HP (Hewlett-Packard) 5300XL ユーザーズマニュアル

HP ProCurve Switch 5300xl Series Reviewer’s Guide 

 

•  If a 5300 Management VLAN is enabled it cannot be defined as an XRRP interface – SNMP 

management requests to a particular 5300 need to go to that physical switch regardless of 
fail-over status. 

•  XRRP does not interoperate with VRRP, but can coexist in a VRRP environment without 

interference. 

For more details see the HP ProCurve Switch 5300xl Series documentation located at: 

http://www.hp.com/go/hpprocurve

 under the Technical Support section. 

Quality of Service (QoS) mechanisms in the HP ProCurve Switch 5300xl Series provide the network 
manager control over packet flows based on a number of factors. In addition, since the switches can 

override the priority values in the incoming packets, the network manager can maintain QoS control 
over inappropriate priority designations coming from users or applications at the end nodes. 

Conversely, many applications can be given priority treatment through the switch without the end node 

clients having to be aware of QoS, particularly valuable since client operating systems and the 
applications themselves are generally not QoS aware at this time. 
The primary means of control is through priority queues in the switch. Pieces of information in the 
packet that can be used to determine priority queue placement are called classifiers. The mechanism to 

actually store the priority based on the classifiers is through the 802.1Q tags or through the IP 
TOS/Diffserv section in the IP header. The HP ProCurve Switch 5300xl Series do not modify either of 

the packet fields when routing the packet (unless a QoS override is specified in the port config) so the 

QoS status of a packet can be maintained as the packet travels elsewhere in the network. Each of these 
is discussed in the following sections. 
The final section, End-to-End QoS, briefly discusses the value of QoS in networks. 

Each port on an HP ProCurve Switch 5300xl Series module has four priority queues. A packet placed 

into a particular queue will be processed according to the priority of that queue. 
The priority queues are managed through a fair-weighted queuing algorithm that prevents any priority 

queue from getting starved (the packets in it not being serviced by the switch) due to high traffic levels 

in higher priority queues. 

Through user configuration, priority of packets can be specified based on the following classifiers. This 

list is in order of precedence; if there are multiple classifiers that apply to a specific packet, the one 
that is highest on this list takes effect. 

•  Layer 4 TCP/UDP port numbers: allows prioritization based on the application associated 

with the packet. This allows, for instance, VoIP packets using fixed port numbers to be 

prioritized higher than other traffic. It can also be used to downgrade packet flows, such 

as HTTP traffic. Can also be used to remap the diffserv code points (DSCP). 

•  Device Priority (destination or source IP address) : up to 256 addresses can be specified per 

chassis, destination address takes precedence over source address. Can also be used to 
remap the DSCP. 

•  IP Type of Service (ToS) field (IP packets only): support for both the older TOS IP 

Precedence definition, or the newer Differentiated Services (Diffserv) definition. If 
using the TOS IP Precedence, the bits are mapped to packet priority queues using the 

same table as shown in the next section “IEEE 802.1p Priority Support”. The 802.1p bits 
are also set for the outbound packet if the packet goes out of the switch through a port 

 

© Hewlett-Packard Co. 2002, 2003 

Rev 1.1 – 2/11/2003 

http://www.hp.com/go/hpprocurve

 

Page 17 of 35