GroupStudy.com GroupStudy.com - A virtual community of network engineers
 Home  BookStore  StudyNotes  Links  Archives  StudyRooms  HelpWanted  Discounts  Login
Re: Newbie Wireless Question [7:95793] posted 01/08/2005
[Chronological Index] [Thread Index] [Top] [Date Prev][Date Next] [Thread Prev][Thread Next]


Joseph Rubino wrote:
> 
> I was actually referring to some differences between the wired
> and wireless versions.

Very different indeed. Not the same protocol. CSMA/CD != CSMA/CA.

> 
> Question:
> 
> What might happen if 2 wireless clients were within 5 feet or
> less of each other but separated by a concrete wall such that
> they could not detect each other.  Even though they could not
> detect each other; the AP they are associated with has line of
> site to both.

use RTS/CTS as mentioned in the first message.


> 
> 
> Priscilla Oppenheimer wrote:
> > 
> > Joseph Rubino wrote:
> > > 
> > > And don't forget; it's not "real" CSMA/CA because the
> stations
> > > can't actually tell if there has been a collision.  They
> have
> > > to guess when it's their turn to speak.
> > 
> > It _is_ "real" CSMA/CA, but not CSMA/CD, as I mentioned.
> > 
> > Because collision detection isn't easy to do on a wireless
> > network, there's an attempt to avoid collisions. Senders set a
> > random timer and don't start sending until the timer has
> > elapsed. Collisions aren't completely avoided though, so
> frames
> > are acknowledged. All directed (non MAC-layer broadcast)
> frames
> > are ACKed, with the exception of control frames, such as
> Beacon.
> > 
> > So, speaking of bandwidth utilization, not only is it shared,
> > but a reasonably large chunk of bandwidth is used by ACKs and
> > other control frames.
> > 
> > If there seem to be lots of collisions, it's time to subdivide
> > the network, reduce the number of users who can assoicate, put
> > in more access points on different channels that don't
> overlap,
> > etc. (refer to the other branch of the thread, in case I
> didn't
> > say that part quite right. :-)
> > 
> > Also, you can fragment frames at the MAC layer. There will be
> > fewer collisions if nobody hogs the medium for a long time by
> > sending a long frame. For example, if two stations listened,
> > sensed the medium was free, set their random timer and then
> > sent after the timers had elapsed, the second one could
> collide
> > with the first one if the first sender is sending a full-size
> > frame. So, you can fragment at the  MAC layer to avoid this.
> > You may see evidence of this feature in a configurable
> > fragmentation threshold that basically says if collisions
> > increase to a certain threshold, then fragment.
> > 
> > Priscilla
> > 
> > 
> > 
> > 
> > > 
> > > Priscilla Oppenheimer wrote:
> > > > 
> > > > Hi John,
> > > > 
> > > > All varieties of 802.11 provide shared bandwidth. It's
> quite
> > > > noticeable, actually. Using a busy wireless network after
> > > > getting spoiled by switched Ethernet can be a bit
> annoying,
> > > > even with 802.11g (54 Mbps).
> > > > 
> > > > 802.11 uses Carrier Sense Multiple Access with Collision
> > > > Avoidance. Note that it's collision avoidance, whereas
> > shared
> > > > Ethernet is collision detection. With 802.11, the device
> > > > listens before sending, because only one station can send
> > at a
> > > > time. When the device senses that the medium (the air) is
> > > free,
> > > > it waits a short random amount of time and then sends. So
> if
> > > > two devices sense that the medium is free and both start
> > > > sending, they are less likely to collide than with
> Ethernet.
> > > > (Ethernet adds the random delay upon collision detection).
> > > > Collisions aren't completely avoided with CSMA/CA, but
> they
> > > are
> > > > less likely.
> > > > 
> > > > 802.11 can also use a Request to Send/ Clear to Send
> > protocol.
> > > > In this case, the access point must give permission to a
> > > > station to send. Generally, this method should be
> disabled,
> > > > however, it does come in handy when there are two stations
> > > that
> > > > can't see each other but they can see the access point.
> This
> > > is
> > > > called "the hidden node problem."
> > > > 
> > > > Now, there are some proprietary enhancements to the 802.11
> > > > specs. For example, some cards claim to be able to do 108
> > Mbps
> > > > instead of 54 Mbps. I don't know how that works. Anyone
> > know?
> > > > But it's still shared bandwidth. There was also some talk
> a
> > > > while back about switched access points. I'm not
> remembering
> > > > off the top of my head how they worked, but I don't think
> > they
> > > > solved the problem of the airwaves being shared.
> > > > 
> > > > Talk to you later,
> > > > 
> > > > Priscilla
> > > > 
> > > > John Neiberger wrote:
> > > > > 
> > > > > >>> John Neiberger 1/5/05
> > > > > 3:50:53 PM >>>
> > > > > >I really haven't been paying any attention to wireless
> > > > > technologies so
> > > > > >forgive me for asking what must seem like a pretty
> basic
> > > > > question. I was
> > > > > >just in a meeting with some co-workers when the subject
> > of
> > > > > wireless came
> > > > > up,
> > > > > >specifically 802.11g. I was under the impression that
> all
> > > of
> > > > > the 802.11[?]
> > > > > >wireless technologies were multiple access
> technologies,
> > > > > meaning to me
> > > > > that
> > > > > >they are similar to big hubs. In other words, let's
> say a
> > > > > particular
> > > > > access
> > > > > >point is capable of providing 20 Mbps of real
> bandwidth.
> > > > Isn't
> > > > > that shared
> > > > > >bandwidth? After all, only one user can be transmitting
> > at
> > > a
> > > > > time, right?
> > > > > >
> > > > > >Someone rescue me by explaining how this works, or at
> > least
> > > > > pointing me
> > > > > >toward some links. I'm reading a bit about it on the
> web
> > > > right
> > > > > now but I'd
> > > > > >love to get some insight from someone who's actually
> been
> > > > > paying attention
> > > > > >for the past several years. :)
> > > > > 
> > > > > I thought of another way to phrase the question. Let's
> do
> > it
> > > > by
> > > > > example:
> > > > > 
> > > > > Let's say I have an access point a couple of feet away
> > from
> > > my
> > > > > wireless
> > > > > device. Because of the wonderful conditions, I'm able to
> > get
> > > > > real data
> > > > > throughput of around 25 Mbps when transferring a very
> > large
> > > > > file. If I start
> > > > > up another wireless device and initiate another transfer
> > so
> > > > > that I have two
> > > > > transfers going at exactly the same time, will I get 25
> > Mbps
> > > > on
> > > > > each machine
> > > > > or will the perceived bandwidth at each machine drop
> > because
> > > > > they are
> > > > > sharing the media?
> > > > > 
> > > > > John
> > > > > 
> > > > > 
> > > 
> > > 
> > 
> > 
> 
> 


Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=95881&t=95793
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html