question about Shannon capacity : Theory

Can Shannon capacity be used to estimate the rate for all wireless
technologies or just certain ones? I've been calculating channel
capacities for 802.16 and 802.11 fine, but 802.15.4 is really
confusing me. My data rates are an order of 10 bigger. How do you
calculate the capacity of an 802.15.4 channel?

Also, 802.15.4 specifies fixed data rates. But to me, logically,
varying an amount of a physical layer resource, such as power or
bandwidth, should vary your rate. How are 802.15.4 rates fixed?
Doesn't the Shannon capacity prove that you can only fix a data rate
by adjusting other physical layer properties to compensate?

Any guidance on this would be appreciated. To me using the shannon
limit C = Wlog2(1 + P*G/(w*N + I)) should allow me to estimate the
channel capacities for any technology considering their physical layer
resource restrictions.

Thanks,
Omar

On Wednesday 08 October 2008 14:24, Ginu wrote:
> Can Shannon capacity be used to estimate the rate for all wireless
> technologies or just certain ones?

It applies to all information channels.

> I've been calculating channel
> capacities for 802.16 and 802.11 fine, but 802.15.4 is really
> confusing me. My data rates are an order of 10 bigger. How do you
> calculate the capacity of an 802.15.4 channel?
>
> Also, 802.15.4 specifies fixed data rates. But to me, logically,
> varying an amount of a physical layer resource, such as power or
> bandwidth, should vary your rate. How are 802.15.4 rates fixed?
> Doesn't the Shannon capacity prove that you can only fix a data rate
> by adjusting other physical layer properties to compensate?

Unless the fixed rate is well below the Shannon limit.

-paul-

> > Also, 802.15.4 specifies fixed data rates. But to me, logically,
> > varying an amount of a physical layer resource, such as power or
> > bandwidth, should vary your rate. How are 802.15.4 rates fixed?
> > Doesn't the Shannon capacity prove that you can only fix a data rate
> > by adjusting other physical layer properties to compensate?
>
> Unless the fixed rate is well below the Shannon limit.
>
> -paul-

If the fixed rate is well below the Shannon limit, then?

Thanks!

On Thursday 09 October 2008 14:33, Ginu wrote:

>> > Also, 802.15.4 specifies fixed data rates. But to me, logically,
>> > varying an amount of a physical layer resource, such as power or
>> > bandwidth, should vary your rate. How are 802.15.4 rates fixed?
>> > Doesn't the Shannon capacity prove that you can only fix a data rate
>> > by adjusting other physical layer properties to compensate?
>>
>> Unless the fixed rate is well below the Shannon limit.
>
> If the fixed rate is well below the Shannon limit, then?

If the fixed rate is well belowthe Shannon limit, then channel
parameters can vary (noise, bandwidth) and you still get the same
fixed data rate, even without changing other properties.

To be concrete, suppose the limit for a channel is 1 Gb/s and you
only use 1 Mb/s (0.1%). If noise increases, the limit might drop to
100 Mb/s, but you can still stay with 1 Mb/s without, say, increasing
power. Notice that the percentage of channel used does increase, here
to 1%

I have absolutely no idea if 802.15.4 rates are fixed this way or even
if they are really fixed. The 802.15.4 rate may be a maximum rate,
allowed rate, or rate under "good" circumstances, not fixed per se.

-paul-

On Oct 10, 12:28 pm, "Paul E. Black" <p.bl...@acm.org> wrote:
> On Thursday 09 October 2008 14:33, Ginu wrote:
>
> >> > Also, 802.15.4 specifies fixed data rates. But to me, logically,
> >> > varying an amount of a physical layer resource, such as power or
> >> > bandwidth, should vary your rate. How are 802.15.4 rates fixed?
> >> > Doesn't the Shannon capacity prove that you can only fix a data rate
> >> > by adjusting other physical layer properties to compensate?
>
> >> Unless the fixed rate is well below the Shannon limit.
>
> > If the fixed rate is well below the Shannon limit, then?
>
> If the fixed rate is well belowthe Shannon limit, then channel
> parameters can vary (noise, bandwidth) and you still get the same
> fixed data rate, even without changing other properties.
>
> To be concrete, suppose the limit for a channel is 1 Gb/s and you
> only use 1 Mb/s (0.1%).  If noise increases, the limit might drop to
> 100 Mb/s, but you can still stay with 1 Mb/s without, say, increasing
> power.  Notice that the percentage of channel used does increase, here
> to 1%
>
> I have absolutely no idea if 802.15.4 rates are fixed this way or even
> if they are really fixed.  The 802.15.4 rate may be a maximum rate,
> allowed rate, or rate under "good" circumstances, not fixed per se.
>
> -paul-

Thanks for that explanation. It makes a lot of sense. I'm going to try
to find a way to incorporate this in my simulations. Thank you