Tony McGregor

The NLANR Network Analysis Infrastructure

The National Laboratory for Applied Network Research is creating a network analysis infrastructure (NAI) to support network research and engineering of high performance research networks. The NAI includes a passive monitoring project, an active monitoring project, and the collection of network management and control data. Together these projects have deployed more than 120 monitors around the high-performance research networks in the United States. This article describes NAI and the projects using it. The article concludes with a discussion of the future plans for the infrastructure.

The RIPE NCC internet measurement data repository

This paper describes datasets that will shortly be made available to the research community through an Internet measurement data repository operated by the RIPE NCC. The datasets include measurements collected by RIPE NCC projects, packet trace sets recovered from the defunct NLANR website and datasets collected and currently hosted by other research institutions. This work aims to raise awareness of these datasets amongst researchers and to promote discussion about possible changes to the data collection processes to ensure that the measurements are relevant and useful to the community.

Design of a processor to support the teaching of computer systems

Teaching computer systems, including computer architecture, assembly language programming and operating system implementation, is a challenging occupation. At the University of Waikato this is made doubly true because we require all computer science and information systems students study this material at second year. The challenges of teaching difficult material to a wide range of students have driven its to find ways of making the material more accessible. The cornerstone of our strategy for delivering this material is the design and implementation of a custom CPU that meets the needs of teaching. This paper describes our motivation and these needs. We present the CPU and board design and describe the implementation of the CPU in an FPGA. The paper also includes some reflections on the use of a real CPU rather than a simulation environment. We conclude with a discussion of how the CPU can be used for advanced classes in computer architecture and a description of the current status of the project.

Quality in measurement: beyond the deployment barrier

Network measurement stands at an intersection in the development of the science. We explore possible futures for the area and propose some guidelines for the development of stronger measurement techniques. The paper concludes with a discussion of the work of the NLANR and WAND network measurement groups including the NLANR Network Analysis Infrastructure, AMP, PMA, analysis of Voice over IP traffic and separation of HTTP delays into queuing delay, network latency and server delay.

Using custom hardware and simulation to support computer systems teaching

Teaching computer systems, including computer architecture, assembly language programming and operating systems implementation, is a challenging occupation. At the University of Waikato we require all computer science and information systems students study this material at second year. The challenges of teaching difficult material to a wide range of students have driven us to find ways of making the material more accessible. The corner-stone of our strategy for delivering this material is the design and implementation of a custom CPU that meets the needs of teaching. In addition to the custom CPU we have developed several simulators that allow specific topics to be studied in detail. This paper describes our motivation for devloping a custom CPU and supporting tools. We present our CPU and the teaching board and describe the implementation of the CPU in an FPGA. The simulators that that have been developed to support the teaching of the course are then described. The paper concludes with a description of the current status of the project.

In-kernel passive measurement of the performance impact of hidden terminals in 802.11 wireless networks

The negative performance impact of the presence of hidden terminals in wireless networks has been well know for decades. Despite much research in the area, many deployed networks continue to suffer a performance penalty because of hidden terminals. Ad hoc wireless networks are particularly susceptible to hidden terminal collisions because there are fewer opportunities to plan the network in a way that avoids or reduces the number of hidden terminals. Measuring the presence of hidden terminals and the impact they are having on performance is difficult, especially in a network of many nodes. Without such measurements, the users and operators of wireless networks can not tell if performance problems are caused by hidden terminals or some other problem. We introduce new methodology that can detect the presence of hidden terminals and estimate the performance impact they are causing. The methodology requires no additional hardware and is suitable for wide scale deployment and long term operation. The approach is based on in-kernel instrumentation of the wireless network stack. The design, implementation, and testing of the approach are covered. Results from in-lab testing and the measurement of a live commercial 802.11 network are also presented, including a case study where performance was significantly improved.

Case Studies in Computer Network Measurement

A network lies at the heart of many modern computer systems including most distributed and parallel systems. A good understanding of the behaviour and performance of the network is, therefore, a prerequisite for the understanding the behaviour and performance of many computer systems. Understanding network behaviour and performance is challenging because networks are among mankinds largest and most complex creations and because most networks, including the Internet, are not designed with network measurement in mind. As a consequence, much network measurement relies on inference from data that does not directly describe the item of interest. Measurement also often requires a widely spread physical infrastructure, for example a network of monitoring machines. This is expensive both in terms of the hardware required and the resources needed to deploy and maintain the systems.

Automatic Application Object Migration in Sensor Networks

Object migration in wireless sensor networks has the potential to reduce energy consumption for a wireless sensor network mesh. Automated migration reduces the need for the programmer to perform manual static analysis to find an efficient layout solution. Instead, the system can self-optimise and adjust to changing conditions. This paper describes an automated, transparent object migration system for wireless sensor networks, implemented on a micro Java virtual machine. The migration system moves objects at runtime around the sensor mesh to reduce communication overheads. The movement of objects is transparent to the application developer. Automated transparent object migration is a core component of Hydra, a distributed operating system for wireless sensor networks that is currently under development. Performance of the system under a complex performance test scenario using a real-world dataset of seismic events is described. The results show that under both simple and complex conditions the migration technique can result in lower data traffic and consequently lower overall energy cost.