Chinaunix首页 | 论坛 | 博客
  • 博客访问: 405181
  • 博文数量: 59
  • 博客积分: 1510
  • 博客等级: 上尉
  • 技术积分: 857
  • 用 户 组: 普通用户
  • 注册时间: 2009-06-03 11:50
文章分类

全部博文(59)

文章存档

2011年(1)

2010年(36)

2009年(22)

我的朋友

分类: 系统运维

2009-06-24 10:43:23

发现一种新的模式组建无线Mesh网络,多跳之后,带宽并没有下降。之前做的无线mesh网络,最大的问题就是网络经过n跳之后,带宽基本呈1/n的趋势下降。这种模式经过实验,带宽在多跳之后并没有下降,在以后的项目实践过程中应该可以很好地利用起来。

原文:

AHDEMO Mode

Usage

Create the VAP device with:

wlanconfig ath0 create wlandev wifi0 wlanmode ahdemo

See tickets & for patches to correct the reported rate control Stats with ahdemo.

What is Ahdemo (Pseudo-IBSS) mode and why use it?

Ahdemo mode is handy for long shot point to point connections as you avoid the RTT and inefficiency of the control packets in BSS/IBSS mode.

The pseudo ad-hoc mode (Pseudo-IBSS mode, ‘lucent ad-hoc demo’ mode, PIM) is a pre-standards, pre-IBSS ad-hoc mode without beaconing ( no management frames at all: no beacons, no association, no probes, just data), which means that all the nodes in the mesh need to have channel assigned manually and won’t hop channels. This will give you slightly better throughput than regular ad-hoc mode because it doesn’t bother with the beaconing/probe response packets. But since beaconing is ignored you don’t know the capabilities of the other nodes (like WME).

With ahdemo mode you can get rid of beacon synchronization and IBSS merging (coalescing) problems in an easy way, especially in situations where you don’t actually need the “beaconing features” of IBSS, like for non-interoperable mesh solutions, for example.

Adhoc demo is not ibss. There is no ’standard’ for “adhoc demo” “pseudo-ibss” mode, implements it as “ibss minus beacons”

At least with old Lucent cards, pseudo adhoc mode did not use auth/assoc at all, it just started using IBSS (BSSID=00:00:00:00:00:00) immediately.

Features

1. The BSSID will be all zeros.

2. There is no distributed beaconing mechanism

3. As long as the cards on the same channel and within the radio range, all the cards can hear each other. There is no separate cells at the link layer.

4. A card in PIM will not send out any management frames (Probe request/response, beacon, authentication, deauthentication, ATIM, association request/response, reassociation request/response, disassociation)

5. What is the impact on the nearby 802.11 IBSS communication? Ubder testing 802.11 IBSS won in getting bandwidth. Running flood ping on both connections, 802.11 IBSS retry on corrupted packets gained, after a short time, all of bandwidth and pseudo IBSS connection had lost it. So you will need clear channel or very strong signal or send very small packets.

6. There are no Control frames (except the ACK frames).

7. The two cards in PIM set up the communication like classic cable Ethernet? Messages are broadcast to the current channel and the IP layer decide whether it should “receive” the data.

8. Packets are sent at rate you set. One needs to set the same rate on every card.

Testimonials

Mar 09 01:35:38 mattbrown It’s working really well btw.

Jun 15 18:33:16 foobarian i get 27Mbps with UDP when I use cwmin=5

Jun 15 18:34:20 ignuss got over 35mbit UDP and 33mbit TCP without turbo in 5ghz band , well, signal has been perfect

Jun 15 18:34:32 foobarian but when i use cwmin=1 i get 41Mbps

Jun 15 18:43:53 foobarian i think you would run ath0 cwmin 0 1

Ago 01 2007 20:10:48 soulhunter This is great!, improved my throughput from around 100kB/s to around 250 ~ 300kB/s. (10 miles link).

摘自:

阅读(1804) | 评论(0) | 转发(0) |
给主人留下些什么吧!~~