802.11g vs 802.11n

Hi,

I am trying to figure out if going to 802.11n from an 802.11g is really
worth it. My internet service is listed as being upto 7MB. Any thoughts on
the switch?

On Dec 5, 8:21 pm, "jrb" <jrbar...@cox.net> wrote:
> Hi,
>
> I am trying to figure out if going to 802.11n from an 802.11g is really
> worth it. My internet service is listed as being upto 7MB. Any thoughts on
> the switch?

The high speed wifi is more useful between local devices rather than
interfacing with the internet. For instance, you might want to set up
a media server, sending the video from your PC to TV.
http://www.dlink.com/products/?pid=547&sec=1
for example. These devices were G, but N are appearing on the market.
I could see wireless network attached storage being draft N.

I think all the wifi routers that have gigabit on the lan are draft N.
That was the driving force behind me going N over G. I do stream
music, but that is find using G. I will eventually get around to
streaming video, so the N will come in useful.

jrb wrote:
> I am trying to figure out if going to 802.11n from an 802.11g is
> really worth it. My internet service is listed as being upto 7MB. Any
> thoughts on the switch?

Even 802.11b at 10 Mbps would be adequate for a 7 Mbps internet connection.

802.11g is rated at 54 Mbps and would be useful for faster file
exchanges within your own network. While it uses ODFM modulation, it is
more robust (more reliable at longer ranges and less with than optimum
signal levels), that advantage is lost as it requires more than twice
the signal level for the higher speed. Locking down 802.11g to 10 Mbps
would give you a very reliable wireless network.

802.11n could be considered an upgrade to 802.11g in terms of robustness
and range.

If your 802.11b networks works fine the way it is, there's no advantage
in spending the money to move up to 802.11n.

DTC <(E-Mail Removed)> hath wroth:

>jrb wrote:
>> I am trying to figure out if going to 802.11n from an 802.11g is
>> really worth it. My internet service is listed as being upto 7MB. Any
>> thoughts on the switch?

Required reading on MIMO:
<http://www.veriwave.com/gurus/index.asp>
This is very much up to date.

802.11n is all about speed. Great for video and massive local file
transfers. Doesn't add much in the way of range. The rule-of-thumb
is that 4 times the speed, will give you 1/2 the range. By the time
you get to speeds greater than 54Mbits/sec, you're talking about 5 to
10 meter ranges.

However, that's for multiple streams (Airgo). To make sure that it's
a total muddle, 802.11n Draft 2 includes beam forming and steering
(Atheros and Ruckus Wireless), which for some amazing reason is
classified as MIMO. It offers no speed enhancements, but does offer
substantial indoor reflection and interference reduction. One thing
that all the various MIMO mutations have in common is that you cannot
easily add an aftermarket antenna (or antennas). If you can't get the
range with what is supplied by the manufacturer, too bad.

>Even 802.11b at 10 Mbps would be adequate for a 7 Mbps internet connection.

Ummm... nope. Thruput is roughly half the connection speed. An
11mbit/sec connection would yield about 5Mbits/sec thruput, which is
much less than the 7Mbit/sec internet speed. The message header for
the OP shows that he's on Cox.net, which has "PowerBoost" speeds to
about 10Mbits/sec for preferred and premier service levels. To
utilize 10Mbits/sec thruput, he needs at least a 24Mbit/sec connection
speed.

>802.11g is rated at 54 Mbps and would be useful for faster file
>exchanges within your own network. While it uses ODFM modulation, it is
>more robust (more reliable at longer ranges and less with than optimum
>signal levels), that advantage is lost as it requires more than twice
>the signal level for the higher speed. Locking down 802.11g to 10 Mbps
>would give you a very reliable wireless network.

Locking it down at 24Mbits/sec would be required to prevent bandwidth
constipation at 10Mbits/sec thruput. However, that doesn't work too
well. I've been recommending that locking down the speed to the
slower OFDM speeds offers a reliability improvement in that the access
point isn't constantly trying to go as fast as possible. That works
well at the slower speeds (I use 12Mbits/sec OFDM). That will a bit
slow for a 7Mbit/sec cable internet, but will not work if the OP has
burstable service. However, my experiments with locking the speed at
much faster rates (I was testing at 36Mbits/sec) has not been so
wonderful. Packet loss starts to creep up. Susceptibility to noise
interference increases. Signal loss and disconnects are more abrupt
and of course, at a shorter range. It will probably work just fine
for a closed room environment (coffee shop, conference room, bedroom,
etc), but isn't too good when going through walls, floors, or in the
presence of substantial interference.

I haven't tried the same tests with any of the MIMO devices, so I can
tell if MIMO adds anything. I doubt it because outside of fairly
short ranges, the typical Airgo MIMO access point reverts to
802.11b/g.

Incidentally, many low end routers can barely move data at cable
internet speeds, or handle many connections:
<http://www.smallnetbuilder.com/component/option,com_chart/Itemid,189/chart,121/>

>802.11n could be considered an upgrade to 802.11g in terms of robustness
>and range.

Ummm... I hope you're not referring to the 2x, 4x, now up to 12x
stickers that are appearing on the retail packages? When I see the
manufactories run a BER/PER test, at various ranges and speeds, I
might believe that manure. Meanwhile, some MIMO articles:
<http://www.smallnetbuilder.com/content/view/30224/100/>
<http://www.smallnetbuilder.com/content/view/30230/100/>
are quite revealing.

>If your 802.11b networks works fine the way it is, there's no advantage
>in spending the money to move up to 802.11n.

Sure there is. There's great benefits to all the 802.11n Draft 2
confusion. It's bad enough that most of what I buy is obsolete in a
few years. With 802.11n Draft 2, I can buy products that are obsolete
on arrival and with no guarantee of a later upgrade to the final
standard. There's also no incentive to produce such an upgrade path,
because that will cut into future replacement sales.

Bah Humbug;
E. Scrooge and Associates.

--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

~ jrb wrote:
~ > I am trying to figure out if going to 802.11n from an 802.11g is
~ > really worth it. My internet service is listed as being upto 7MB. Any
~ > thoughts on the switch?

~ Even 802.11b at 10 Mbps would be adequate for a 7 Mbps internet connection.

No it would not. The maximum nominal signaling rate of 802.11b is 11Mbps.
However, the 802.11 MAC layer has a lot of overhead, so even with a perfect
11Mbps 802.11b physical/MAC layer, this will translate (at the IP layer)
to a best case throughput rate of 5.5 - 6Mbps.

So, 802.11b would be a bottleneck for a 7Mbps Internet connection.

(Terminology note: when I say Mbps, by that I mean MegaBITS per second,
where "Mega" denotes one million [1,000,000].)

~ 802.11g is rated at 54 Mbps and would be useful for faster file
~ exchanges within your own network. While it uses ODFM modulation, it is
~ more robust (more reliable at longer ranges and less with than optimum
~ signal levels), that advantage is lost as it requires more than twice
~ the signal level for the higher speed.

802.11g OFDM is more robust, and with a perfect 54Mbps nominal signaling
rate will yield an IP layer throughput of up to 27Mbps or so.

~ Locking down 802.11g to 10 Mbps
~ would give you a very reliable wireless network.

10Mbps is not an 802.11g rate. You could cap 802.11g at the 24Mbps
signaling rate which would give you a maximum throughput of about
12Mbps, which should be sufficient to prevent the wireless link from
being a bottleneck for your Internet traffic.

However, if there are 802.11b-only clients present in your cell, then
the 802.11g protection mechanisms may reduce your maximum throughput,
such that your nominal 24Mbps signaling rate may no longer be sufficient
to yield 7Mbps of throughput.

Therefore, I would probably stick with the maximum 802.11g rate of
54Mbps, unless you have reason to be that the clients and/or APs are
doing a suboptimal job of rate selection and would therefore benefit from
a rate cap.

~ 802.11n could be considered an upgrade to 802.11g in terms of robustness
~ and range.
~
~ If your 802.11b networks works fine the way it is, there's no advantage
~ in spending the money to move up to 802.11n.

Roger that.

References:

Capacity Coverage & Deployment Considerations for IEEE 802.11g
Cisco whitepaper
http://www.cisco.com/en/US/products/...801d61a3.shtml

When Is 54 Not Equal to 54? A Look at 802.11a, b, and g Throughput
Article by Michael Gast
http://www.oreillynet.com/pub/a/wire...hroughput.html

Aaron

Aaron Leonard wrote:
>
> ~ jrb wrote:
> ~ > I am trying to figure out if going to 802.11n from an 802.11g is
> ~ > really worth it. My internet service is listed as being upto 7MB. Any
> ~ > thoughts on the switch?
>
> ~ Even 802.11b at 10 Mbps would be adequate for a 7 Mbps internet connection.
>
> No it would not. The maximum nominal signaling rate of 802.11b is 11Mbps.
> However, the 802.11 MAC layer has a lot of overhead, so even with a perfect
> 11Mbps 802.11b physical/MAC layer, this will translate (at the IP layer)
> to a best case throughput rate of 5.5 - 6Mbps.
>
> So, 802.11b would be a bottleneck for a 7Mbps Internet connection.
>
> (Terminology note: when I say Mbps, by that I mean MegaBITS per second,
> where "Mega" denotes one million [1,000,000].)
>
> ~ 802.11g is rated at 54 Mbps and would be useful for faster file
> ~ exchanges within your own network. While it uses ODFM modulation, it is
> ~ more robust (more reliable at longer ranges and less with than optimum
> ~ signal levels), that advantage is lost as it requires more than twice
> ~ the signal level for the higher speed.
>
> 802.11g OFDM is more robust, and with a perfect 54Mbps nominal signaling
> rate will yield an IP layer throughput of up to 27Mbps or so.
>
> ~ Locking down 802.11g to 10 Mbps
> ~ would give you a very reliable wireless network.
>
> 10Mbps is not an 802.11g rate. You could cap 802.11g at the 24Mbps
> signaling rate which would give you a maximum throughput of about
> 12Mbps, which should be sufficient to prevent the wireless link from
> being a bottleneck for your Internet traffic.
>
> However, if there are 802.11b-only clients present in your cell, then
> the 802.11g protection mechanisms may reduce your maximum throughput,
> such that your nominal 24Mbps signaling rate may no longer be sufficient
> to yield 7Mbps of throughput.
>
> Therefore, I would probably stick with the maximum 802.11g rate of
> 54Mbps, unless you have reason to be that the clients and/or APs are
> doing a suboptimal job of rate selection and would therefore benefit from
> a rate cap.
>
> ~ 802.11n could be considered an upgrade to 802.11g in terms of robustness
> ~ and range.
> ~
> ~ If your 802.11b networks works fine the way it is, there's no advantage
> ~ in spending the money to move up to 802.11n.
>
> Roger that.
>
> References:
>
> Capacity Coverage & Deployment Considerations for IEEE 802.11g
> Cisco whitepaper
> http://www.cisco.com/en/US/products/...801d61a3.shtml
>
> When Is 54 Not Equal to 54? A Look at 802.11a, b, and g Throughput
> Article by Michael Gast
> http://www.oreillynet.com/pub/a/wire...hroughput.html
>
> Aaron
>
Let me jump in here a minute.
My interest in 'n' is, by far, more for the increased range.
Are you, and others, saying the range is not that much greater?
I need to reach wirelessly about 80-100 feet through 3 walls and one
floor. My 'g' can't reach that far.

Jeff Liebermann wrote:
> Locking it down at 24Mbits/sec would be required to prevent bandwidth
> constipation at 10Mbits/sec thruput. However, that doesn't work too
> well.

I've heard of that with consumer gear. We lock the speed down for our
tier two (three to six mile distance) radios to minimize complaints of
widely varying speeds. At six miles, we can certainly throw a faster
signal, but with only a 10 dB fade margin.

> Ummm... I hope you're not referring to the 2x, 4x, now up to 12x
> stickers that are appearing on the retail packages?

Good grief no! Besides, I don't use the consumer crap.

On Dec 6, 9:15 am, Jeff Liebermann <je...@cruzio.com> wrote:
> DTC <m...@nothingtoseehere.zzx> hath wroth:
>
> >jrb wrote:
> >> I am trying to figure out if going to 802.11n from an 802.11g is
> >> really worth it. My internet service is listed as being upto 7MB. Any
> >> thoughts on the switch?
>
> Required reading on MIMO:
> <http://www.veriwave.com/gurus/index.asp>
> This is very much up to date.
>
> 802.11n is all about speed. Great for video and massive local file
> transfers. Doesn't add much in the way of range. The rule-of-thumb
> is that 4 times the speed, will give you 1/2 the range. By the time
> you get to speeds greater than 54Mbits/sec, you're talking about 5 to
> 10 meter ranges.
>
> However, that's for multiple streams (Airgo). To make sure that it's
> a total muddle, 802.11n Draft 2 includes beam forming and steering
> (Atheros and Ruckus Wireless), which for some amazing reason is
> classified as MIMO. It offers no speed enhancements, but does offer
> substantial indoor reflection and interference reduction. One thing
> that all the various MIMO mutations have in common is that you cannot
> easily add an aftermarket antenna (or antennas). If you can't get the
> range with what is supplied by the manufacturer, too bad.
>
> >Even 802.11b at 10 Mbps would be adequate for a 7 Mbps internet connection.
>
> Ummm... nope. Thruput is roughly half the connection speed. An
> 11mbit/sec connection would yield about 5Mbits/sec thruput, which is
> much less than the 7Mbit/sec internet speed. The message header for
> the OP shows that he's on Cox.net, which has "PowerBoost" speeds to
> about 10Mbits/sec for preferred and premier service levels. To
> utilize 10Mbits/sec thruput, he needs at least a 24Mbit/sec connection
> speed.
>
> >802.11g is rated at 54 Mbps and would be useful for faster file
> >exchanges within your own network. While it uses ODFM modulation, it is
> >more robust (more reliable at longer ranges and less with than optimum
> >signal levels), that advantage is lost as it requires more than twice
> >the signal level for the higher speed. Locking down 802.11g to 10 Mbps
> >would give you a very reliable wireless network.
>
> Locking it down at 24Mbits/sec would be required to prevent bandwidth
> constipation at 10Mbits/sec thruput. However, that doesn't work too
> well. I've been recommending that locking down the speed to the
> slower OFDM speeds offers a reliability improvement in that the access
> point isn't constantly trying to go as fast as possible. That works
> well at the slower speeds (I use 12Mbits/sec OFDM). That will a bit
> slow for a 7Mbit/sec cable internet, but will not work if the OP has
> burstable service. However, my experiments with locking the speed at
> much faster rates (I was testing at 36Mbits/sec) has not been so
> wonderful. Packet loss starts to creep up. Susceptibility to noise
> interference increases. Signal loss and disconnects are more abrupt
> and of course, at a shorter range. It will probably work just fine
> for a closed room environment (coffee shop, conference room, bedroom,
> etc), but isn't too good when going through walls, floors, or in the
> presence of substantial interference.
>
> I haven't tried the same tests with any of the MIMO devices, so I can
> tell if MIMO adds anything. I doubt it because outside of fairly
> short ranges, the typical Airgo MIMO access point reverts to
> 802.11b/g.
>
> Incidentally, many low end routers can barely move data at cable
> internet speeds, or handle many connections:
> <http://www.smallnetbuilder.com/component/option,com_chart/Itemid,189/...>
>
> >802.11n could be considered an upgrade to 802.11g in terms of robustness
> >and range.
>
> Ummm... I hope you're not referring to the 2x, 4x, now up to 12x
> stickers that are appearing on the retail packages? When I see the
> manufactories run a BER/PER test, at various ranges and speeds, I
> might believe that manure. Meanwhile, some MIMO articles:
> <http://www.smallnetbuilder.com/content/view/30224/100/>
> <http://www.smallnetbuilder.com/content/view/30230/100/>
> are quite revealing.
>
> >If your 802.11b networks works fine the way it is, there's no advantage
> >in spending the money to move up to 802.11n.
>
> Sure there is. There's great benefits to all the 802.11n Draft 2
> confusion. It's bad enough that most of what I buy is obsolete in a
> few years. With 802.11n Draft 2, I can buy products that are obsolete
> on arrival and with no guarantee of a later upgrade to the final
> standard. There's also no incentive to produce such an upgrade path,
> because that will cut into future replacement sales.
>
> Bah Humbug;
> E. Scrooge and Associates.
>
> --
> Jeff Liebermann je...@cruzio.com
> 150 Felker St #D http://www.LearnByDestroying.com
> Santa Cruz CA 95060http://802.11junk.com
> Skype: JeffLiebermann AE6KS 831-336-2558

Those photos prompted me to open my dead Netgear WNR854T. It uses a
Marvel board. The interesting thing is they used ferrite chokes on the
antenna coax.

On Thu, 06 Dec 2007 15:48:04 -0500, jay lunis <(E-Mail Removed)>
wrote:

>My interest in 'n' is, by far, more for the increased range.
>Are you, and others, saying the range is not that much greater?

I just happen to dribble by Office Max. On some 802.11n products
being sold, the box declared:
4x the range. 12x the speed
Sounds great? Well, they lie. (Everybody lies, but that's ok because
nobody can understand the numbers anyway).

First, you're NOT going to get both 4x the range and 12x the speed at
the same time. If you believe the hype, it's one or the other.

2nd, 4 time the range of what? Compared to what device and under what
conditions? I'm a bit busy right now (leaky office roof) so I'm going
to suggest that you do the necessary Googling and see if any of the
vendors that use this 4x and 12x manure bother to specify test
conditions on their web piles. They probably do, but see if it
actually resembles something you can use for comparison. Look for at
what range they did the test, with what error rate, and using what
client device for testing.

>I need to reach wirelessly about 80-100 feet through 3 walls and one
>floor. My 'g' can't reach that far.

Neither can an 802.11n Draft 2 router go through 3 assorted walls of
unspecified material and one floor of more of the same. If it's
concrete, stucco, chicken wire, or aluminum foil backed insulation,
you're lucky if it can go through one wall. The only way I know of
going through 3 walls and a floor is with an electric drill and CAT5
cable. If desperate, think about power line (HomePlug) or phone line
(HomePNA) networking.

Incidentally, my rule of thumb is where the 4x and 12x crap came from.
The way Airgo style MIMO (spatial mux) works is to transmit multiple
streams of data at the same time. So, if you're getting perhaps 100ft
of reliable range at 25 Mbits/sec, then with Airgo style MIMO, you'll
get two streams or twice the thruput. However, you can always trade
speed for range. A 2nd stream will give you SQRT(2) times the range,
if you drop the TOTAL speed of the two streams back down to the
previous 25Mbits/sec. If you have 4 streams, and slow things down to
25Mbits/sec, you'll go 2x as far. To go 4x the range, you need 16
streams, which I don't think any of the current incantations are able
to deliver. Similarly, if you're expecting 12x times the speed,
you'll need 12 streams, which is also stretching things a bit. Of
course the spatial mux is far from perfect and tends to create some
self interference. Your mileage may vary.


--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558 (E-Mail Removed)
# http://802.11junk.com (E-Mail Removed)
# http://www.LearnByDestroying.com AE6KS

jay lunis wrote:
> Let me jump in here a minute.
> My interest in 'n' is, by far, more for the increased range.
> Are you, and others, saying the range is not that much greater?
> I need to reach wirelessly about 80-100 feet through 3 walls and one
> floor. My 'g' can't reach that far.

Increased range? sure...in an open environment.

Bubble pack FRS radios can go 12 miles...from a mountain top down into a
valley.