Re: [STDS-802-16] One question about bandwidth distribution
Dear Richard,
As I know, the main access scheme in WLAN is CSMA which is
a contention-based manner. Therefore, the overall throughput
decreases as the number of users increase because the cost
on contention collision depends on the number of users.
On the other hand, WiMAX adopts two scheme for access: contention-based
and centralized polling. The former is only used for transmission of
initial ranging and bandwidth request messages while the latter is used
to schedule the resource allocation for data transmission.
As to your testing case by using FTP, a WLAN CPE firstly partitions the
FTP stream into many packets and transmit request by contention-based CSMA
for EACH packet. On the other hand, a WiMAX CPE only transmit ONE bandwidth
request for the FTP stream and then listen the bandwidth allloction info
in each periodic frame broadcast message (named MAP in WiMAX).
That is main reason to have different performance for WLAM and WiMAX
while you test by two-users scenario.
--
Fang-Ching Ren（任芳慶）
Tel: 886-3-5914786
Fax: 886-3-5820371
Email: frank_ren@itri.org.tw
Section Manager, M100 ICL/ITRI
Bldg.14,195 Sec.4, Chung Hsing Rd.
Chutung, HsinChu, Taiwan
---------------- Original Message ----------------
> Richard <baixin@GMAIL.COM> 2006-04-20 05:44:15 AM wrote:
收件人: STDS-802-16@listserv.ieee.org
主旨: [STDS-802-16] One question about bandwidth distribution
Hi, Everyone,
I did some tests with a WiMAX solution which is FDD full duplex.
I did the throughput tests with one CPE and two CPEs respectively by using
FTP. I want to know whether it works the same as WLAN that the bandwidth
is going to be shared by the CPEs.
For both uplink and downlink, the amount of the throughput of the
clients connected to the two CPEs should almost equal to the throughput of
one cliet with one CPE. Is it the expected result?
Since I am a little comfused about the result I got.
So, please anyone knows the way WiMAX works, give me some help!
Thank you very much!
Richard