隨著網路的發達、科技的進步，雲端科技也越來越受到人們的關注。雲端運算是一種電腦運算的概念，它本質上就是代表分散式運算的概念。而分散式運算簡單的說，就是讓一些不同的電腦同時去幫你做事情、進行運算，所以不管你擁有幾台電腦，可以讓他們可以互相溝通，一起同時幫你做事情，這就是分散式運算。這也就是說隨著許多人開始使用雲端提供的服務，雲端內部的虛擬機器或實體機器們就要彼此溝通並且一起作運算，但這伴隨著是雲端內部的網路流量暴增。有鑑於此，本論文將致力於減輕雲端內部流量的壓力。
本論文將分析雲端內部封包欄位的使用情形，之後依據統計結果來為雲端內部網路的封包作客製化，把雲端內部不需要的header 欄位拿掉以此壓縮封包的大小。也因為客製化的封包其他機器會看不懂，我們修改了Linux 核心網路部份，使其具有發送以及接收客製化封包的能力，並且藉由控制Open vSwitch 來轉送封包。但這些都僅限於雲端內部流通的網路封包，也就是產生於雲端內部並送往雲端內部其他機器的封包形式。實驗結果顯示，壓縮率為1.9，空間節省率約為47.4%。簡而言之，此論文目的在於為雲端內部網路減少非必要的封包欄位，使其通訊量降低。

The term ”cloud” has become more and more popular. Due to its’popularity, more and more people are talking about the issue of cloud computing. With the rapid development of cloud computing, there are many new services and technology arise. But it goes along with the explosion of internal network traffic in the cloud. Because of this consequence, in this thesis we aim to alleviate the pressure of the internal traffic of cloud.
In this thesis, we are going to analyze the usage of fields of internal packet of cloud, and customize the packets based on the statistical results. We remove unneeded fields of header in order to compress the packet size. Because other machines can not read customized packets, we modify the network portion of the Linux kernel, and make it have capability to send and receive customized packets. In addition, we control the Open vSwitch to forward packets, too. But these are limited to the cloud internal network packets, which are produced in the cloud and sent to other machines within the cloud. The result of experiment shows that the compression ratio is 1.9, space-saving rate is about 47.4%. In brief, the goal of this thesis is to reduce non-essential fields of internal packets of cloud and lower down the amount of network communication.

Table of Contents i
List of Figures iii
List of Tables v
1 Introduction 1
2 Background 3
2.1 Cloud computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Network packet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Cloud network traffic . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.1 Cloud external traffic . . . . . . . . . . . . . . . . . . . . . . . 6
2.3.2 Cloud internal traffic . . . . . . . . . . . . . . . . . . . . . . . 6
2.4 Linux kernel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.5 OpenStack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.6 Software-defined networking . . . . . . . . . . . . . . . . . . . . . . . 8
2.7 Open vSwitch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3 Related work 10
3.1 Solutions for improving the network traffic . . . . . . . . . . . . . . . 10
3.1.1 Traffic Redundancy Elimination (TRE) . . . . . . . . . . . . . 10
3.1.2 Packet compression . . . . . . . . . . . . . . . . . . . . . . . . 11
3.1.3 Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4 Design Architecture 14
4.1 The building of cloud environment with SDN . . . . . . . . . . . . . 14
4.2 Customized packet . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.2.1 Packet analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.2.2 How to customize packet header . . . . . . . . . . . . . . . . 19
4.3 Customized Linux kernel for virtual machine . . . . . . . . . . . . . . 21
4.3.1 The sender algorithm . . . . . . . . . . . . . . . . . . . . . . . 21
4.3.2 The receiver algorithm . . . . . . . . . . . . . . . . . . . . . . 22
4.4 New function for virtual switch . . . . . . . . . . . . . . . . . . . . . 23
4.5 Limitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5 Implementation 27
5.1 Building Environment . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.1.1 OpenStack . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.1.2 Open vSwitch . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.2 Modifying Linux kernel source code . . . . . . . . . . . . . . . . . . . 29
5.2.1 Sending packet . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.2.2 Receiving packet . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.2.3 Recompile Linux kernel from our modified kernel . . . . . . . 31
5.3 Modifying Open vSwitch source code . . . . . . . . . . . . . . . . . . 31
6 Analysis 38
7 Conclusion 40
7.1 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
7.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

[1] Albert Greenberg, James Hamilton, David A Maltz, and Parveen Patel. The cost of a cloud: research problems in data center networks. ACM SIGCOMM Computer Communication Review, 39(1):68–73, 2008.
[2] Talking about new technological knowledge ”cloud computing”@nhu, 2010. http://www.nhu.edu.tw/~society/e-j/86/13.htm.
[3] Cloud computing and applied life, 2011. http://www.dgbas.gov.tw/public/Data/11281538171.pdf.
[4] Network packet@wiki, 2013. http://en.wikipedia.org/wiki/Network_packet.
[5] Ipv4@wiki, 2013. https://en.wikipedia.org/wiki/IPv4.
[6] Linux kernel@wiki, 2013. http://en.wikipedia.org/wiki/Linux_kernel.
[7] Linux kernel diagram@makelinux, 2007. http://www.makelinux.net/kernel/diagram.
[8] Openstack@wiki. http://en.wikipedia.org/wiki/OpenStack.
[9] Software-defined networking@wiki, 2013. http://en.wikipedia.org/wiki/Software-defined_networking.
[10] What is open vswitch?@open vswitch. http://openvswitch.org/.
[11] Neil T Spring and David Wetherall. A protocol-independent technique for eliminating redundant network traffic. In ACM SIGCOMM Computer Communication Review, volume 30, pages 87–95. ACM, 2000.
[12] Ashok Anand, Vyas Sekar, and Aditya Akella. Smartre: an architecture for coordinated network-wide redundancy elimination. In ACM SIGCOMM Computer Communication Review, volume 39, pages 87–98. ACM, 2009.
[13] C Tye and G Fairhurst. A review of ip packet compression techniques. In Proceedings of PostGraduate Networking Conference, Liverpool. Citeseer, 2003.
[14] Takahiro Miyamoto, Michiaki Hayashi, and Hideaki Tanaka. Customizing network functions for high performance cloud computing. In Network Computing and Applications, 2009. NCA 2009. Eighth IEEE International Symposium on, pages 130–133. IEEE, 2009.
[15] Mathur Mohit. A comprehensive solution to cloud traffic tribulations.
[16] Ethernet frame@wiki, 2013. http://www.nhu.edu.tw/~society/e-j/86/13.htm.
[17] Basic knowledge of network. http://linux.vbird.org/linux_server/0110network_basic.php.
[18] Christian Benvenuti. Understanding Linux network internals. O’Reilly, 2009.
[19] Analysis of open vswitch source code, 2001. http://security.riit.tsinghua.edu.cn/~bhyang/publication/ovs_code.pdf.