

Figure 9 The structure of Tiny210V2

3.1.1.3 Assessment

Tiny210V2 bases on Linux platform which highly supports to customize and develop the application softwares It would be a good development platform for the multimedia applications, but Tiny210V2 costs about 95$ with the integrated heavy LCD as Figure 8 shown

3.1.2 Raspberry Pi

3.1.2.1 What Is Raspberry Pi?

Raspberry Pi [6] is a powerful, tiny controller board The board length is only 85mm with width is only 56mm and height is only 17mm (85mm x 56mm x 17mm) Its size is only as a credit card It can do many things as a PC does, like HD video playing, spreadsheets, word processing It saves power consumption, and more useful for embedded projects Raspberry Pi has the attractive price only with

$25

Figure 10 Raspberry Pi model B+

3.1.2.2 The Structure of Raspberry Pi

The core is the Broadcom BCM2835 system on a chip (SoC)clocked at 700MHz This chip is integrated with the VideoCore IV GPU graphics core as Figure 11 shown This GPU has strong enough for playing the full HD video standards

Figure 11 The structure of Raspberry Pi model B+

3.1.2.3 Assessment

Raspberry Pi bases on Linux platform which highly supports to customize and develop the application softwares It would be a good development platform for the multimedia applications, and Raspberry Pi costs about 25$ with a moderate strength

3.1.3 Difference between ARM and x86

As well as Tiny210V2, the core of Raspberry Pi is the Broadcom BCM2835 system

on a chip (SoC), all components including CPU, GPU, audio, communication chip are integrated in one The SoC located below the black memory chip of 512MB in middle board This SoC is different with CPU in a normal PC in that it is built on the Instruction Set Architecture (ISA) [7], which is different with the x86 architecture The ARM architecture has a shortened form Reduced Instruction Set Computer (RISC), which uses low power consumption The x86 architecture has a shortened form Complex Instruction Set Computer (CISC), which operates with higher capacity and power consumption Because of the basic differences, so the x86 architecture and the ARM architecture has its own using purposes For example, ARM chip on Raspberry Pi operates with 5V, 700mA, 3.5W per hour while a laptop consumes at least a 20 watts per hour Raspberry Pi with moderate strength for tiny shape without fan, so the ARM architecture is suitable for mobile devices so far Due to difference of the Instruction Set Architecture, source codes are written for the x86 platform would not work on the ARM platform This would

be solved by converting source codes to achieve compatibility

2 Comparison of Apache Web Server Technologies

In the September 2014, a survey is received responses from 1,022,954,603 sites from NetCraft [8] Although the share market is decreasing, Apache still holds share

of the web server market with 50% of all hostnames Nginx has been achieving share with 14% of all hostnames as Figure 12 shown

Figure 12 Market share of the top million busiest sites What technologies makes Apache still dominate the market share of websites? Why does Nginx becomes an emerging candidates have equal market share as Microsoft? The two questions have the same answer That is the Event Multi Processing Module (MPM) This MPM was designed for more requests which are served simultaneously by dedicating the threads for new requests

3.1.4 Apache Web Server 2.0 and The Earlier Versions

Apache Web Server 2.0 and the earlier versions create processes and threads for connections The maximum number of processes would be configured exactly by the administrator for the best performance Those configurations depend on the available CPU and memory on the server There are many processes would exhaust CPU and memory When the limit number of processes is reached, Apache would reject additional connections That is the way Apache Web Server 2.0 and the earlier versions [9] have been configured to run the Prefork Multi Processing Module (MPM) and the Worker Multi Processing Module (MPM)

3.1.4.1 Apache MPM Prefork

In the Prefork Multi Processing Module(MPM) [10], a single parent process would used to control a pool of child processes, and responsible for observing the status of the child processes It creates new child processes when necessary, and terminates child processes when no longer necessary The child processes are responsible for

handling incoming requests from the clients Each child process is single threading and would handle just one request concurrently That is the way Apache works with the Prefork Multi Processing Module, there would have one process at a time The maximum number of requests would be processed concurrently is equal to the number of child processes as Figure 13 shown

Advantage:

The Prefork Multi Processing Module (MPM) works with multiple child processes, one thread per one child process, one child process handles one request, so it has separation and stability

Figure 13 The Prefork Multi Processing Module

3.1.4.2 Apache MPM Worker

In the Worker Multi Processing Module (MPM) [11], a single parent process would used to control a pool of child processes, and the child processes are responsible for handling incoming requests from the clients However, different with the Prefork Multi Processing Module (MPM), in the Worker Multi Processing Module (MPM), each child process is multithreading and would handle multiple requests concurrently That is the way Apache works with the Worker Multi Processing Module, each process handles multiple threads The maximum number of requests would be processed concurrently is the number of child processes multiplied by the number of threads per child process as Figure 14 shown

httpd process

thread thread thread

Disadvantage:

The Worker Multi Processing Module would not provide the same method as a process-based MPM If a single threading is delayed or lost control, the entire process will be terminated, so this process would affect all its threads

3.1.5 Apache Web Server 2.4 with Event MPM

The Event Multi Processing Module (MPM) [12] is designed for more requests which are served simultaneously by dedicating the threads for new requests The Event MPM has been created with stable in Apache 2.4 version The Event MPM has the same technique as the Worker MPM The most important difference is the Event MPM would dedicate the threads for the new requests, not for the entire HTTP connection This MPM would fix the problem of keeping alive for HTTP connections When a client finishes the request, this client would retain the connection opened, and send the extra requests by using the same socket This

technique would save overhead for creating HTTP connections Apache Web Server 2.0 and the earlier versions retain the entire HTTP connection for data is transmitted from the clients, this is the most disadvantage The Event MPM would dedicate the threads for handling the listening sockets, all sockets would be in a keep alive state, and would only wait for the data is transmitted from the clients That is the most important advantage of the Event MPM, the threads would only used for the requests and would free up instantly as the requests are handled by the parent process So the threads would be used for the new requests rapidly

3.2 Comparison of Transcoding Technology

There are many solutions for transcoding the digital contents for personal purposes and industry Especially, in television and advertisng field require to create the customize workflows which transcode the original multimedia contents into a diversity of formats, and deliver the encoded multimedia contents to destinations specified, including Storage Area Network (SAN), network folders, and FTP sites automatically There are three typical technologies such as Grass Valley with Canopus Procoder, Harmonic is the industry's leading service provider for encoding, stream processing and video networking solutions for high quality broadcast and multiscreen applications, and the last is Telestream FlipFactory which

is the powerful way to transcode and publish the digital contents automatically This part elaborates the major drawbacks of each technology, and answers why DCDDS chooses those technologies

3.2.1 Grass Vally with Canopus Procoder

Grass Valley with Canopus Procoder [13] is a specialized tool for converting the diversity of video formats It fully supports for editing frame rates, size, and resolution with multiple video codecs It is a great tool to create the digital contents for the video applications in internet

3.2.1.1 Assessment

Advantage: Canopus Procoder has enhanced the digital codec technologies, which makes the conversion from this video format into the other video formats quickly and easily

Disadvantage: with the feature for multi-user watch folders, the software creates the customize watch folders and publishes them in the network environment This feature allows to move or copy the encoded contents in the network environment But what we want is create the customize workflows which transcode the original multimedia contents into a diversity of formats, and deliver the encoded multimedia contents to destinations specified, including Storage Area Network (SAN), network folders, and FTP sites automatically

3.2.2 Harmonic with Promedia Carbon

Harmonic with ProMedia Carbon is a transcoding solution for the conversion of the multimedia contents to a diversity of formats ProMedia Carbon run as a stand alone application or as a transcoding farm, fully automated rendering farm under the control of the Harmonic file-based workflow system (WFS) [14] The WFS is an open and extensible task based engine for managing file-based workflows The nature of WFS allows for automatic process of transcoding tasks, managing job distribution, load balancing, delivering content, status monitoring