afonso/core-extra
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

quick pass with some doc cleanup and reformatting to fix various issues

Browse source
This commit is contained in:
Blake Harnden 2020-04-09 14:34:52 -07:00
parent a5c412b594
commit ca039946a6
37 changed files with 1160 additions and 891 deletions
Download patch file Download diff file Expand all files Collapse all files

37
docs/performance.md
View file

@ -5,9 +5,10 @@
## Overview
The top question about the performance of CORE is often *how many nodes can it handle?* The answer depends on several factors:
The top question about the performance of CORE is often *how many nodes can it
handle?* The answer depends on several factors:
| Factor | How that factor might affect performance |
| 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. |
@ -16,16 +17,30 @@ The top question about the performance of CORE is often *how many nodes can it h
| 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.
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.
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?"*.**
> :warning: 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:
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.
* 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.