Random early detection gateways for congestion avoidance
IEEE/ACM Transactions on Networking (TON)
Evaluation of TCP Vegas: emulation and experiment
SIGCOMM '95 Proceedings of the conference on Applications, technologies, architectures, and protocols for computer communication
Dummynet: a simple approach to the evaluation of network protocols
ACM SIGCOMM Computer Communication Review
Mobile IP; Design Principles and Practices
Mobile IP; Design Principles and Practices
HAWAII: a domain-based approach for supporting mobility in wide-area wireless networks
IEEE/ACM Transactions on Networking (TON)
REAL: A Network Simulator
A framework for alternate queueing: towards traffic management by PC-UNIX based routers
ATEC '98 Proceedings of the annual conference on USENIX Annual Technical Conference
The design and implementation of the NCTUns 1.0 network simulator
Computer Networks: The International Journal of Computer and Telecommunications Networking
Virtual routers: a tool for networking research and education
ACM SIGCOMM Computer Communication Review
Time Jails: A Hybrid Approach to Scalable Network Emulation
Proceedings of the 22nd Workshop on Principles of Advanced and Distributed Simulation
Hi-index | 0.00 |
This paper proposes a new methodology for easily constructing extensible and high-fidelity TCP/IP network simulators. The methodology uses a kernel-reentering technique to reuse the existing real-life network protocol stacks, real application programs that generate traffic, and real utility programs that configure, monitor, or gather network statistics to the maximum extent. Only an event scheduler and some modifications to the kernel are needed to "glue" these existing components to collectively simulate a network.A simulator constructed this way has many advantages that a traditional network simulator cannot provide. First, reuse of real-life implementation in the simulator can generate more accurate results than a traditional simulator that abstracts a lot of away from the real implementation. Second, it can save much time and effort that would be needed if a high-fidelity simulator is developed from scratch. Third, because real application programs cannot distinguish a simulated network constructed by the simulator from a real one, all existing real-life and future application programs can directly run on any node in a simulated network.