Patton electronic SmartNode 4110 Series ユーザーズマニュアル

ページ / 664
Introduction
154
SmartWare Software Configuration Guide 
13 • Link scheduler configuration
service that is only available for a limited bandwidth e.g. low delay. When connecting the SmartNode to a Diff-
Serv
 network shaping might be a required operation.
Burst tolerant shaping or wfq
For weighted fair queuing and shaping there is a variation of the scheduler that allows to specify if a traffic class 
may temporarily receive a higher rate as long as the average stays below the limit. This burstiness measure 
allows the network to explicitly assign buffers to bursty sources. 
When you use shaping on the access link the shaper sometimes has the problem that multiple sources are 
scheduled for the same time - and therefore some of them will be served too late. If the rate of every source had 
to strictly obey its limit, all following packets would also have to be delayed by the same amount, and further 
collisions would reduce the achieved rate even further. To avoid this effect, the SmartWare shaper assumes that 
the burstiness needed for sources to catch up after collisions is implicitly allowed. Future versions of SmartWare 
might allow setting the burst rate and bursting size if more control over its behavior is considered necessary.
Burst tolerance has a different effect when used with weighted fair queuing. Think of it as a higher initial rate 
when a source device starts transmitting data packets. This allows giving a higher weight to short data transfers. 
This feature is sometimes referred to as a service curve.
Hierarchy
An arbiter can either use wfq or shaping to determine which source to serve next. If you want the scheduler to 
follow a combination of decision criteria you can combine different schedulers in hierarchy to do a multi-level 
arbitration.Hierarchical scheduling is supported in SmartWare with service-policy profiles used inside service-
policy profiles.In 
 an example of hierarchical scheduling is illustrated. The 1
st
 level arbiter Level_1 uses 
weighted fair queuing to share the bandwidth among source classes VPN, Web and incorporates the traffic 
from the 2
nd
 level arbiter Low_Priority, which itself uses shaping to share the bandwidth among source classes 
Mail and Default.