Radwin 2000+ SERIES, Руководство пользователя

Страница: 211 / 276

RADWIN 2000+ Series User Manual Release 4.2.40 16 ‐ 4
Assigning Queue Priorities
16.4.2 Assigning Queue Priorities
Four mechanisms are available to control queue performance:
• Strict ‐ An over ‐ the ‐ air packet is loaded first with data from Real Tome queue (see
Figure 16 ‐ 2 ), and then from the remaining queues in order. If there is too much data for
the first queue, it will “starve” the lower queues and so on.
• Weight ‐ One solution to such “starvation” is to weight input flow to the four queues.
This could still lead to wasted resources. If in the example below, the actual traffic is all
Near Real Time, only 20% will be allocated to Near real time and the rest will be held up.
• Maximum Information Rate (MIR) ‐ This is a simple throttle mechanism, which suffers
from a drawback similar to the previous case.
• Time to Live (TTL) in ms ‐ This device is use to reduce re ‐ transmits of real time data (such
as voice or viewed video).

To configure queue priorities for uplink and downlink in turn:
1. For each queue, both downlink and uplink enter the required weight, MIR and TTL.
The MIR and TTL may respectively be left Unlimited and Disabled in which case the
system will use a “best effort” method.
If you exceed 100% total weight, you will receive an error message.
Figure 16 ‐ 2: The four QoS data queues
You will be required to correct this before leaving the window other than by cancel ‐
lation.
If you are under ‐ booked, for example by setting a queue to zero, the unused weight
will be distributed to the remaining queues. The effect of doing this will only
become apparent under congestion. In particular, a queue set to zero weight will
become nearly blocked under congestion with packets passing through on a best
effort basis.
2. When you complete your entries, in the QoS Configuration window ( Figure 16 ‐ 1 ),
click OK to save them and continue.