To study the Internet in a quantitative fashion, we will need a variety of tools. Some tools are mathematical: linear algebra, probability, statistics, and graph theory.
3.1 Linear Algebra
please refer to Elementary Linear Algebra 9ed by Anton Horward, John Wiley and Sons.
2.2 Details of Internet Operation
2.2.1 End systems, links and routers
2.2.2 Autonomous Systems (AS)
2.2.3 Routing
2.3 Protocols
2.3.1 IP
2.3.2 TCP
2.3.3 UDP
2.3.4 Routing Protocols
2.3.5 ICMP
2.3.6 SNMP
2.3.7 IP Multicast
2.3.8 DNS
2.3.9 HTTP
2.3.10 P2P
2.4 Applications
To measure the Internet properly and to correctly interpret the results of Internet measurement, it is essential to understand the Internet’s architecture. The design of the Internet affects our ability to measure it and determines many of the questions that we seek to answer via measurement.
2.1 The Internet’s Architecture
2.1.1 The history of Internet
ARPANET -> ARPANET re-organization -> NSF-Net
2.1.2 The organization of the Internet
packet switching
protocols and layers
Headers
TCP
2.1.3 Design Principles of the Internet
decentralized design and operation
the IP hourglass
stateless switching
the end to end argument
Many different people and organizations have an interest in measuring the network or in obtaining Internet measurement.But broadly speaking, interest in Internet measurements arises for three kinds of reasons: commercial, social, and technical.
Commercially, the ability to sell a product or provide information about a product to a large number of people requires a variety of Internet measurements. For example, demographic information is important: what is the reach of the Internet ? How many individuals are connected in a given area ? What fraction of users have high-speed connectivity and how many are dependent on dial-up connectivity ? Where should network access point be placed ?
Effective commerce on the Internet also requires an understanding of the network’s performance properties. How long do web pages take to download from a vendor’s site ? What is the capacity of the path from a customer to a vendor’s web server ? How often do network problems prevent efficient transfer of information through the network ?
Socially, the need for statistics is clear: understanding the amount of network activity incolving various sites and protocols gives considerable insight into social issues. Government, scientists, and corporations may desire to have information about social implications of Internet user.
For example, popular web sites attract a considerable amount of traffic. Given the millions of sites from which information may be available, it is important to have characterizations of popularity and content. In another example, emerging protocols are indicators of popularity of a new application. This is shown by the sudden appearance of Napster and later KaZaa, which pointed to the imminent explosion in file sharing.
Finally, there are many technical reasons for Internet measurement. The design of network components and protocols is strongly driven by the nature of Internet workloads. For example, router designs depend strongly on the statistical properties of network traffic and packet size distribution. The statistical properties of web pages influence the performanc and design of web servers and browsers. Understanding the topology of the network helps identify the places where performance problems may arise and how applications may choose to adapt to the network. The popularity of new applications (such as network games) leads to efforts to build variants (new games). The popularity of new applications can also drive improvements to associated protocols – as in the case of the explosion of web traffic, which motivated the improvement of the basic HTTP/1.0 protocol to yield HTTP/1.1
Internet Measurement: Infrastructure, traffic, and applicantions by Mark Crovell and Balachander Krishnamurthy. ISBN 0-470-01461-X
The Internet as a Large-Scale Complex System, edited by Kihong Park and Walter Willinger, ISBN 0-19-515721-4
A Practical Guide to Heavy Tails: statistical techniques and applications, editors: Robert J Adler, Raisa E Feldman, Murad S Taqqu. ISBN 0-8176-3951-9
some papers from: ACM SIGCOMM, ACM SIGMETRICS, IEEE INFOCOM, IEEE MASCOTS, NANOG, etc.
Welcome to WordPress.com. This is your first post. Edit or delete it and start blogging!
