| 1 |
I don't know about tc, but for wireless overhead in general I've come to expect about 50% of the theoretical max. |
| 2 |
|
| 3 |
-- Matthew Thode |
| 4 |
|
| 5 |
On Aug 9, 2011, at 9:43 PM, Pandu Poluan wrote: |
| 6 |
|
| 7 |
> Hello list! |
| 8 |
> |
| 9 |
> I am planning to deploy a GLGF (Gentoo Linux Gateway/Firewall -- a |
| 10 |
> term I just made up). Now, I'm also supposed to put in some traffic |
| 11 |
> management in it. But there are some subtleties I'd like to clarify |
| 12 |
> first. |
| 13 |
> |
| 14 |
> 1. tc bandwidth measurement |
| 15 |
> |
| 16 |
> Does tc measure bandwidth based on the wire speed (e.g., datagram size |
| 17 |
> + L4 header + IP header + L2 header), or effective throughput speed |
| 18 |
> (datagram only)? |
| 19 |
> |
| 20 |
> 2. Wireless 802.11g overhead |
| 21 |
> |
| 22 |
> Related to the above. The link I'm having to traffic-managed is a |
| 23 |
> point-to-point link (a pair of dish antennae) using 802.11g. LOS |
| 24 |
> exists, Fresnel clear. Distance is approx. 200 meters. |
| 25 |
> |
| 26 |
> With this near-ideal situation, what's the expected overhead against |
| 27 |
> the 802.11g's theoretical max of 54 Mbps? |
| 28 |
> |
| 29 |
> Thank you beforehand for any input. |
| 30 |
> |
| 31 |
> Rgds, |
| 32 |
> |
| 33 |
> |
| 34 |
> |
| 35 |
> |
| 36 |
> -- |
| 37 |
> -- |
| 38 |
> Pandu E Poluan - IT Optimizer |
| 39 |
> My website: http://pandu.poluan.info/ |
| 40 |
> |
Archives