Quality of Service Using HP Virtual Connect Technologies Connected to Cisco Networks

Transcription

1 Technical white paper Quality of Service Using HP Virtual Connect Technologies Connected to Cisco Networks Last updated: 08/29/2014 Table of contents Introduction... 2 What is Quality of Service (QoS)?... 2 What is HP Virtual Connect?... 2 The networking basics of QoS... 2 Why do we need to think about QoS?... 3 Understanding the basics of QoS... 3 Layer 2 markings... 3 Layer 3 markings... 4 Bringing the markings together... 5 Virtual Connect and QoS... 5 Virtual Connect QoS modes of operation... 5 Understanding Virtual Connect QoS behaviors... 6 Virtual Connect QoS architecture... 7 Architectural topology example using Cisco Voice... 9 QoS voice reference architectural drawing... 9 How to configure Virtual Connect to provide QoS for Cisco Voice Cisco QoS for Nexus Overview Useful links Summery

2 Introduction This guide provides an understanding of how to configure custom Quality of Service (QoS) for Virtual Connect with a focus on Cisco Nexus Networking. It will also explain the need for QoS in today s data networks. In addition, we will include networking basics of QoS followed by Layer 2 and Layer 3 markings. Furthermore, we discuss the features of the HP Virtual Connect (VC) Solution including the VC QoS architecture, modes of operation, and behavior. Then the guide concludes with a QoS Voice example configuring VC to a Cisco Nexus. Finally, we summarize and include a listing of useful information for further reference. What is Quality of Service (QoS)? Quality of service (QoS) is the ability to provide different networking priorities to different applications, users, or data flows or to guarantee a certain level of application performance. For example, an application must be guaranteed throughput or minimum response time for service levels mandated such as a financial database transaction. However, on-demand applications such as voice or video (media-intensive) and real-time streaming consume huge amounts of bandwidth causing other applications to suffer and vice versa. Quality of service guarantees are important if the network capacity is insufficient. You might consider increasing the network speed and bandwidth to higher levels like 40 Gb Ethernet to solve a network performance or congestion challenge, but increasing the network speed could potentially cause the problem more quickly. QoS shapes the network traffic based on rules and policies set by your organizational priorities, especially for real-time streaming multimedia applications such as voice over IP, online games, and IP-TV. These often require a fixed bit rate and are sensitive to delay. QoS is of particular importance in networks where the capacity is limited. Having more bandwidth like a 10/40 Gigabit Ethernet does not eliminate the need for QoS, as we still want to guarantee performance of certain traffic flows over others in heavy utilization situations. What is HP Virtual Connect? HP Virtual Connect (VC) is a set of interconnect modules and embedded software for HP BladeSystem c-class enclosures that simplifies the setup and administration of server connections. VC implements server edge virtualization between the server and the data center infrastructure so networks can communicate with individual servers or pools of HP BladeSystem servers. You can upgrade, replace, or move server blades within the enclosures without visible changes to the external LAN and SAN environments. The external networks connect to a shared resource server pool rather than to individual servers. VC cleanly separates server enclosure administration from LAN and SAN administration. Virtual Connect includes the following components: HP Virtual Connect Manager HP Virtual Connect Manager (VCM) centralizes connection management for HP Blade Servers that use Virtual Connect Modules to access LANs, SANs, and converged network infrastructures. VCM is embedded on VC-Enet and FlexFabric modules. VC-Enet modules VC-Enet and FlexFabric modules enable connectivity to data center Ethernet switches. VC-Enet and FlexFabric modules can also be directly connected to other types of devices, such as printers, laptops, rack servers, and network storage devices. VC-FC modules The VC-FC and FlexFabric modules enable connectivity of the enclosure to data center FC switches. Every FC fabric is limited in the number of switches it can support, but the VC-FC modules do not appear as switches to the FC fabric and do not count against FC fabric limits. Note: The VC-FC modules require the use of a VC-Enet Module to configure and manage the VC-FC modules. The networking basics of QoS Quality of Service on modern network equipment typically leverages a Modular QoS Command-line Interface (MQC). MQC traffic is classified using a combination of packet markings tied together with class maps and match commands that roll into a policy defining how traffic will be prioritized and optimized throughout a network of routers and switches. In short, this breaks down into Layer 2 (reliable switch-like direct point-to-point data connection) and Layer 3 (addressing, routing, and not necessarily reliable) packet markings for things such as voice and video to get priority over something like web surfing. 2

3 Why do we need to think about QoS? Quality of Service is the most overlooked configuration item until you reach a performance bottleneck or have a bad experience with choppy voice or video. The general school of thought assumes that sufficient bandwidth avoids needing to think about Quality of Service. In reality, bandwidth does not solve the problem. Increasing bandwidth temporarily fixes the quality until stressed resources and bottlenecks recur. Configuring Quality of Service delivers a business policy to the network and the applications by setting a standard expectation of performance and reliability regardless of the bandwidth, ensuring resources are fully utilized to maximum capabilities. Understanding the basics of QoS Layer 2 markings Quality of Service Layer 2 Markings are defined by the IEEE 802.1p a mechanism for marking packets at the Media Access Control (MAC) Layer of the Open System Interconnection (OSI) model. A switch operates at Layer 2. Some switches operate at Layer 3 as well, but we will focus on Layer 2. The QoS technique developed in the IEEE 802.1p, also known as class of service (CoS), is a 3-bit field called the Priority Code Point (PCP). It specifies a priority value of between 0 and 7 inclusive that can be used by QoS disciplines/policies to differentiate traffic. Although this technique is commonly referred to as IEEE 802.1p, there is no standard or amendment by that name published by the IEEE. Rather than change the 802.1p standard the technique is incorporated into IEEE 802.1Q standard, which specifies the tag inserted into an Ethernet frame. Priority levels Eight different classes of service are available as expressed through the 3-bit PCP field in an IEEE 802.1Q header added to the frame. The way traffic is treated when assigned to any particular class is undefined and left to the implementation. Table 1: CoS value to PCP bit-field mapping CoS Value Bit Field CoS CoS CoS CoS CoS CoS CoS CoS IP precedence values Table 2: Precedence values Value Description Traffic Characteristics 000 (0) Routine or Best Effort Background 001 (1) Priority Best Effort 010 (2) Immediate Excellent Effort 011 (3) Flash (mainly used for voice signaling or for video) Critical Applications (FCOE, Call Control) 100 (4) Flash Override Video, < 100 ms latency 101 (5) Critical (mainly used for voice RTP) Voice, < 10 ms latency 110 (6) Internet Internetwork Control 111 (7) Network Network Control (OSPF, BGP) The IEEE 802.1Q specification defines a standards-based mechanism for providing VLAN tagging and class of service (CoS) across Ethernet networks. This task is accomplished through an additional 4-byte tag, which carries VLAN and frame prioritization information, inserted within the header of a Layer 2 Ethernet frame, as shown in Figure 1. 3

4 Figure 1: Ethernet frame with 802.1Q tag. Preamble Dest. MAC Addr. Source MAC Addr Q Tag Type/ Length Payload FCS/ CRC The 802.1Q tag has a specific format, consisting of four fixed-length fields. Two of the four fields carry the frame prioritization and VLAN information. Figure 2 illustrates the format of the 802.1Q tag itself. Figure 2: IEEE 802.1Q tag format Tag Protocol Identifier (TPID) = 0x8100 PCP C F I VLAN Identifier (VID) The following are the fields within the 802.1Q tag: Tag Protocol Identifier (TPID): 802.1Q tagged Ethernet frames are indicated by a value of 0x8100 within this 16-bit field. Priority Code Point (PCP): A 3-bit field that indicates the frame priority level Canonical Format Indicator (CFI): A 1-bit field set to 0 for Ethernet; used for compatibility between Ethernet and Token Ring VLAN Identifier (VID): A 12-bit field that specifies the VLAN to which the frame belongs The following priority assignment is an example of how administrators might configure traffic. Since the operation of a network requires network control traffic such as OSPF or BGP, it is given the highest priority. Voice is next because it is considered inelastic traffic, which cannot handle fluctuations in throughput and is sensitive to delay and packet loss. Video can withstand more delay than voice and has the ability to degrade quality in the event of network congestion. Critical Applications is high priority data traffic with a guaranteed minimum bandwidth. Excellent-effort traffic is data traffic that is not critical but more important than the default/non-prioritized best-effort traffic. Finally, background traffic consists of bulk transfers, which should not affect the service of any other traffic classes. Layer 3 markings Since packets are marked in the IP header for Layer 3 routing, QoS (Layer 2) needs a method to ensure that performance and reliability defined in the QoS policy maps to the various application priorities. Differentiated Services Code Point (DCSP) is the Layer 3 IP header marking, and it uses a 6-bit (DS Field) in the IP header for packet classification purposes. Table 3: Commonly used DSCP values DSCP Value Decimal Value Meaning Best Effort AF AF AF AF AF AF AF AF AF AF AF AF High Priority Expedited Forwarding (EF) 4

5 Figure 3: Sample IP Packet (Note that the field labeled as ToS is the IP Header packet field that will have the first 6 bits signifying the DSCP marking). Bringing the markings together Now that we have talked about the key Layer 2 and Layer 3 markings for QoS, it is important to understand how they come together. Depending on how the network is laid out, we will have boundaries defined for Layer 2 and Layer 3, and with these we need to consider a trust for Layer 2 boundaries. If you uplink a Layer 2 boundary switch into a switch that will service both routing and switching you will need to trust your Layer 2 CoS markings and prepare to transition them into Layer 3 DSCP Markings. This allows you to maintain performance and reliability driven by the business policy. Layer 2 markings transition to Layer 3, and Layer 3 markings can transition to Layer 2 to leverage hardware resources like buffers and queues. For this paper, we will not get into buffers and queues because they differ between hardware platforms and vendors. Virtual Connect and QoS Virtual Connect QoS modes of operation Virtual Connect has three primary modes of operation for setting Quality of Service markings. The modes of operation are outlined below: Passthrough: Incoming non-fcoe packets are not classified or altered. PCP values from the user s VLAN tag will be passed unchanged and will not be used for QoS. There is no classification and traffic management. This configuration is for customers who do not want to use QoS in VC but want to use QoS in other switches. Custom (with FCoE Lossless): Enable QoS and allow a customized configuration that includes FCoE class. The configuration defines two system classes: Best Effort and FCoE Lossless. You can configure six additional classes for non- FCoE Ethernet traffic. You must configure traffic class parameters and traffic classification. Custom (without FCoE Lossless): Enable QoS and allow a customized configuration. The configuration defines one system class (Best Effort), and you can configure seven additional classes for non-fcoe Ethernet traffic. You must configure traffic class parameters and traffic classification. You cannot switch to this type when the domain has a fabric associated with an FCoE-capable interconnect module, a shared uplink set as an FCoE network, or a server profile which has an FCoE connection. 5

6 Figure 4: Virtual Connect QoS Modes/Types. HP Virtual Connect Manager User Administrator Home Sign Out Domain Status Define Configure Tools Help x Domain Status View Legend...? i Quality of Service (QoS) Find Configuration Items.? IP Address QoS Configuration Type: Passthrough Enclosures Backup/Restore Storage Mgmt Credentials Traffic Classes Passthrough Custom (with FCoE Lossless) Custom (without FCoE Lossless) Apply Cancel SNMP Configuration System Log Stacking Links + Users/Authentication - Ethernet MAC Addresses Port Monitoring Advanced Settings Quality of Service (QoS) For the purposes of this paper, we are going to focus on the Custom (with FCOE Lossless) configuration. Understanding Virtual Connect QoS behaviors The first step for QoS is to categorize different kinds of traffic. Various L2/L3/L4-7 parameters are used to classify the traffic. This phase is called classification. In the next step, classified traffic is managed within the switch chip to provide different priorities and scheduling. For end-to-end QoS all hops are configured with similar QoS policies of classification and traffic management. Figure 5: The following figure shows how traffic flow currently works on Virtual Connect without QoS. Traffic is simply processed in a First-In First-Out (FIFO) approach. VC Enet Module FIFO Ingress Ethernet Traffic Egress Ethernet Traffic Traffic without QoS Figure 6 shows how traffic flows when QoS is enabled on Virtual Connect. With QoS configured, traffic flows according to resource allocation and priority. In the previous figure, packets enter in the order 1,2,3,4,5, and 6 but exit in the order 3,5,1,4, and 6. During congestion, packet 2 is dropped based on low resources allocated for it. Virtual Connect leverages Weighted Round Robin (WRR) from a queuing perspective. Figure 6: VC with Traffic QoS Enabled VC Enet Module Classification Traffic class 1 Traffic management Ingress Ethernet Traffic Traffic class 2 Egress Ethernet Traffic Traffic class 3 Traffic class 4 6 Traffic with QoS

7 Traffic from server to VC uses minimum and maximum bandwidth. This bandwidth management is enforced by FlexNIC firmware. Traffic from VC to server on flex ports does not use min-max. VC will enforce QoS policies on the physical port level. All flex ports are aggregated to a physical port. Virtual Connect QoS architecture High reference architecture Figure 7 illustrates how a packet flows in the Virtual Connect Module regarding QoS from the perspective of ingress and egress: During ingress, traffic on incoming ports is classified into different traffic classes. Classified traffic then goes through traffic management: This traffic is queued on the appropriate traffic class/queue of the ingress port. Then classified traffic will be queued on the appropriate traffic class/queue of the destination (egress) port. Considering the bandwidth allocation for each class/queue, the switch scheduler will select the appropriate packet for egress. Before the packet leaves the egress port, an egress PCP value associated with traffic class will be marked on the packet. This only happens if the egress packet has a VLAN tag. For tunneled traffic, the original PCP value will be untouched. Virtual Connect reserves appropriate bandwidth for high priority traffic so packets from that class get preference during egress. Figure 7: VC QoS reference architecture Ingress Port Real Time Ingress VC Ethernet Traffic (includes FCoE, iscsi) Ingress VLAN Translation Determine traffic class Inress Logic Lossless Custom 5 Custom 4 Custom 3 Custom 2 Egress Logic Custom 1 Best Effort Egress Logic Real Time Max Share Lossless Share, Max Share Custom 5 Share, Max Share Custom 4 Share, Max Share Custom 3 Share, Max Share Custom 2 Share, Max Share Custom 1 Share, Max Share Priority plus Weight Deficit Weighted Round Robin Egress Port Engress VLAN translation. Mark Co5 on outer VLAN tag if mapped network. FCoE to FC Ethernet Traffic (includes FCoE, iscsi) FC Traffic Egress VC Best Effort Share, Max Share VC traffic classification Traffic can be classified in different ways in VC, and these methods are referred to as ingress/egress traffic classifiers. For VC, the following classification mechanisms/classifiers are supported: PCP: A 3-bit value in VLAN tag. PCP value from VLAN tags will be used for classification and not the FlexNIC VLAN tag. DSCP: A byte-long Differentiated Services Code Point (DSCP) filed in IP header. 12 Assured forwarding, 1 Expedited Forward, and 8 Class Selector Fields Supported in VC: All other traffic is considered best effort. 7

8 Figure 8: Traffic classification flow When DSCP and PCP are both in use, DSCP will be used to classify IP traffic, and PCP will be used for non-ip traffic. The classification is supported for Ethernet and iscsi traffic. FCOE in Virtual Connect does not provide classification. Within Virtual Connect Manager (VCM) interface, the ability to choose different ingress classifiers for uplinks and downlinks is available. Configuration is required to set which PCP/DSCP value goes to which traffic class. Figure 9: Traffic-class PCP flow 8

9 System classes cannot be disabled. Traffic classes have the following properties: Share: The share value is the minimum guaranteed bandwidth that each traffic class gets. The share value is configurable except for system classes. The FCoE_Lossless class gets its share from profile connection configuration. The Best Effort class share is automatically calculated. The sum of shares of all enabled classes and Best Effort class add up to 100. Maximum share: The maximum share is the maximum bandwidth available to use when other traffic classes are not using their maximum share. The FCOE_Lossless class gets its maximum share from fabric speed settings. Real time: Only one user-defined class can be real time. Traffic from this class will have highest priority and will be scheduled in strict priority order. The maximum share and the share for real-time classes must be equal to or less than 10. Egress Dot1p (PCP) Priority: Traffic classified in a particular class will egress with egress PCP markings on a VLAN tag. System classes have predefined egress PCP (3 for FCoE Lossless and 0 for Best Effort). The user needs to configure the egress PCP value for other four classes. Egress classes are not applicable to tunneled traffic. Figure 10: Traffic classes Architectural topology example using Cisco Voice QoS voice reference architectural drawing 9

10 How to configure Virtual Connect to provide QoS for Cisco Voice QoS custom mode with FCOE Figure 11: QoS custom mode with FCOE QOS Custom Mode: With FCOE 2 Systems Classes 6 User Defined Classes Switch to Server Direction Real Time (Highest Priority) Uplink Classification Based on Dot1p DSCP FCOE FCOE Lossless Custom Class 0 (Default Name: Medium) Custom Class 1 (Disabled by Default) Custom Class 2 (Disabled by Default) Custom Class 3 (Disabled by Default) Dot1p=3 Dot1p Marking Ingress Port Custom Class 4 (Disabled by Default) Best Effort Dot1p=0 Egress Port Virtual Connect Server to Switch Direction Dot1p=3 Dot1p Marking Dot1p=0 Real Time (Highest Priority) FCOE Lossless Custom Class 0 (Default Name: Medium) Custom Class 1 (Disabled by Default) Custom Class 2 (Disabled by Default) Custom Class 3 (Disabled by Default) Custom Class 4 (Disabled by Default) Best Effort FCOE Egress Port Virtual Connect Downlink Classification Based on Dot1p DSCP Ingress Port QoS configuration in Virtual Connect Figure 12: Step 1: Choose a QoS mode QoS Configuraton Step 1: Choose QoS Mode HP Virtual Connect Manager User Administrator Home Sign Out Domain Status Domain Status View Legend... x? i Define Configure Tools Help Quality of Service (QoS) Find Configuration Items.? Domain Settings Configuration IP Address Enclosures QoS Configuration Type: QoS customization with FCoE Traffic Classes Passthrough Passthrough Custom (with FCoE Lossless) Custom (without FCoE Lossless) Default mode QoS customization without FCoE Backup/Restore Storage Mgmt Credentials SNMP Configuration System Log Stacking Links + Users/Authentication - Ethernet MAC Addresses Port Monitoring Advanced Settings Quality of Service (QoS) IGMP Settings 10

11 Figure 13: Step 2: Egress queue definition QoS Config Step 2: Egress Queue Definition HP Virtual Connect Manager Domain Status Define Configure Tools Help x Domain Status View Legend...? i Quality of Service (QoS) Find Configuration Items.? Configuration QoS Configuration Type: Custom (with FCoE Lossless) Reset IP Address Enclosures Backup/Restore Storage Mgmt Credentials Traffic Classes Traffic Classes Ingress Traffic Classifiers For each user defined class, you can optionally config engress dot1p value SNMP Configuration System Log Stacking Links + Users/Authentication - Ethernet MAC Addresses Port Monitoring Name Real Time Share Max Share Egress DOT1P Priority Enabled FCoE_Lossless Per Connection* Per Fabric* 3 Best_Effort System Classes, users Medium can t change Class Class Class User defined Classes Class Real time *For the FCoE_Lossless traffic class, the Share is based on the profile connection configuration and the Max Share is based on the fabric configuration. Advanced Settings Quality of Service (QoS) IGMP Settings Config desired Share and Max Share value. The value is percentage based on physical link speed. For example, for VC egress to Blade 1 with 10G physical link, Real Time class 10% share will have 1G of guaranteed share upon congestion. Users can choose enable or disable the class Figure 14: Step 3: ingress classification VCM CLI checks and verification show QoS-configuration The command shows the Quality of Service configuration including configuration type, traffic classes, traffic classifications, and classification mappings. All the QoS command outputs will be sorted in a consistent manner before being displayed. show QoS-configtype This command displays the QoS configuration type that is configured. show QoS-class The command shows the Quality of Service traffic classes settings. show QoS-classifier This command displays the traffic classifier configuration. show QoS-map The command displays the traffic classification maps. 11

12 Cisco QoS for Nexus Overview Trust boundaries The incoming interface enforces the trust boundary as follows: All Fiber Channel and virtual Fiber Channel interfaces are automatically classified into the FCoE system class. By default, all Ethernet interfaces are trusted interfaces. A packet tagged with an 802.1p CoS value is classified into a system class using the value in the packet. Any packet that is not tagged with an 802.1p CoS value is classified into the default drop system class. If the untagged packet is sent over a trunk, it is tagged with the default untagged CoS value, which is zero. You can override the default untagged CoS value for an Ethernet interface or port channel. After the system applies the untagged CoS value, QoS functions the same as for a packet that entered the system tagged with the CoS value. Ingress queuing policies You can associate an ingress policy map with an Ethernet interface to guarantee bandwidth for the specified traffic class or to specify a priority queue. The ingress policy is applied in the adapter to all outgoing traffic that matches the specified CoS value. When you configure an ingress policy for an interface, the switch sends the configuration data to the adapter. If the adapter does not support the DCBX protocol or the ingress policy type-length-value (TLV), the ingress policy configuration is ignored. Ingress classification policies You use classification to partition traffic into classes. You classify the traffic based on the port characteristics (CoS field) or the packet header fields that include IP precedence, Differentiated Services Code Point (DSCP), and Layer 2 to Layer 4 parameters. The values used to classify traffic are called match criteria. When you define a traffic class, you can specify multiple match criteria, you can choose not to match on a particular criterion, or you can determine traffic class by matching any or all criteria. Traffic that fails to match any class is assigned to a default class of traffic called class-default. Egress queuing policies You can associate an egress policy map with an Ethernet interface to guarantee the bandwidth for the specified traffic class or to configure the egress queues. The bandwidth allocation limit applies to all traffic on the interface including any FCoE traffic. Each Ethernet interface supports up to six queues, one for each system class. In addition to the six queues, control traffic that is destined for the CPU uses strict priority queues. These queues are not accessible for user configuration. FCoE traffic (traffic that maps to the FCoE system class) is assigned a queue. This queue uses weighted round-robin (WRR) scheduling with 50% of the bandwidth. Standard Ethernet traffic in the default drop system class is assigned a queue. This queue uses WRR scheduling with 50% of the bandwidth. If you add a system class, a queue is assigned to the class. You must reconfigure the bandwidth allocation on all affected interfaces. Bandwidth is not dedicated automatically to user-defined system classes. You can configure a strict priority queue. This queue is serviced before all other queues except the control traffic queue (which carries control rather than data traffic). 12

13 Example configuration These examples can be used in conjunction with the Nexus 7000, 5000, and They are provided as a sample configuration, not a complete configuration, as you may need to adjust specifics to your environment. Configuring global traffic classification based on CoS This classification will be for global CoS to qos-group mapping. class-map type qos match-all cos3-hpn-global match cos 3 class-map type qos match-all cos5-hpn-global match cos 5 Assign each global CoS class to a qos-group in a qos policy-map policy-map type qos HPN-Classify-Global class cos5-hpn-global set qos-group 5 class cos3-hpn-global set qos-group 3 Configuring traffic classification based on DSCP ip access-list HPN-Qos-Signaling remark Deny voice, video, and audio streams expressly NOT related to remark QoS call signaling functions. remark SCCP and Secure SCP permit tcp any any eq 2000 permit tcp any any eq 2443 remark H323 Q931 and H225 Call Signaling permit udp any any eq 1719 permit tcp any any eq 1720 ip access-list HPN-Qos-Voice remark Phone media traffic between valid endpoints (IP phones, IPTgateways, remark & IPT servers) permit udp 10.x.x.x 0.x.x.x range x.x.x 0.x.x.x range ip access-list HPN-Qos-MM-Stream remark Multimedia Streaming traffic from sanctioned streaming audio/video endpoints remark at the access layer permit udp 10.x.x.x 0.x.x.x gt x.x.x 0.x.x.x gt 1023 Class-maps will use these ACLs class-map type qos match-any HPN-Voice-In description Voice/VoIP/IPT match access-group name HPN-Qos-Voice class-map type qos match-any HPN-MM-Stream-In description Multimedia Streaming match access-group name HPN-Qos-MM-Stream class-map type qos match-any HPN-Voice-Signal-In description Call Signaling match access-group name HPN-Qos-Signaling 13

14 These classes are assigned to a qos-group and marked with a DSCP value using a policy map. The policy map will later be applied to the N5K interfaces. Note that multiple traffic classes are assigned to the same qos-group: policy-map type qos HPN-Inbound-Nexus description Inbound classification/marking policy for trust boundaries. class HPN-Voice-In set dscp ef set qos-group 5 class HPN-Inbound-MM-Stream set dscp af31 set qos-group 3 class HPN-Voice-Signal-In set dscp cs3 set qos-group 3 Apply policy to interface Note: For a port channel member, the service-policy needs to be configured under the interface port channel, not to individual channel ports. interface ethernet mod/port[-port] or port-channel # Use a range of interfaces if possible. Use a range of interfaces if possible. servicepolicy HPN-Inbound-Nexus end Configuring egress scheduling and queuing based on QoS-Groups The following type-queuing configuration commands support the transmit queue assignments and bandwidth allocations: class-map type queuing HPN-Queue-Cos5 match qos-group 5 class-map type queuing HPN-Queue-Cos4 match qos-group 4 class-map type queuing HPN-Queue-Cos6-3-2 match qos-group 3 class-map type queuing HPN-Queue-Cos1 match qos-group 2 policy-map type queuing HPN-Queue-Global class type queuing class-default bandwidth percent 15 class type queuing class-fcoe bandwidth percent 30 class type queuing HPN-Queue-Cos5 priority bandwidth percent 10 class type queuing HPN-Queue-Cos4 bandwidth percent 10 class type queuing HPN-Queue-Cos6-3-2 bandwidth percent 30 class type queuing HPN-Queue-Cos1 bandwidth percent 5 14

15 Marking CoS based on QoS-Groups The following type network-qos configuration commands support the global policy that marks CoS values based on existing QoS-groups: class-map type network-qos cos5 match qos-group 5 class-map type network-qos cos4 match qos-group 4 class-map type network-qos cos6-3-2 match qos-group 3 class-map type network-qos cos1 match qos-group 2 policy-map type network-qos HPN-Global-Net class type network-qos cos5 set cos 5 mtu 9216 class type network-qos cos4 set cos 4 mtu 9216 class type network-qos cos6-3-2 set cos 3 mtu 9216 class type network-qos cos1 set cos 1 mtu 9216 Establishing the system QoS configuration The global system QoS configuration on a Nexus 5000/2000 switch is used for the N2K ports. system qos service-policy type qos input HPN-Classify-Global The input QoS policy allows the CoS-based classification to be applied globally to support the N2K ports. Follow these guidelines when globally establishing the system QoS configuration on a Nexus 5000 switch. system qos service-policy type queuing output HPN-Queue-Global service-policy type queuing input HPN-Queue-Global service-policy type network-qos HPN-Global-Net Applying a traffic-classification policy based on DSCP to support the non-n2k ports. The last step is to apply the interface level QoS configuration based on DSCP. interface ethernet mod/port[-port] or port-channel # Use a range of interfaces if possible. service-policy input HPN-Inbound-Nexus end Configuring Auto QoS on Catalyst 3850 to IP Phone HPN(config)# interface GigabitEthernet1/0/5 HPN(config-if)# auto qos voip cisco-phone HPN(config-if)# end HPN# show policy-map interface GigabitEthernet1/0/5 GigabitEthernet1/0/5 Service-policy input: AutoQos-4.0-CiscoPhone-Input-Policy 15

16 Class-map: AutoQos-4.0-Voip-Data-CiscoPhone-Class (match-any) Match: cos 5 QoS Set dscp ef police: cir bps, bc 8000 bytes conformed 0 bytes; actions: transmit exceeded 0 bytes; actions: set-dscp-transmit dscp table policed-dscp conformed 0000 bps, exceed 0000 bps Class-map: AutoQos-4.0-Voip-Signal-CiscoPhone-Class (match-any) Match: cos 3 QoS Set dscp cs3 police: cir bps, bc 8000 bytes conformed 0 bytes; actions: transmit exceeded 0 bytes; actions: set-dscp-transmit dscp table policed-dscp conformed 0000 bps, exceed 0000 bps Class-map: AutoQos-4.0-Default-Class (match-any) Match: access-group name AutoQos-4.0-Acl-Default QoS Set dscp default Class-map: class-default (match-any) Match: any Service-policy output: AutoQos-4.0-Output-Policy queue stats for all priority classes: Queuing priority level 1 (total drops) 0 (bytes output) 0 16 Class-map: AutoQos-4.0-Output-Priority-Queue (match-any) Match: dscp cs4 (32) cs5 (40) ef (46) Match: cos 5 Priority: 30% ( kbps), burst bytes ,

17 Priority Level: 1 Class-map: AutoQos-4.0-Output-Control-Mgmt-Queue (match-any) Match: dscp cs2 (16) cs3 (24) cs6 (48) cs7 (56) Match: cos 3 Queueing queue-limit dscp 16 percent 80 queue-limit dscp 24 percent 90 queue-limit dscp 48 percent 100 queue-limit dscp 56 percent 100 (total drops) 0 (bytes output) 0 bandwidth remaining 10% queue-buffers ratio 10 Class-map: AutoQos-4.0-Output-Multimedia-Conf-Queue (match-any) Match: dscp af41 (34) af42 (36) af43 (38) Match: cos 4 Queuing (total drops) 0 (bytes output) 0 bandwidth remaining 10% queue-buffers ratio 10 Class-map: AutoQos-4.0-Output-Trans-Data-Queue (match-any) Match: dscp af21 (18) af22 (20) af23 (22) Match: cos 2 Queuing (total drops) 0 (bytes output) 0 bandwidth remaining 10% queue-buffers ratio 10 Class-map: AutoQos-4.0-Output-Bulk-Data-Queue (match-any) Match: dscp af11 (10) af12 (12) af13 (14) Match: cos 1 Queuing 17

18 (total drops) 0 (bytes output) 0 bandwidth remaining 4% queue-buffers ratio 10 Class-map: AutoQos-4.0-Output-Scavenger-Queue (match-any) Match: dscp cs1 (8) Queuing (total drops) 0 (bytes output) 0 bandwidth remaining 1% queue-buffers ratio 10 Class-map: AutoQos-4.0-Output-Multimedia-Strm-Queue (match-any) Match: dscp af31 (26) af32 (28) af33 (30) Queuing (total drops) 0 (bytes output) 0 bandwidth remaining 10% queue-buffers ratio 10 Class-map: class-default (match-any) Match: any Queuing (total drops) 0 (bytes output) 0 bandwidth remaining 25% queue-buffers ratio 25 Configuring Auto Qos on Catalyst 6800 for Marked Traffic HPN-Routers(config)# interface GigabitEthernet1/0/5 HPN-Router(config-if)# auto qos voip trust HPN-Router(config-if)# end 18

19 Useful links Cisco Enterprise SRND QoS: cisco.com/c/en/us/td/docs/solutions/enterprise/wan_and_man/qos_srnd/qos-srnd-book.html Nexus 7000 QoS Example: cisco.com/c/en/us/support/docs/switches/nexus-7000-series-switches/ nex7000-qos-config-00.html Nexus 5000 QoS Example: QoS_Config_7x_chapter_01010.html Virtual Connect and HP BladeSystem Resources: Summary In summary, this whitepaper provides a detailed understanding of Quality of Service for Virtual Connect along with an explanation of how it integrates with other vendors. Even with the increases in bandwidth, QoS is still required as a safety mechanism to ensure key traffic such as voice or video in a network always has priority over other nonessential traffic. Sign up for updates hp.com/go/getupdated Share with colleagues Rate this document Copyright 2014 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein. 4AA4-xxxxENW, August 2014