Research carried out using NetBSD
Several groups, organizations, and individuals have conducted research or performed demonstrations utilizing NetBSD as a vehicle. If you have such material that might be appropriate for this section, and would like to make it available here, please let us know!
- TCP in satellite networks (NASA Lewis)
- IPv6/IPsec (KAME)
- Institute for Media Communication (IMK) Value-added Solutions
- TCP Vegas with Live Experiments
- NEC Europe Network Laboratories
- NetBSD sets Internet2 Land Speed World Record
- A Scaleable Monitoring Platform for the Internet
- The UVM Virtual Memory System
- Space Acceleration Measurement System II (SAMS-II)
- Massively-parallel And Real-time Storage
- New Callout and Timer Facilities
- Group Awareness in Distributed Software Development
- Using Application-Driven Checkpointing Logic for Hot Spare High Availability
NASA Lewis Research Center - Satellite Networks and Architectures Branch use NetBSD almost exclusively in their investigation of TCP for use in satellite networks. They are currently examining several proposed modifications to the Transmission Control Protocol (TCP) protocol. Extensions are being tested in an attempt to improve satellite communication. However, the extensions are also being tested in terrestrial environments. Among the extensions being tested are restransmission mechanisms based on selective acknowledgements (e.g., FACK TCP) and TCP with larger initial windows. They are currently testing the mechanisms' performance benefits and their fairness to other traffic. They are also working with the Internet Engineering Task Force (IETF)'s TCP Over Satellite and TCP Implementations Working Groups.
KAME project . A research group for implementing IPv6, IPsec and other recent TCP/IP related technologies into BSD UNIX kernels, under BSD license. KAME stack got merged into NetBSD-current tree in June 1999.
The Institute for Media Communication (IMK) Value-added Solutions department . The current set of projects carries out on NetBSD:
Implementing an access control language which is integretated into the ip stack and can be configured by the system administrator. This gives the system manager fine grained control over who is allowed to access the network, at what time, when logged in from a remote machine or not, etc...
Certain users can be prohibited network access altogether ('nobody', 'bin', etc...). This reduces the risk of abuse on the machines.
Setting up a tunneling server which allows users to tunnel insecure protocols (pop3, smtp, etc...) through a secure tunnel transparently.
They have developed a tool that protects Internet users from receiving unsolicited bulk mail (aka spam) and are currently evaluating whether it is a feasible solution.
They are developing a sendmail-replacement which allows an arbitrary number of simultaneous SMTP connects with only one process, using asynchronous i/o. First beta versions are -substantially- faster than sendmail, qmail or any of the other free MTAs.
NetBSD TCP Vegas with Live Experiments . TCP Vegas is an extended slowstart TCP flow control from Lawrence Brakmo et al. at the University of Arizona. Prof. Peter B. Danzig and his group at usc.edu ported TCP Vegas to NetBSD 1.0.
NEC Europe Ltd. established the Network Laboratories in Heidelberg, Germany in 1997, as NEC's third research facility in Europe. Research and development functions are integrated into the same organization to shorten the time to market of cutting-edge network technologies. The laboratories place special emphasis on solutions meeting the needs of NEC Network's European customers.
The Heidelberg labs focus on software-oriented research and development for the next generation Internet. New communication architectures and protocols supporting multimedia and mobility over the Internet, together with intelligent Internet services, are the core of our work. A small market research team continuously analyses market trends and market requirements to assure that R&D activities address actual market needs.
NetBSD is used in the Network Laboratories for different task in the fields of research and IPv6 network operation among other free available BSD systems. The availability of NetBSD for different platforms makes it an ideal candidate for several activities, like:
- running IPv6 routers and services on older computer systems, e.g. Sparc systems
- testing research result in real life
The free available sources, the tight coupling of the sources to the well documented 4.4 BSD and the use of innovative techniques makes NetBSD an excellent software research platform.
Researchers of the Swedish University Network ( SUNET ) have beaten the Internet2 Land Speed Record using two Dell 2650 machines with single 2GHz CPUs running NetBSD 2.0 Beta. SUNET has transferred around 840 GigaBytes of data in less than 30 minutes, using a single IPv4 TCP stream, between a host at the Luleň University of Technology and a host connected to a Sprint PoP in San Jose, CA, USA. The achieved speed was 69.073 Petabit-meters/second. According to the research team, NetBSD was chosen "due to the scalability of the TCP code" .
More information about this record including the NetBSD configuration can be found at: http://proj.sunet.se/LSR2/
The website of the Internet2 Land Speed Record (I2-LSR) competition is located at: http://lsr.internet2.edu/
SCAMPI is a two-and-a-half-year European project to develop a scaleable monitoring platform for the Internet. It also aims to promote the use of monitoring tools for improving services and technology.
The project develops a network adapter, initially at 10 Gbps speeds, tailored to the needs of monitoring tools. This includes development of an open and extensible monitoring architecture to support a secure and programmeable shared monitoring infrastructure. It will also investigate the technical challenges of developing monitoring systems for 100 Gbps speeds and beyond.
The Liberouter COMBO6 Card is considered to be adopted as a base of the SCAMPI hardware monitoring adapter for speeds up to 10 Gbps.
UVM is a new virtual memory system specifically designed to provide the I/O and IPC systems with a range of flexible data movement mechanisms. Implemented in the NetBSD operating system, UVM completely replaces the Mach based 4.4BSD VM system. In addition to featuring flexible data movement mechanisms, UVM also improves virtual memory performance over BSD VM in traditional areas such as forking and pageout. UVM is implemented entirely within the framework of BSD and thus maintains all the features and standard parts of the traditional Unix environment that programmers have come to expect.
SAMS-II Project - Space Acceleration Measurement System II
NASA will be measuring the microgravity environment on the International Space Station using a distributed system, consisting of:
RTS (Remote Triaxial
Several PC104 ISA boxes running NetBSD 1.2.1 or 1.3.2.
ICU (Interium Control
An IBM ThinkPad 760XD as a temporary main box.
The first RTS was scheduled to be delivered in September 1998, for launch to the space station in Jan. 2000.
Cluster based architectures that employ inexpensive Personal Computers (PCs) interconnected by high speed commodity interconnect have been recognized as a cost-effective way of building high performance scalable Multimedia-On-Demand (MOD) storage servers. Typically, the PCs in these architectures run operating systems such as UNIX that have traditionally been optimized for interactive computing and lack fast disk-to-network data paths and support for guaranteed CPU and storage access. In this work we report design, implementation and performance measurements of innovative enhancements to 4.4 BSD UNIX carried out to rectify these limitations in the context of our Massively-parallel And Real-time Storage (MARS) project. We have proposed and implemented the following enhancements to a 4.4 BSD compliant public domain NetBSD UNIX operating system:
- A new kernel buffer management system called Multimedia M-buf (mmbuf) which shortens the data path from a storage device to network interface,
- fair queuing within the SCSI driver for equitable resource sharing between real-time and non-real-time streams, and
- integration of these new OS services with a CPU scheduling mechanism called Real Time Upcall and a software disk striping driver called Concatenated Disk (ccd).
Our experimental results demonstrate that these enhancements provide throughput improvements and QOS guarantees on the data path from the disk to network.
There is a paper describing the OS enhancements that appeared in IEEE Multimedia98, and there are several papers on the "Project MARS-Scalable, Web Based Multimedia-On-Demand Servers and Services" of which this work was a part:
Research carried out at the Washington University in St. Louis: Current BSD kernels (4.4BSD-Lite and derivatives) take time proportional to the number of outstanding timers (``callouts'') to set or cancel timers. This implementation takes constant time to start, stop, and maintain timers leading to a highly scalable design that can support thousands of outstanding timers without much overhead. Unlike the existing implementation, it is guaranteed to lock out interrupts only for a small, bounded amount of time.
Open-source software development projects are almost always collaborative and distributed. Despite the difficulties imposed by distance, these projects have managed to produce large, complex, and successful systems. However, there is still little known about how open-source teams manage their collaboration. In this paper we look at one aspect of this issue: how distributed developers maintain group awareness. This paper aims at analyzing various Open Source projects (NetBSD, Apache, Subversion) and analyzing group structure and communication.