Nortel Networks Recording Equipment 1 User Manual
Voice over Wireless LAN Solution Guide
v1.0
December 2005
______________________________________________________________________________________________________
Page 40
Auto-RF is another set of features that can have a negative impact on the overall quality of calls.
As a general rule, dynamic adjustments can create many transient problems with voice calls. If
the channel changes, the handset is forced to roam to another AP that may not be close enough
to maintain quality. Or the other APs in the neighborhood may already be at full call capacity. If
power is decreased, the handset may suddenly be outside of nominal signal strength
expectations and also have to roam. The same problems may occur as with the channel change
case. If power is increased, handsets tend to be more “sticky,” meaning they will not roam away
from their current AP, because the cell is larger. It becomes easier to oversubscribe an AP.
As a general rule, dynamic adjustments can create many transient problems with voice calls. If
the channel changes, the handset is forced to roam to another AP that may not be close enough
to maintain quality. Or the other APs in the neighborhood may already be at full call capacity. If
power is decreased, the handset may suddenly be outside of nominal signal strength
expectations and also have to roam. The same problems may occur as with the channel change
case. If power is increased, handsets tend to be more “sticky,” meaning they will not roam away
from their current AP, because the cell is larger. It becomes easier to oversubscribe an AP.
Because of these effects, Nortel generally recommends that you either disable auto-tuning of
channels or to maximize the timers involved in channel changes so that they occur very
infrequently. Nortel also recommends that auto-tuning of power either be disabled or carefully
controlled in terms of baseline power levels and maximum power increase thresholds.
Specifically, you do not want a radio that is at approximately one-quarter power, for example, to
increase all the way to maximum power in response to client connectivity problems. This would
likely result in the cascading of those problems to neighboring APs using the same channel.
channels or to maximize the timers involved in channel changes so that they occur very
infrequently. Nortel also recommends that auto-tuning of power either be disabled or carefully
controlled in terms of baseline power levels and maximum power increase thresholds.
Specifically, you do not want a radio that is at approximately one-quarter power, for example, to
increase all the way to maximum power in response to client connectivity problems. This would
likely result in the cascading of those problems to neighboring APs using the same channel.
For the WLAN handsets to work properly, you must enable long preambles on the WLAN 2300
series. The longer preamble increases the detectability of voice calls from one handset to
another, reducing the number of collisions and increasing the number of calls per AP. Setting the
preamble to long may create compatibility issues with other device types that require the short
preamble.
series. The longer preamble increases the detectability of voice calls from one handset to
another, reducing the number of collisions and increasing the number of calls per AP. Setting the
preamble to long may create compatibility issues with other device types that require the short
preamble.
2.5.2 Ethernet
Switch
family
The products referred to in this section as Ethernet Switch family are the ES 460-24T-PWR, ES
470-24/48T, and BPS 2000. There are other ES products that have QoS features, but they are
outside the scope of this document. The ES family products can be located in a number of places
in a network of WSS 2300s and AP 2330s. Shown in Figure 14 are some of the potential
interfaces between an ES and various components of the VoWLAN solution. This information
provided here is not meant to imply that an ES must be installed as an interface to each and
every one of these devices. Rather, its intent is to clarify that if you do connect one or more of
these products to an ES, then these are some of the things you can do to help maintain end-to-
end QoS.
470-24/48T, and BPS 2000. There are other ES products that have QoS features, but they are
outside the scope of this document. The ES family products can be located in a number of places
in a network of WSS 2300s and AP 2330s. Shown in Figure 14 are some of the potential
interfaces between an ES and various components of the VoWLAN solution. This information
provided here is not meant to imply that an ES must be installed as an interface to each and
every one of these devices. Rather, its intent is to clarify that if you do connect one or more of
these products to an ES, then these are some of the things you can do to help maintain end-to-
end QoS.
Figure 14 also illustrates the basic architecture of a QoS capable network. Devices on the edge of
the QoS domain, here noted as the DiffServ domain, such as the ES family products in this
diagram, are responsible for identifying traffic, classifying it in terms of QoS or CoS, and marking
the traffic accordingly using the DiffServ Code Point (DSCP) field in the IP packet and/or 802.1p
field in an Ethernet header. These DiffServ edge devices connect to non-DiffServ devices through
an untrusted interface. This interface, and in particular the ingress direction of the interface, is
responsible for classifying and marking traffic as it arrives. The backbone or core of the DiffServ
domain network consists of any number of other DiffServ capable networking devices. The
difference is that these core devices interconnect with each other and the edge DiffServ devices
through trusted interfaces: they simply respect the traffic classifications that are already marked
on the packets. Because of the marking, the core devices are able to give each packet the correct
per-hop forwarding behavior (PHB) without the burden of deep packet inspection of all packets.
the QoS domain, here noted as the DiffServ domain, such as the ES family products in this
diagram, are responsible for identifying traffic, classifying it in terms of QoS or CoS, and marking
the traffic accordingly using the DiffServ Code Point (DSCP) field in the IP packet and/or 802.1p
field in an Ethernet header. These DiffServ edge devices connect to non-DiffServ devices through
an untrusted interface. This interface, and in particular the ingress direction of the interface, is
responsible for classifying and marking traffic as it arrives. The backbone or core of the DiffServ
domain network consists of any number of other DiffServ capable networking devices. The
difference is that these core devices interconnect with each other and the edge DiffServ devices
through trusted interfaces: they simply respect the traffic classifications that are already marked
on the packets. Because of the marking, the core devices are able to give each packet the correct
per-hop forwarding behavior (PHB) without the burden of deep packet inspection of all packets.
In the context of the ES family switches, the three main interfaces into the DiffServ capable QoS
network are those that connect to the AP 2330, the WSS 2300, and the WTM 2245. The following
sections concern traffic ingressing to an ES on each of these three interfaces.
network are those that connect to the AP 2330, the WSS 2300, and the WTM 2245. The following
sections concern traffic ingressing to an ES on each of these three interfaces.