We are collaborating with several Japanese ISPs to understand the traffic behavior of broadband users.
The Gulliver Project is a distributed measurement project aiming to observe the global behavior of the Internet from all over the world, especially from peripheries in the Internet topology. The existing measurements are taken mostly at places close to the Internet core, and measurements from peripheries are very limited. However, views from peripheries such as developing regions are crucial to understanding the global behavior of the Internet.
There are many challenges to deploy measurement platforms in developing countries due to the lack of human resources and infrastructure. A measurement platform needs easy installation and remote maintenance, to be robust against instability and failures of power and connectivity, and should be secure even against physical burglaries. Another important challenge is how to remotely manage a large number of platforms.
We use a small diskless box as an active measurement probe. The probe runs a variant of NetBSD and has remote management framework built into the firmware. These probes are remotely managed by a set of servers. For example, the probes can be instructed to perform measurement and report the result to the server or to update its firmware through the GUI on the server.
More information is available from the gulliver project page.
This project is an ongoing effort to investigate the root name server performance from various locations of the Internet. We use simple probe programs to measure the responsetime of the root servers. We also measure the response time of the ccTLD servers to compare them with the root servers.
More information is available from the DNS measurement page.
One of the major hurdles limiting IPv6 adoption is the existence of poorly managed experimental IPv6 sites that negatively affect the perceived quality of the IPv6 Internet. To assist network operators in improving IPv6 networks, we are exploring methods to identify wide-area IPv6 network problems. Our approach makes use of parallel IPv4 and IPv6 connectivity to dual-stacked nodes.
We identify the existence of an IPv6 path problem by comparing IPv6 delay measurements to IPv4 delay measurements. Our test results indicate that the majority of IPv6 paths have delay characteristics comparable to those of IPv4, although a small number of paths exhibit a much larger delay with IPv6. Thus, we hope to improve the quality of the IPv6 Internet by identifying the worst set of problems.
Our methodology is simple. We create a list of systems with IPv6 and IPv4 addresses in actual use by monitoring DNS messages. We then measure delay to each address in order to select a few systems per site based on their IPv6:IPv4 response-time ratios. Finally, we run traceroute with Path MTU discovery to the selected systems and then visualize the results for comparative path analysis.
More information is available from the dualstack path analysis page.
Packet trace data is essential to study network dynamics, usage characteristics, and growth patterns, as well as various protocol design. Despite the increasing demands, it is difficult to obtain packet traces, especially at a backbone network. Packet traces could contain user privacy information so that they are usually available only under non-disclosure agreements.
As a chair of the MAWI (Measurement and Analysis on the WIDE Internet) working group of the WIDE Project, I have been working on building a public traffic trace repository. Our challenges include:
The WIDE traffic data repository contains packet traces from the WIDE backbone.
Visualization is essential to network traffic analysis. We are exploring techniques for visualizing network related information.
Queue management is an essential component in managing network traffic. A large number of queueing disciplines have been proposed to date in order to meet a wide range of requirements. However, FIFO queueing is the only queueing discipline used in traditional UNIX systems.
The ALTQ project is aimed at providing a flexible queueing platform in order to promote research and operational experience in the field. The ALTQ software release includes a queueing framework and a set of advanced queueing disciplines such as CBQ, HFSC, RED, WFQ BLUE, and RIO. ALTQ also includes traffic conditioning elements for diffserv. The ALTQ release for FreeBSD has been publicly available since March 1997, and used by many groups world wide. NetBSD and OpenBSD are also supported.
The H-FSC implementation in ALTQ is a joint work with Hui Zhang and his group members at CMU during my stay at CMU in July 1999.