Blab Tech
NetBSD Powers Diverse Research Projects from Satellite Networks to Spaceflight
NetBSD’s reputation as a lightweight, highly portable BSD‑derived operating system has turned it into a secret weapon for researchers pushing the boundaries of networking, spaceflight, storage, and security. From NASA’s satellite‑link experiments to a record‑breaking Internet2 data transfer, the system’s clean code base and broad hardware support have made it the go‑to platform for scientists and engineers worldwide.
At NASA’s Lewis Research Center, NetBSD dominates the study of TCP performance over satellite links. The Satellite Networks and Architectures Branch is deploying extensions such as selective‑acknowledgement–based retransmission (FACK TCP) and larger initial windows, running tests in both satellite and terrestrial environments. These efforts run in close coordination with the Internet Engineering Task Force’s TCP Over Satellite and TCP Implementations working groups.
In June 1999, the KAME project—an alliance of six Japanese organizations—merged an IPv6/IPsec stack into NetBSD. The code, originally crafted for the WIDE Hydrangea IPv6 implementation, was integrated into the main NetBSD tree, yielding a free, BSD‑licensed reference implementation of IPv6 and IPsec.
Germany’s NEC Europe Network Laboratories in Heidelberg use NetBSD to prototype IPv6 routers and conduct real‑world testing. The labs focus on multimedia and mobility protocols, and NetBSD’s support for legacy architectures such as SPARC makes it an ideal fit for experiments on older hardware.
The Institute for Media Communication (IMK) in Germany has built a suite of applications atop NetBSD. Their Network Access Control project embeds an access‑control language directly into the IP stack, enabling administrators to restrict network use by user, time, or remote login status. IMK also operates a tunneling server that transparently forwards insecure protocols through secure tunnels, a spam‑protection tool, and a sendmail replacement capable of handling many simultaneous SMTP connections with a single process.
NetBSD’s networking stack has long served as a testbed for congestion‑control research. The TCP Vegas algorithm—originating at the University of Arizona—was ported to NetBSD 1.0 by Peter B. Danzig’s group at USC, and live experiments have evaluated its delay‑based congestion avoidance in controlled settings.
In 2004, researchers at the Swedish University Network (SUNET) set an Internet2 Land Speed record using two Dell 2650 machines running NetBSD 2.0 beta. They transferred 840 GB of data in under 30 minutes over a single IPv4 TCP stream between Luleå University of Technology and a Sprint PoP in San Jose, achieving 69.073 petabit‑meters per second. The record was credited to NetBSD’s scalable TCP code.
The European SCAMPI project, spanning two and a half years, employs NetBSD to develop a scalable monitoring platform for high‑speed Internet traffic. The project builds a 10 Gbps network adapter and explores 100 Gbps monitoring challenges, with the Liberouter COMBO6 card serving as the hardware foundation.
Other initiatives have leveraged NetBSD’s kernel to improve performance and functionality. The UVM virtual‑memory system replaced the Mach‑based 4.4BSD VM, boosting performance for forking and pageout while preserving Unix compatibility. The Space Acceleration Measurement System II (SAMS‑II) runs on PC104 boxes with NetBSD 1.2.1 or 1.3.2 to measure microgravity aboard the International Space Station.
The Massively‑parallel And Real‑time Storage (MARS) project extended NetBSD with a multimedia buffer (mmbuf), fair‑queueing SCSI drivers, and a real‑time scheduling mechanism to support high‑throughput multimedia‑on‑demand servers.
Additional kernel enhancements include a constant‑time callout and timer implementation from Washington University, and a study of group awareness in distributed open‑source projects that examined NetBSD, Apache, and Subversion.
Finally, Antti Kantee’s work on application‑driven checkpointing for hot‑spare high availability was presented at EuroBSDCon 2004, showcasing NetBSD’s suitability for fault‑tolerant systems.
Across these projects, NetBSD’s portability, clean architecture, and active community have enabled researchers to prototype and validate new networking protocols, spaceflight systems, and storage solutions on a single, reliable platform.
At NASA’s Lewis Research Center, NetBSD dominates the study of TCP performance over satellite links. The Satellite Networks and Architectures Branch is deploying extensions such as selective‑acknowledgement–based retransmission (FACK TCP) and larger initial windows, running tests in both satellite and terrestrial environments. These efforts run in close coordination with the Internet Engineering Task Force’s TCP Over Satellite and TCP Implementations working groups.
In June 1999, the KAME project—an alliance of six Japanese organizations—merged an IPv6/IPsec stack into NetBSD. The code, originally crafted for the WIDE Hydrangea IPv6 implementation, was integrated into the main NetBSD tree, yielding a free, BSD‑licensed reference implementation of IPv6 and IPsec.
Germany’s NEC Europe Network Laboratories in Heidelberg use NetBSD to prototype IPv6 routers and conduct real‑world testing. The labs focus on multimedia and mobility protocols, and NetBSD’s support for legacy architectures such as SPARC makes it an ideal fit for experiments on older hardware.
The Institute for Media Communication (IMK) in Germany has built a suite of applications atop NetBSD. Their Network Access Control project embeds an access‑control language directly into the IP stack, enabling administrators to restrict network use by user, time, or remote login status. IMK also operates a tunneling server that transparently forwards insecure protocols through secure tunnels, a spam‑protection tool, and a sendmail replacement capable of handling many simultaneous SMTP connections with a single process.
NetBSD’s networking stack has long served as a testbed for congestion‑control research. The TCP Vegas algorithm—originating at the University of Arizona—was ported to NetBSD 1.0 by Peter B. Danzig’s group at USC, and live experiments have evaluated its delay‑based congestion avoidance in controlled settings.
In 2004, researchers at the Swedish University Network (SUNET) set an Internet2 Land Speed record using two Dell 2650 machines running NetBSD 2.0 beta. They transferred 840 GB of data in under 30 minutes over a single IPv4 TCP stream between Luleå University of Technology and a Sprint PoP in San Jose, achieving 69.073 petabit‑meters per second. The record was credited to NetBSD’s scalable TCP code.
The European SCAMPI project, spanning two and a half years, employs NetBSD to develop a scalable monitoring platform for high‑speed Internet traffic. The project builds a 10 Gbps network adapter and explores 100 Gbps monitoring challenges, with the Liberouter COMBO6 card serving as the hardware foundation.
Other initiatives have leveraged NetBSD’s kernel to improve performance and functionality. The UVM virtual‑memory system replaced the Mach‑based 4.4BSD VM, boosting performance for forking and pageout while preserving Unix compatibility. The Space Acceleration Measurement System II (SAMS‑II) runs on PC104 boxes with NetBSD 1.2.1 or 1.3.2 to measure microgravity aboard the International Space Station.
The Massively‑parallel And Real‑time Storage (MARS) project extended NetBSD with a multimedia buffer (mmbuf), fair‑queueing SCSI drivers, and a real‑time scheduling mechanism to support high‑throughput multimedia‑on‑demand servers.
Additional kernel enhancements include a constant‑time callout and timer implementation from Washington University, and a study of group awareness in distributed open‑source projects that examined NetBSD, Apache, and Subversion.
Finally, Antti Kantee’s work on application‑driven checkpointing for hot‑spare high availability was presented at EuroBSDCon 2004, showcasing NetBSD’s suitability for fault‑tolerant systems.
Across these projects, NetBSD’s portability, clean architecture, and active community have enabled researchers to prototype and validate new networking protocols, spaceflight systems, and storage solutions on a single, reliable platform.
