6

I come from a computer science background but physics is not presently one of my strengths. I've been reading some research regarding multi-hop routing of wireless (ad-hoc) networks such as this paper and this one.

One item that the authors seem to take for granted is the fact that a wireless signal will interfere with itself across a single node that is responsible for receiving and forwarding the signal (datagram) about 50% of the time. WHAT!? A 50% loss rate sounds like the most important thing to devote research toward mitigating! Yet, these authors aren't dunces. There must be a really good reason why self-interference is extremely difficult to work around.

So, for someone familiar with networking and computer science, but not as much with wireless networking and certainly not the physical layer's nuances, can anyone explain why packets being forwarded to and then from a mid-point node will interefere with itself (and perhaps even why it is hard/impossible to avoid it)?

BlackVegetable
  • 185
  • 5
  • 1
    FYI http://networkengineering.stackexchange.com/questions/10715/why-does-wlan-use-collision-avoidance-and-not-collison-detection/10727 – Mike Pennington Jan 20 '15 at 19:01

2 Answers2

6

Self-interference is an identity issue within mesh wireless networks, primarily with pre 802.11b wireless networks. The standards prior to 802.11n required only 3 non-overlapping channels of the 2.4GHz band. In large mesh deployments you would have multiple cells conflicting with one another and reporting back to the primary host controller.

802.11n improves this situation by opening up the additional channels in the 5 GHz band thus introducing many more new channels that can mitigate this self-interference challenge. However, while more channels helps, they do not eliminate the problem entirely. The use of smart antennas in 802.11n, particularly with beam forming, dramatically increases range but not necessarily in predictable or manageable patterns.

While the range increase is a benefit, it also brings with it increased interference at distance with other nodes. These issues suggest that the problem of interference is reduced with the advances of 802.11n; it will continue to be a key challenge in the design of large scale, high performance 802.11n Wi-Fi networks.

Citizen
  • 387
  • 5
  • 16
  • Thank you, this helps quite a bit. Looks like I'll need to look up the topics of: smart antennas, beam forming, the key differences between 802.11n vs. a/g. – BlackVegetable Jan 21 '15 at 16:16
  • 2
    No. You have 11n and 11ac confused. 11n is a different encoding method just like 11g was. It adds 40MHz channel support that makes 2.4GHz even more crowded -- it drops back to 20 if there's anything detected in the wide channel. 11n operates on both 2.4 and 5GHz. 11ac only operates at 5GHz. And it doesn't "open additional channels", it's the same channels 802.11a defined years ago. ('tho it can use wider channels) – Ricky Jan 22 '15 at 04:13
  • 802.11a uses the 5Ghz band... – cpt_fink May 19 '15 at 03:59
4

If you receive a packet on frequency X, you can't also transmit the same info on that same frequency. So you need to wait. that's why if you use the newer 802.11 protocol's who use more then one frequency you are going to have less interference(its not really interference its more like a busy phone line :P).

A side from Co-Channel interference(http://www.metageek.com/support/why-channels-1-6-and-11/) that's why they accept 50% lose of transmit bandwidth.. its not the rate that's slowing its the wireless bandwidth that's getting cut in about half :(

TheSec
  • 126
  • 3