Talk:Hidden node problem

From Wikipedia, the free encyclopedia

Section "Background": What are "adaptive acknowledgements"? NotInventedHere 12:47, 27 May 2006 (UTC)

Section "Use Omnidirectional Antennas:" The advice to use omnidirectional antennas is not good in all cases. Directional antennas by nature reduce the interference potential of transmitters and increase selectivity of receivers. If the hidden nodes are not sending traffic to the same receiver (requires a minimum four-node network example), directional antennas can be a key component of frequency reuse inside a network. Also, in the case where an access point IS the recipient of two hidden nodes' data streams, two directional antennas in spatial diversity mode will greatly increase the odds that at least one of the two colliding transmissions will be correctly received (since the receiver will choose one antenna and recieve the transmission on that antenna without interference, requiring the retransmission of just one packet from the hidden node that "lost." 63.164.202.130 13:18, 13 April 2007 (UTC)

[edit] I do not understand the problem

After reading this article, I still do not understand why there is a problem when the nodes can't see each other - the nodes are communicating with the hub, not directly with each other, aren't they? Even if node A wants to send data to node B, that data goes (as far as I have understood computer networks) through the hub, so why is there a problem when A and B can't see each other?

The article should be edited to clarify this, by someone who knows the answer. Cybotoro 13:42, 17 August 2007 (UTC)

Imagine the two nodes are TV transmitters, transmitting different pictures on the same frequency. The "TV signal" from node A isn't strong enough to be received or cause interference at node B, because of the distance, but because the AP is in the middle and in range of both, the two signals will interfere with each other there and prevent reception.
If the nodes were not hidden from each other, they would sense that they were both talking at the same time, and one would defer, but because they are far enough apart to be out of each other's range, they cannot detect this situation. Steved424 (talk) 08:25, 6 May 2008 (UTC)

[edit] Picture Perfect?

Logically, this picture is both wrong and makes no sense, because there is a case where there exists a union between A and B. Even-though there is a case where A union B exists, but only when A union B union C. To be correct, there must be no cases where A union B intersect. If the A and B circles did not overlap at all, that would be better.
Image:A_union_c_b_union_c.JPG

there is no case where A is union with B Kassorlae (talk) 05:09, 25 January 2008 (UTC)

The point is that the actual nodes (the text labels) are not within the ranges of each other's signals (the circles), so cannot detect each other, though there exist points where both signals can be detected - such as the AP. It's not a Venn diagram :) -- Steved424 (talk) 08:41, 6 May 2008 (UTC)

[edit] Please explain

I agree with Cybotoro - who cares if the nodes (clients) cannot see each other? They aren't talking to each other, they are talking to the AP. And the AP has a buffer so that if he gets packets at the same time, they are queued, and not a problem. I don't see a problem here. What is the point of this article?

The AP can't get packets at the same time, at least not from one network (= one frequency), assuming it's using an omnidirectional antenna. The two signals would interfere with other, just as they would on old-fashioned, bus-style coax Ethernet. As the comment above mentions, you could use two highly directional antennas at the AP, but unless you used multiple MACs you'd have to to choose only one to listen to at a time. -- Steved424 (talk) 08:51, 6 May 2008 (UTC)