44 lines
3.1 KiB
Markdown
44 lines
3.1 KiB
Markdown
# CORE Performance
|
|
|
|
## Overview
|
|
|
|
The top question about the performance of CORE is often *how many nodes can it
|
|
handle?* The answer depends on several factors:
|
|
|
|
| Factor | Performance Impact |
|
|
|--------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------|
|
|
| Hardware | the number and speed of processors in the computer, the available processor cache, RAM memory, and front-side bus speed may greatly affect overall performance. |
|
|
| Operating system version | distribution of Linux and the specific kernel versions used will affect overall performance. |
|
|
| Active processes | all nodes share the same CPU resources, so if one or more nodes is performing a CPU-intensive task, overall performance will suffer. |
|
|
| Network traffic | the more packets that are sent around the virtual network increases the amount of CPU usage. |
|
|
| GUI usage | widgets that run periodically, mobility scenarios, and other GUI interactions generally consume CPU cycles that may be needed for emulation. |
|
|
|
|
On a typical single-CPU Xeon 3.0GHz server machine with 2GB RAM running Linux,
|
|
we have found it reasonable to run 30-75 nodes running OSPFv2 and OSPFv3
|
|
routing. On this hardware CORE can instantiate 100 or more nodes, but at
|
|
that point it becomes critical as to what each of the nodes is doing.
|
|
|
|
Because this software is primarily a network emulator, the more appropriate
|
|
question is *how much network traffic can it handle?* On the same 3.0GHz
|
|
server described above, running Linux, about 300,000 packets-per-second can
|
|
be pushed through the system. The number of hops and the size of the packets
|
|
is less important. The limiting factor is the number of times that the
|
|
operating system needs to handle a packet. The 300,000 pps figure represents
|
|
the number of times the system as a whole needed to deal with a packet. As
|
|
more network hops are added, this increases the number of context switches
|
|
and decreases the throughput seen on the full length of the network path.
|
|
|
|
!!! note
|
|
|
|
The right question to be asking is *"how much traffic?"*, not
|
|
*"how many nodes?"*.
|
|
|
|
For a more detailed study of performance in CORE, refer to the following
|
|
publications:
|
|
|
|
* J\. Ahrenholz, T. Goff, and B. Adamson, Integration of the CORE and EMANE
|
|
Network Emulators, Proceedings of the IEEE Military Communications Conference 2011, November 2011.
|
|
* Ahrenholz, J., Comparison of CORE Network Emulation Platforms, Proceedings
|
|
of the IEEE Military Communications Conference 2010, pp. 864-869, November 2010.
|
|
* J\. Ahrenholz, C. Danilov, T. Henderson, and J.H. Kim, CORE: A real-time
|
|
network emulator, Proceedings of IEEE MILCOM Conference, 2008.
|