User's Manual

AN-50e
user manual
70-00033-02 Proprietary Redline Communications © 2004 November 2004
Page 17 of 96
A single sector AN-50e PMP implements a distributed wireless L2 switch, with
one uplink port located on the sector controller (master) distributing bandwidth to
a variable number of subscriber stations (slaves).
- Each subscriber station (remote-end) is considered a separate wireless link.
- Each configured wireless link adds to the switch one remote-end port that is the
Ethernet port on the corresponding subscriber station.
- Each link (remote-end) is assigned one ID that is used to manage both the
connection traffic and the wireless link.
- The switch supports one broadcast/multicast group called default group, which
is automatically assigned a fixed ID.
The master-end MAC utilizes a request/grant polling mechanism to determine
which subscriber station requires bandwidth. This is achieved by periodically
polling each subscriber station to determine if there is a request for bandwidth. If
the subscriber station requests bandwidth then the MAC allocates the appropriate
number of time slots, in both the downstream and upstream direction, in
accordance with the CIR rate limits specified for that particular subscriber station.
The AN-50e PMP system supports 18 programmable CIR levels. Through an
external calculator the minimum Committed Information rate CIR rates allowed
for each subscriber station can be determined. Note that a minimum CIR of 8 Kbps
implies almost no CIR, and is referred to as Best Effort (BE) service.
With multiple subscriber stations compete for bandwidth, the MAC ensures that
time slots are allocated in a balanced manner, according to the different CIR
levels, during periods of over-subscription. For example, during peak times, the
MAC will first deny time slot allocations for everything above the provisioned
CIR, and then reduce bandwidth in a proportional manner to the remaining units
with varying CIR levels and priorities. With the support of CIR, the service
provider can offer different grades of service to each end user in a controlled
manner based on their service level agreements or contracts.