Flashcards in QoS Deck (410):
List the four traffic characteristics that QoS tools can affect.
Bandwidth, delay, jitter, and loss.
Describe some of the characteristics of voice traffic when no QoS is applied in a network.
Voice is hard to understand; voice breaks up, sounds choppy; calls are disconnected; large delays make it difficult to know when the other caller has finished talking.
Describe some of the characteristics of video traffic when no QoS is applied in a network.
Picture displays erratically; picture shows jerky movements; audio not in sync with video; movements slow down; video stream stops; picture black due to missing pixels; frozen video.
Describe some of the characteristics of data traffic when no QoS is applied in a network.
Data arrives too late to be useful; erratic response times cause users to stop using application; customer care agents waiting on screen refresh, so customer waits.
Interpret the meaning of the phrase, “QoS is both ‘managed fairness,’ and at the same time ‘managed unfairness’.”
QoS tools improve QoS characteristics for particular packets. However, improving one packet’s behavior typically comes at the expense of another packet. The terms “managed fairness” and “managed unfairness” just refer to the fact that QoS policies may be fair to one packet but unfair to another.
Define bandwidth. Compare and contrast bandwidth concepts over point-to-point links versus Frame Relay.
Bandwidth refers to the number of bits per second that can reasonably be expected to be successfully delivered across a network. With point-to-point networks, bandwidth is equal to the speed of the link—the clock rate. With Frame Relay, the actual bandwidth is difficult to define. Typically, the minimum bandwidth equals the CIR of a VC. However, engineers at the provider and at the customer typically expect more than CIR to get through the network. The maximum bandwidth would be bounded by the slower of the two access rates on the access links.
Compare and contrast bandwidth and clock rate in relation to usage for QoS.
Bandwidth refers to the router’s perceived bandwidth on the interface/subinterface, and is referenced by QoS tools. Clock rate defines the physical encoding rate on router interfaces that provide clocking; QoS tools ignore the clock rate setting.
List the QoS tool types that affect bandwidth, and give a brief explanation of why each tool can affect bandwidth.
Compression, CAC, and queuing affect bandwidth. Compression reduces the number of bits needed to transmit a frame, allowing more frames to be sent over the same amount of bandwidth. CAC reduces the overall load of voice and video traffic in the network by disallowing new calls. Queuing can reserve subsets of the bandwidth on a link for a particular queue, guaranteeing a minimum amount of bandwidth for that queue.
Define delay, compare/contrast one-way and round-trip delay, and characterize the types of packets for which one-way delay is important.
Delay is the time taken from when a frame/packet is sent until it is received on the other side of the network. One-way delay just measures the delay for a packet from one endpoint in the network to its destination. Round-trip delay measures the time it takes to send a packet to one destination and for a response packet to be received. Voice and video are concerned with one-way delay.
List the categories of delay that could be experienced by all three types of traffic: data, voice, and video.
Serialization, propagation, queuing, forwarding/processing, shaping, network. Note that codec, packetization, and de-jitter delays are unique to voice and video, so technically these delays should not have been part of your answer for this question.
Define, compare, and contrast serialization and propagation delay.
Serialization delay defines the time it takes to encode a frame onto the physical link. For instance, on a point-to-point link of 56 kbps, a bit is encoded every 1/56,000 seconds; therefore, a frame that is 1000 bits long takes 1000/56000 seconds to encode on the link. So, serialization delay is a function of link speed and length of the frame. Propagation delay defines the time taken for a single bit to be delivered across some physical medium, and is based solely on the length of the physical link, and the speed of energy across that medium. If that same point-to-point link were 1000 km (approximately 620 miles) in length, the propagation delay would be 1,000,000m/2.1 * 108 ms, or 4.8 milliseconds.
Define network delay.
Network delay refers to the delay incurred by a packet inside a packet network, like ATM, Frame Relay, or MPLS networks. Because the customer does not know the details of these networks, and because many customers’ packets share the carrier network, variable delays occur.
List the QoS tool types that affect delay and give a brief explanation of why each tool can affect delay.
Queuing, link fragmentation and interleaving, compression, and traffic shaping. Queuing methods use an algorithm to choose from which queue to take the next packet for transmission, which can decrease delay for some packets and increase delay for others. LFI tools break large frames into smaller frames, so that smaller delay-sensitive frames can be sent after the first short fragment, instead of having to wait for the entire large original frame to be sent. Compression helps delay because it reduces the overall load in the network, reducing congestion, reducing queue lengths, and reducing serialization delays. Finally, traffic shaping actually increases delay, but it can be applied for one type of traffic, allowing other traffic to be sent with less delay. Also, policing can have an indirect impact on delay by preventing other traffic from consuming too much of a link, thereby lessening the length of queues, and reducing queuing delay.
Define jitter. Give an example that shows a packet without jitter, followed by a packet with jitter.
Jitter measures the change in delay experienced by consecutive packets. If a PC sends four packets one after the other, practically at the same time, say 1 ms apart, so the departure times are T=0, T=1, T=2, and T=3, for instance, packets arrive at T=70, T=71, T=80, T=81, respectively. The second packet was sent 1 ms after the first, and arrived 1 ms after the first packet—so no jitter was experienced. However, the third packet arrived 9 ms after the second packet, after being sent 1 ms after the second packet—so 8 ms of jitter was experienced.
List the QoS tool types that affect jitter and give a brief explanation of why each tool can affect jitter.
Queuing, link fragmentation and interleaving, compression, and traffic shaping. These same QoS tools can be used for addressing delay issues. Queuing can always be used to service a jitter-sensitive queue first if packets are waiting, which greatly reduces delay and jitter. LFI decreases jitter by removing the chance that a jitter-sensitive packet will be waiting behind a very large packet. Compression helps by reducing overall delay, which has a net effect of reducing jitter. Traffic shaping may actually increase jitter, so it should be used with care—but if shaping is applied to jitter-insensitive traffic only, jitter-sensitive traffic will actually have lower delays and jitter.
Define packet loss and describe the primary reason for loss for which QoS tools can help.
Packet loss means that a packet, which has entered the network, does not get delivered to the endpoint—it is lost in transit. Routers and switches drop packets for many reasons. However, QoS tools can affect the behavior of loss when packets will be lost due to queues being too full. When a queue is full, and another packet needs to be added to the queue, tail drop occurs.
List the QoS tool types that affect loss and give a brief explanation of why each tool can affect loss.
Queuing and RED. Queuing tools allow definition of a longer or shorter maximum queue length; the longer the queue, the less likely that drops will occur. Also by placing traffic into different queues, more variable traffic may experience more loss, because those queues will be more likely to fill. RED tools preemptively discard packets before queues fill, hoping to get some TCP connections to slow down, which reduces the overall load in the network—which shortens queues, reducing the likelihood of packet loss.
Describe the contents of an IP packet carrying the payload for a G.729 VoIP call.
The IP packet contains an IP header, a UDP header, an RTP header, and the voice payload. With G.729, the payload uses 20 bytes, with an 8-byte UDP header, and a 12-byte RTP header. The IP header is 20 bytes long.
Describe the amount of bandwidth required for G.711 and G.729 VoIP calls, ignoring data-link header/trailer overhead.
G.711 consumes 64 kbps for payload, plus another 16 kbps for the IP, UDP, and RTP headers, for a total of 80 kbps. G.729 consumes 8 kbps for payload, plus another 16 kbps for IP, UDP, and RTP headers, for a total of 24 kbps.
List the delay components that voice calls experience, but which data-only flows do not experience.
Codec delay, packetization delay, and de-jitter initial playout delay.
Define the meaning of the term “packetization delay” in relation to a voice call.
Voice must be converted from sound waves to analog electrical signals, and finally to digital signals, and then placed into a packet. Before 20 ms of voice digital payload can be placed into a packet, the speaker must speak for 20 ms. Packetization delay refers to the (default) 20 ms of delay, waiting for the speaker to speak long enough to fill the packet with the correctly sized payload.
List the different one-way delay budgets as suggested by Cisco and the ITU.
The ITU in document G.114 suggests a budget of up to 150 ms for quality voice calls; Cisco suggests a delay budget of up to 200 ms one-way if you cannot meet the 150-ms goal.
Define the term “codec delay” and discuss the two components when using a G.729 codec.
Voice calls incur codec delay when the codec converts the analog signal into digital voice payload. Every codec requires some time to process the incoming signal, which adds delay. With G.729, because it is predictive, it must also wait for some additional incoming voice to arrive, because it is algorithm-processing the voice sample to be encoded, plus a part of the next sample that will be encoded. The delay waiting for the additional voice is called “look-ahead” delay.
Describe the affects of a single lost packet versus two consecutive lost packets, for a G.729 voice call.
Lost voice packets result in the receiver having a period of silence corresponding the length of voice payload inside the lost packet(s). With two consecutive G.729 packets lost, 40 ms of voice is lost; the G.729 codec cannot predict and generate replacement signals when more than 30 ms of consecutive voice is lost. A single lost G.729 packet would only cause a 20-ms break in the voice, which could be regenerated. So, a single lost packet is not perceived as loss in a G.729 call.
Describe a typical video payload flow in terms of packet sizes and packet rates.
Video payloads use variable-length packets. The packet rates are also typically variable.
Discuss the delay requirements of video traffic.
Interactive video (video conferencing, for instance) requires low delay because it is interactive. Delay budgets up to 200 ms are the norm. However, streaming video—one-way video—can tolerate long delays. When playing an e-learning video, for instance, the playout may start after 30 seconds of video has been received into a de-jitter buffer—but each packet may have experienced several seconds of delay.
List the basic differences between TCP and UDP traffic.
TCP performs error recovery, whereas UDP does not. TCP also uses dynamic windowing to perform flow control, whereas UDP does not. Both use port numbers to multiplex among various applications running on a single computer.
Contrast the QoS characteristics needed by interactive data applications, as compared to the QoS needs of voice payload flows.
Bandwidth demands vary greatly for data applications, whereas a single voice call uses a static amount of bandwidth. Delay for interactive data can be relatively longer than for voice, but the key measurement for data is application response time, which includes round-trip packet delays. Finally, data applications are much more tolerant of packet loss, because either the application will resend the data, or rely on TCP to resend the data, or just not care whether some data is lost.
What are the three steps suggested in this chapter for planning QoS policy implementation?
Identify traffic and its requirements
Divide traffic into classes
Define QoS policies for each class.
List and give a brief description of the two audit steps for developing QoS policy.
The two steps are a network audit and a business audit. The network audit identifies the applications and prototocols that are used in the network, using tools like Sniffers and NBAR. The business audit examines compares the business needs for these applications, compared to the protocols found in the network audit, in order to decide of the importance of each applicataion.
What is QoS?
QoS stands for quality of service. In one sense, it is “managed fairness,” and at the same time it is “managed unfairness”—you purposefully choose to favor one packet over another. To quote the Cisco QoS course: “The ability of the network to provide better or “special” service to a set of users/applications to the detriment of other users/ applications.” In either case, the goal is to improve the behavior of one class of traffic, knowing that it will most likely degrade another type of traffic.
What is the purpose of service classes when implementing a QoS policy?
To align business priorities with network resrouces.
List the two classification and marking tools mentioned in this chapter, including the full names and popular acronyms.
Class-Based Marking (CB Marking), Network-Based Application Recognition (NBAR).
List four queuing tools, including the full names and popular acronyms.
Priority Queuing (PQ), Custom Queuing (CQ), Weighted Fair Queuing (WFQ), IP RTP Priority, Class-Based WFQ (CBWFQ), Low Latency Queuing (LLQ), Modified Deficit Round-Robin (MDRR).
List the two shaping tools mentioned in this chapter, including the full names and popular acronyms.
Frame Relay traffic shaping (FRTS) and Class-Based shaping (CB shaping).
List three Congestion Avoidance tools, including the full names and popular acronyms.
Random Early Detection (RED), Weighted RED (WRED), Explict Congestion Notification (ECN).
List four link efficiency tools, including the full names and popular acronyms.
Payload compression, RTP header compression (cRTP), TCP header compression, Multilink PPP fragmentation and interleaving (MLPPP LFI), Frame Relay fragmentation (FRF), link fragmentation and interleaving for Frame Relay and ATM VCs.
List the QoS tools that perform some classification function.
This is a bit of a trick question. Almost all IOS QoS tools perform classification—for instance, to place two different types of packets into two different queues, the queue tool performs classification.
Which of the following tools can be used for classification and marking? CB marking, PQ, CB shaping, WFQ, WRED, FRTS, LLQ, MLPPP LFI, NBAR, QPM, cRTP
CB marking. NBAR can be used for classification in conjunction with CB Marking.
Which of the following tools can be used for queuing? CB marking, PQ, CB shaping, WFQ, WRED, FRTS, LLQ, MLPPP LFI, NBAR, QPM, cRTP
WFQ, LLQ, PQ
Which of the following tools can be used for shaping? CB marking, PQ, CB shaping, WFQ, WRED, FRTS, LLQ, MLPPP LFI, NBAR, QPM, cRTP
CB shaping and FRTS
Which of the following tools can be used for link efficiency? CB marking, PQ, CB shaping, WFQ, WRED, FRTS, LLQ, MLPPP LFI, NBAR, QPM, cRTP
cRTP, MLPPP LFI
Define the DiffServ term behavior aggregate.
According to RFC 2475, a behavior aggregate is “a collection of packets with the same DS code point crossing a link in a particular direction.” The key points are that the DSCP has been set; the packets all move the same direction; and the packets collectively make up a class.
Define the DiffServ term DSCP, including what the acronym stands for.
According to RFC 2475, DSCP refers to “a specific value of the DSCP portion of the DS field, used to select a PHB.” The acronym stands for differentiated services code point. It is the 6-bit filed in the redefined ToS byte in the IP header used for marking packets for DiffServ.
Define the DiffServ term PHB, including what the acronym stands for.
According to RFC 2475, PHB refers to “the externally observable forwarding behavior applied at a DS-compliant node to a DS behavior aggregate.” The acronym stands for per-hop behavior. It is the collection of QoS actions that occur at one router (hop) in a network for a particular BA.
Define the DiffServ term MF classifier, including what the acronym stands for.
According to RFC 2475, an MF classifier is “a multi-field (MF) classifier which selects packets based on the content of some arbitrary number of header fields; typically some combination of source address, destination address, DS field, protocol ID, source port and destination port.” It is the classification function used to classify packets before the DSCP has been set.
Define the DiffServ term DS ingress node, including what the acronym stands for.
According to RFC 2475, a DS ingress node is “a DS boundary node in its role in handling traffic as it enters a DS domain.” DS stands for differentiated services. The term defines a node at which packets enter the DiffServ domain.
Compare and contrast the terms BA classifier and MF classifier, according to DiffServ specifications. Suggest typical points in the network where each is used.
A classifier is a DiffServ function that classifies or categories packets based on the contents of fields in the packet headers. A BA classifier performs this function only based on the DSCP field. An MF classifier can look at many fields in the packet header. MF classifiers typically classify ingress traffic near the edge of a network, and work with markers to set the DSCP field. BA classifiers are used at points in the network after an MF classifier and marker have set the DSCP field values.
Compare and contrast the contents of the IP ToS byte before and after the advent of DiffServ.
Before DiffServ, the ToS byte contained a 3-bit Precedence field, 4 bits in a ToS field, and 1 reserved bit. DiffServ redefined the ToS byte to contain a 6-bit DSCP field, which contains the DSCP values, and 2 reserved bits.
Describe the QoS behavior at a single DS node when using the AF PHB. Also explain what the acronym AF PHB represents and identify the RFC that defines it.
The assured forwarding per-hop behavior, as defined in RFC 2597, defines a PHB with two components. The first part defines four BAs or classes, each which should be placed in a separate queue and given a configured guaranteed minimum amount of bandwidth. The second component provides three different drop probabilities for a Congestion Avoidance tool such as RED.
Explain (by comparing and contrasting) whether AF and CS PHB DSCPs conform to the concept that “bigger DSCP values are better than smaller values.”
CS uses values that have three binary 0s at the end, and the eight IP precedence values for the first three bits. In other words, CS includes the eight binary values for a 6-bit number for which the last three digits are 0s. CS conforms to the idea that a bigger value is better, to be backward compatible with IP precedence. AF uses 12 different values. Of the three AF DSCPs in each class, the highest of the three values actually receives the worst drop preference.
Describe the QoS behavior at a single DS node when using the EF PHB. Also explain what the acronym EF PHB represents and identify the RFC that defines it.
The expedited forwarding per-hop behavior, as defined in RFC 2598, defines a PHB with two components. The first part defines queuing, with features that reserve bandwidth for a single BA, with the added feature on minimizing latency, delay, and loss. The other action of the PHB provides a policing/dropper function, disallowing traffic beyond a configured maximum bandwidth for the class.
Describe the process used by RSVP to reserve bandwidth in a network.
A host signals to the network using an RSVP reservation request using an RSVP path message. The request passes along the route to the destination host; at each intermediate router, if that router can guarantee the right bandwidth, the request is forwarded. When received by the destination host, it replies with an RSVP resv message. The process is reversed, with each router passing the reserve message if it can guarantee the bandwidth in the opposite direction. If the original host receives the reservation message, the bandwidth has been reserved.
Compare and contrast DiffServ and IntServ in terms of using classes, flows, and scalability.
IntServ applies to individual flows, whereas DiffServ differentiates traffic into classes. With large networks and the Internet, the number of IntServ-managed flows does not scale, because information retained about each flow, and the RSVP signaling messages for each flow, continues throughout the life of each flow. DiffServ uses classes, and the number of classes does not increase when packet volumes increase, which allows better scalability.
List and describe the two key advantages of the Best Effort model for QoS.
Best Effort (BE) scales well, because routers and switches do not have to perform any extra work for each packet or frame. And because Best Effort does no specific PHB, it also requires no specific QoS tools.
List and describe the two key advantages of the DiffServ model for QoS.
DiffServ scales well mainly due to its Class-Based operation. Also, DiffServ provides a large number of different classes, ensuring that most networks will have plenty of different classes for their network traffic.
List and describe the two key disadvantages of the DiffServ model for QoS.
DiffServ tools can be complicated, which requires more training and higher skill levels. Also, DiffServ does attempt to provide the appropriate bandwidth, delay, jitter, and loss characteristics, but it does not absolutely guarantee those characteristics.
List and describe the two key disadvantages of the IntServ model for QoS.
The main problem is poor scalability. IntServ scales poorly because it is flow-based, it signals repetitively for each flow, and the nodes must keep flow state information for each flow.
List the three major configuration steps, and the main command used in each step, for the configuration of a QoS feature using MQC.
irst, classification is configured with class-map commands. Then, PHBs are configured using a policy-map. Finally, the policy map is enabled for input or output packets on an interface using the service-policy command.
Describe two different ways with which you could classify packets with DSCP AF31, AF32, and AF33 into a single class using MQC commands.
Inside a class-map, the match dscp AF31 AF32 AF33 command would match packets that had any of the three DSCP values. Alternately, you could use a class-map with the match-any parameter, followed by three match dscp commands, one for each DSCP value.
List 3 benefits of MQC as compared with non-MQC-based QoS features.
Reduces the effort taken to configure QoS. Configuration of classification and PHBs are separated from the interfaces, allowing more concise configuration and more flexibility. Uniform command syntax across multiple QoS features in a single device. Uniform command syntax across router and IOS-based switch platforms. Class maps are reusable for multiple QoS policy maps.
List the two SNMP MIBs included in Cisco router IOS that can be used by QPM to improve the statistics presented to a QPM user. List the long version of the names and the acronyms.
The Class-Based QoS MIB (CBQoSMIB) and the Cisco NBAR Protocol Discovery (CNPD) MIB.
What information can be seen using the CBQoSMIB that cannot be seen with show commands on the device being managed?
The CBQoSMIB allows you to see statistics about packets before and after the application of a policy.
How many classes can be associated with a single policy map in Cisco IOS Software Release 12.2(15)T?
On a router using AutoQoS, what command enables the feature for Frame Relay VCs that use Frame Relay to ATM service interworking?
auto qos voip fr-atm
On a router using AutoQoS, what command enables the feature on a serial interface when the router can trust the DSCP settings of incoming packets?
auto qos voip trust
Describe the classification configuration created by a router when enabling AutoQoS on a serial interface, with all default values chosen on the auto qos command.
The router classifies voice payload into one service class, voice signaling into another, and all other traffic into a third. It uses NBAR and ACL for matching the voice traffic.
Describe the marking actions created by a router when enabling AutoQoS on a serial interface, with all default values chosen on the auto qos command.
The router classifies marks voice payload with DSCP EF, voice signaling with DSCP AF31, and all other traffic as DSCP BE.
List three of the requirements on router AutoQoS that need to be true before actually configuring AutoQoS.
IP CEF must be enabled on each interface or ATM VC, unless the trust option will be used. The bandwidth command should be configured correctly on each interface or VC. Any existing service-policy commands should be removed from interfaces on which AutoQoS will be enabled. Frame Relay must also use only point-to-point subinterfaces.
List the data link protocols on a router that support AutoQoS.
ATM, Frame Relay, HDLC, and PPP.
List the PHBs created by a router when the auto qos voip command is used on a PPP serial interface with default bandwidth setting.
Classification and Marking, Queuing (LLQ)
List the PHBs created by a router when the auto qos voip command is used on a PPP serial interface with bandwidth 768 is configured.
Classification and Marking, Queuing (LLQ), MLP LFI, and cRTP. All links at 768 kbps or less also have cRTP and LFI added as PHBs.
List the PHBs created by a router when the auto qos voip command is used on a Frame Relay PVC with bandwidth 832 is configured.
Classification and Marking, Queuing (LLQ), plus Shaping with FRTS. Had bandwidth 768 been configured, Frame Relay fragmentation and cRTP would also have been configured.
When configuring AutoQoS on a router, with a Frame Relay interface, what configuration mode must you be in before using the auto qos command? What command gets you into that configuration mode?
You must have used the frame-relay interface-dlci command to get into DLCI configuration mode.
When configuring a 2950 switch with the auto qos voip trust command, what PHBs are configured on the interface?
Queuing with queue 4 as the low latency queue, CoS-to-DSCP maps that correlate AF31 to CoS 3 and EF to Cos 5, and the trusting of incoming CoS values.
When configuring a 2950 switch with the auto qos voip cisco-phone command, what PHBs are configured on the interface?
Queuing with queue 4 as the low latency queue, CoS-to-DSCP maps that correlate AF31 to CoS 3 and EF to CoS 5, and the trusting of incoming CoS values. However, it also includes extending the trust boundary to the IP Phone, so if an IP Phone is not found, all frames are considered to be CoS 0.
When configuring a 2950 switch with the auto qos voip cisco-phone command, what version of CDP is required in order for AutoQoS to work at all?
CDP Version 2
When planning to use AutoQoS on a 2950 switch, what types of ports are generally configured with the trust option, and what type are generally configured with the cisco-phone option?
Ports that are connected to end users are configured with cisco-phone. Ports connected via 802.1Q trunks to other switches (or to trusted servers), for which those switches or servers have already marked CoS correctly, are configured with the trust option.
Comparing the CLI of older QoS options in a Cisco router, MQC, and AutoQoS, which takes the least time to implement?
Comparing the CLI of older QoS options in a Cisco router, MQC, and AutoQoS, which is considered to be the most modular?
AutoQoS and MQC are both modular, whereas older QoS features that do not use MQC are not.
Comparing the CLI of older QoS options in a Cisco router, MQC, and AutoQoS, which is considered to be the most difficult to use?
Describe the difference between classification and marking.
Classification processes packet headers, or possibly other information, to differentiate between multiple packets. Marking changes a field inside the frame or packet header.
Describe, in general, how a queuing feature could take advantage of the work performed by a classification and marking feature.
Queuing features can perform their own classification function to place different packets into different queues. After a classification and marking tool has marked a packet, the queuing feature can look for the marked value when classifying packets.
Characterize what must be true before the CoS field may be useful for marking packets.
CoS only exists in 802.1P/Q headers and ISL headers. In turn, these headers are used only on Ethernet links that use trunking. Therefore, the CoS field can only be marked or reacted to for Ethernet frames that cross an 802.1Q or ISL trunk.
Most other QoS tools, besides classification and marking tools, also have a classification feature. Describe the advantage of classification, in terms of overall QoS design and policies, and explain why classification and marking is useful, in spite of the fact that other tools also classify the traffic.
lassification and marking, near the ingress edge of a network, can reduce the amount of work required for classification by other QoS tools. In particular, many QoS tools can classify based on marked fields without using an ACL, which reduces overhead for each QoS tool. By marking packets near the ingress edge, QoS policies can be more consistently applied. In addition, configurations for most other QoS tools become simpler, which can reduce configuration errors in the network.
Which of the following classification and marking tools can classify based on the contents of an HTTP URL: class-based marking (CB Marking), QoS Pre-classification, network-based application recognition (NBAR), or cos-to-dscp maps?
NBAR actually performs the classification based on HTTP header contents. CB Marking is the only tool that marks based on NBAR’s match of the URL string.
Describe the differences between IP extended ACLs as compared with NBAR for matching TCP and UDP port numbers.
You can use both tools to match packet based on well-known port numbers. However, some higher-layer protocols allocate dynamic port numbers, making the use of extended ACLs difficult at best. NBAR can look further into the packet contents to identify what dynamic ports are currently in use by certain protocols, and match packets using those dynamic ports.
Which of the following QoS marking fields are carried inside an 802.1Q header: QoS, CoS, DE, ToS byte, User Priority, ToS bits, CLP, Precedence, QoS Group, DSCP, MPLS Experimental, or DS?
CoS and User Priority. CoS is the more general name, with User Priority specifically referring to the 3-bit field in the 802.1P header.
Which of the following QoS marking fields are carried inside an IP header: QoS, CoS, DE, ToS byte, User Priority, ToS bits, CLP, Precedence, QoS Group, DSCP, or MPLS Experimental?
ToS byte, ToS bits, Precedence, DSCP.
Which of the following QoS marking fields are never marked inside a frame that exits a router: QoS, CoS, DE, ToS byte, User Priority, ToS bits, CLP, Precedence, QoS Group, DSCP, MPLS Experimental, or DS?
QoS Group is only used for internal purposes in GSR and 7500 series routers.
Describe the goal of marking near the edge of a network in light of the meaning of the term “trust boundary.”
ood QoS design calls for classification and marking, based on well-defined QoS policies, as near to the ingress edge of the network as possible. However, packets marked in devices near the edge of the network may be able to be re-marked by devices whose administrators cannot be trusted. A packet can be marked by the end-user PC, for instance, but the end user can configure the value to be marked. An IP Phone, however, can mark packets, and the marked values cannot be overridden by the user of the phone. Therefore, the goal of marking near the edge must be tempered against the fact that some devices can be reconfigured for QoS by those outside the group responsible for QoS.
What configuration command lists the classification details when configuring CB Marking? What configuration mode must you use to configure the command? What commands must you issue to place the configuration mode user into that mode?
The match command defines the details of what must be matched to classify a packet. The command is a subcommand under the class-map global configuration command.
What configuration command lists the marking details when configuring CB Marking? What configuration mode must you use to configure the command? What commands must you issue to place the configuration mode user into that mode?
The set command defines what value to mark in the frame or packet header once a packet is classified. The command is a subcommand under the class command, which is a subcommand under the policy-map global configuration command.
What configuration command enables CB Marking? What configuration mode must you use to configure the command? What commands must you issue to place the configuration mode user into that mode?
The service-policy command enables CB Marking for either input or output packets on an interface. The command refers to the policy map, which in turn refers to the class maps. The command is a subcommand under the interface global configuration command.
Describe how you can match multiple DSCP values with a single class map. How many can you match with a single command?
The match ip dscp command allows for up to 8 DSCP values to be listed, so a single command can match 8 values. If you want to match more in a single class map, you could use multiple match ip dscp commands, with the match-any option configured on the class-map command.
What configuration command lets you match RTP audio without also matching RTP video traffic?
The match protocol rtp audio command.
Describe the process by which NBAR can be updated to support new protocols, without upgrading IOS.
Cisco builds Packet Descriptor Language Modules (PDLMs). These PDLMs define new protocols to NBAR. By downloading a copy of these from Cisco, and putting the PDLM in Flash memory, and reloading the router, NBAR knows how to identify new protocols, without requiring an updated IOS image.
What CB Marking command implies that a policy map requires NBAR in order to match packets?
The match protocol command means that the policy map will use NBAR for matching the packets.
What command enables NBAR on an interface for incoming packets? For outgoing packets?
The ip nbar protocol-discovery command enables NBAR for packets in each direction.
Describe the reason why you might see multiple set commands inside a single service class in a policy map, and give one example
Multiple set commands means that the CB Marking policy is marking more than one header field. That may be useful when later devices might look at different marked fields. For example, a router fastethernet interface might have a policy-map that marks DSCP EF in the IP header, while marking CoS 5 in the Ethernet 802.1p header.
Imagine you are supposed to update a router configuration. The current configuration includes a class-map that refers to ACL 101, which has 23 ACL clauses (separate access-list commands). How could you easily create a new class map that matches the traffic denied by the ACL?
You could create a class map, with a match not access-group 101 command in it. This command matches all packets not permitted by ACL 101—in other words, packets denied by the ACL.
A router is configred to create a VPN tunnel. Explain the required steps you must take to cause a router to copy the ToS byte of the original packet into the ToS byte of the new IP header used to encapsulate the packet.
No additional overt action is required—Cisco IOS automatically copies the ToS byte into the newly-created IP header.
A router is configured to create a VPN tunnel, with unencrypted packets entering interface Fa0/0, and the encrypted packets going over a link to the internet (S0/0). Assuming as many defaults as possible were taken, could a policy map for packets entering the router’s FA0/0 interface examine the packet headers as originally created by the end user device? Why?
The packet will not have been processed by the VPN feature of the router yet, so all the original packet headers will be available for matching.
A router is configred to create a VPN tunnel, with unencrypted packets entering interface Fa0/0, and the encrypted packets going over a link to the internet (S0/0). Assuming as many defaults as possible were taken, could a policy map for packets exiting the router’s S0/0 interface examine the packet headers as originally created by the end user device? Why or why not?
The original packet headers will not be available for matching, because the router will have already encapsulated, and probably encrypted, those headers.
A router is configred to create a VPN tunnel, with unencrypted packets entering interface Fa0/0, and the encrypted packets going over a link to the Internet (S0/0). Assuming the qos pre-classify command was configured correctly, could a policy map for packets entering the router’s FA0/0 interface examine the packet headers as originally created by the end user device? Why or why not?
The packet headers will be available for matching, because the qos pre-classify command tells the router to keep a copy of the headers available for the purpose of performing QoS features.
Name the three configuration areas in which you might use the qos pre-classify command in order to enable pre-classification.
Under a tunnel interface, under a crypto map, and under a virtual-template interface.
Describe the benefits of having a single FIFO output queue.
The most basic benefit of queuing is to provide a means to hold a packet while the interface is busy. Without at least a single FIFO queue, routers would have to discard packets if the outgoing interface were busy.
Explain the effects of changing a single FIFO queue’s length to twice its original value. Include comments about how the change affects bandwidth, delay, jitter, and loss.
With a longer queue, more packets can be enqueued before the queue fills. Therefore, the tail-drop process drops packets less often. However, with more packets in the queue, the average delay increases, which also can increase jitter. There is no impact on bandwidth.
Explain the purpose of a TX Ring and TX Queue in a Cisco router.
By design, routers want to be able to begin immediately sending the next packet when the preceding packet’s last bit is sent. To do this, the interface hardware must have access to a queue structure with the next packet, and not be impeded by waiting on service from other processes. On Cisco routers, the TX Ring and TX Queue provide queue structures that are available to the interface directly, without relying on the main processor.
Explain how a long TX Ring might affect the behavior of a queuing tool.
Output queuing does not occur until the TX Ring is full. If the TX Ring is long, the Queuing tool might not be enabled. Because the TX Ring always uses FIFO logic, packets will not be reordered. With a short TX Ring, output queuing may be queuing the packets, and have an opportunity to reorder the packet exit sequence based on the queuing scheduling algorithm.
Describe the command output that identifies the length of the TX Ring or TX Queue, and whether the length was automatically lowered by IOS.
The show controllers command lists output that includes the output line that reads something like “tx_limited=0(16).” The first number is 0 or 1, with 0 meaning that the statically-configured value is being used, and the number in parenthesis representing the length of the TX Ring/TX Queue. If the first number is 1, the TX Ring/ TX Queue has been automatically shortened by the IOS as a result of having a queuing tool enabled on the interface.
Explain under what circumstances the TX Ring, interface output queues, and subinterface output queues both fill and drain, and to where they drain.
The TX Ring fills when the packets needing to exit an interface exceed the line (clock) rate of the interface. When the TX Ring fills, the interface output queues begin to fill. The subinterface output queues only fill if traffic shaping is enabled on the subinterfaces or individual VCs, and if the offered traffic on a subinterface or VC exceeds the shaped rate. The VC or subinterface queues drain into the interface queues, the interface queues into the TX Ring, and the TX Ring onto the physical interface.
Assume a queuing tool has been enabled on interface S0/0. Describe the circumstances under which the queuing tool would actually be used.
Congestion must occur on the interface first, which causes packets to be held in the TX Ring/TX Queue. When the TX Ring/TX Queue fills, IOS enables the queuing function on the interface.
Explain the circumstances under which it would be useful to enable a queuing tool on a subinterface.
Queues only form on subinterfaces when traffic shaping is enabled on the subinterface.
Describe the process and end result of the scheduling feature of Priority Queuing.
Always service higher-priority queues first; the result is great service for the High queue, with the potential for 100 percent of link bandwidth. Service degrades quickly for lower-priority queues, with possile total starvation of the lower queues.
Describe the process and end result of the scheduling feature of Custom Queuing.
Scheduler services packets from a queue until a byte count is reached; round-robins through the queues, servicing the different byte counts for each queue. The effect is to reserve a percentage of link bandwidth for each queue.
Describe how the Modified Deficit Round-Robin scheduler works, and specifically why the word “deficit” refers to part of the scheduler logic.
DRR schedules some number of bytes per pass through the queues. MDRR takes packets from the queue, which means it may take more than the allotted number of bytes; this excess is called the deficit. The deficit is subtracted from the number of bytes taken from that queue in the next round. As a result, MDRR can accurately predict the percentage bandwidth assigned to a queue.
WFQ classifies packets based on their flow. Other than a typical flow from an end user device, identify the other two types of flows recognized by WFQ.
WFQ reserves 8 flow queues for system overhead traffic. It also adds flows in conjunction with RSVP, helping to reserve bandwidth for those flows.
Characterize the effect the WFQ scheduler has on different types of flows.
Lower-volume flows get relatively better service, and higher-volume flows get worse service. Higher-precedence flows get better service than lower-precedence flows. If lower-volume flows are given higher precedence values, the bandwidth, delay, jitter, and loss characteristics improve even more.
Describe the WFQ scheduler process. Include at least the concept behind any formulas, if not the specific formula.
Each new packet is assigned a sequence number, which is based on the previous packet’s SN, the length of the new packet, and the IP precedence of the packet. The formula is as follows:
Previous SN + (weight * New packet length)
The scheduler just takes the lowest SN packet when it needs to de-queue a packet.
You previously disabled WFQ on interface S0/0. List the minimum number of commands required to enable WFQ on S0/0.
Use the fair-queue interface subcommand.
What commands list statistical information about the performance of WFQ?
The show interfaces and the show queueing fair commands list statistics about WFQ.
Define what comprises a flow in relation to WFQ.
A flow consists of all packets with the same source and destination IP address, transport layer protocol, and transport layer source and destination port. Some references also claim that WFQ includes the ToS byte in the definition of a flow.
You just bought and installed a new 3600 series router. Before adding any configuration to the router, you go ahead and plug in the new T1 Frame Relay access link to interface S0/0. List the minimum number of commands required to enable WFQ on S0/0.
No commands are required. WFQ is the default on E/1 and slower interfaces in a Cisco router.
Describe the CBWFQ scheduler process, both inside a single queue and among all queues.
The scheduler provides a guaranteed amount of bandwidth to each class. Inside a single queue, processing is FIFO, except for the class-default queue. In class-default, Flow-Based WFQ can be used, or FIFO, inside the queue.
Describe how LLQ allows for low latency while still giving good service to other queues.
LLQ is actually a variation of CBWFQ, in which the LLQ classes are always serviced first—in other words, the low-latency queues are a strict-priority queues. To prevent the low-latency queues from dominating the link, and to continue to guarantee bandwidth amounts to other queues, the LLQ classes are policed.
Compare and contrast the CBWFQ command that configures the guaranteed bandwidth for a class with the command that enables LLQ for a class.
The bandwidth command enables you to define a specific bandwidth, or a percentage bandwidth. The priority command, which enables LLQ in a class, appears to reserve an amount or percentage of bandwidth as well. However, it actually defines the policing rate, to prevent the LLQ from dominating the link. The priority command enables you to set the policing burst size as well.
Describe the CBWFQ classification options. List at least five fields that can be matched without using an ACL.
CBWFQ uses the Modular QoS CLI, and therefore can match on any fields that can be matched with other MQC tools, like CB marking. Other than referring to an ACL, CBWFQ can classify based on incoming interface, source/destination MAC, IP Precedence, IP DSCP, LAN CoS, QoS group, MPLS Experimental bits, and anything recognizable by NBAR.
Name the two CBWFQ global configuration commands that define classification options, and then the per-hop behaviors, respectively. Also list the command that enables CBWFQ on an interface.
The class-map command names a class map and places the user into class map configuration mode. Classification parameters can be entered at that point. The policy-map command names a policy and enables you to refer to class maps and then define actions. The service-policy command enables the policy map for packets either entering or exiting the interface.
All the policy maps except pmap4 would perform LLQ on voice payload. In some cases, the policy map would match more than just voice payload. Only pmap1 would match just RTP voice payload traffic.
If some other classification and marking tool were configured, and it marked all voice payload traffic as DSCP EF, pmap4 would match all voice packets in the low-latency queue.
Which of the following queuing tools can always service a particular queue first, even when other queues have packets waiting? First-In, First-Out Queuing (FIFO); Priority Queuing (PQ); Custom Queuing (CQ); Weighted Fair Queuing (WFQ); Class-Based WFQ (CBWFQ); Low Latency Queuing (LLQ).
PQ and LLQ.
Which of the following queuing tools allows for a percentage bandwidth to be assigned to each queue? First-In, First-Out Queuing (FIFO); Priority Queuing (PQ); Custom Queuing (CQ); Weighted Fair Queuing (WFQ); Class-Based WFQ (CBWFQ); Low Latency Queuing (LLQ).
CBWFQ and LLQ. CQ effectively does this as well, but you cannot specify the exact percentage.
Which queuing tools could be configured to provide the lowest possible latency for voice traffic? Of these, which does Cisco recommend as the best option for voice queuing today?
PQ and LLQ. PQ would probably not be a good option in many networks today, but it could provide the lowest possible latency for voice. Cisco currently recommends LLQ.
Which of the following queuing tools can use flow-based classification? First-In, First-Out Queuing (FIFO); Priority Queuing (PQ); Custom Queuing (CQ); Weighted Fair Queuing (WFQ); Class-Based WFQ (CBWFQ); Low Latency Queuing (LLQ).
WFQ and CBWFQ in the class-default queue.
Which of the following queuing tools uses the Modular QoS CLI? First-In, First-Out Queuing (FIFO); Priority Queuing (PQ); Custom Queuing (CQ); Weighted Fair Queuing (WFQ); Class-Based WFQ (CBWFQ); Low Latency Queuing (LLQ).
Which of the following queuing tools allows for a value to be configured, which then results in a specific number of bytes being taken from each queue during a round-robin pass through the queues? First-In, First-Out Queuing (FIFO); Priority Queuing (PQ); Custom Queuing (CQ); Weighted Fair Queuing (WFQ); Class-Based WFQ (CBWFQ); Low Latency Queuing (LLQ).
What model of Cisco router supports WFQ inside CBWFQ classes other than class-default?
7500 series routers.
Give an explanation for the following comment: “WFQ can become too fair when it has a large number of active flows”?
With many flows, WFQ will give some bandwidth to every flow. In an effort to give each flow some of the link bandwidth, WFQ may actually not give some or most of the flows enough bandwidth for them to survive.
Explain the points during the process of a single router receiving and forwarding traffic at which shaping and policing can be enabled on a router.
Shaping can be enabled for packets exiting an interface, subinterface, or individual VC. Policing can be performed both on packets entering an interface or exiting an interface.
Compare and contrast the actions that shaping and policing take when a packet exceeds a traffic contract.
Shaping queues packets when the shaping rate is exceeded. Policing either discards the packet, just transmits the packet, or it re-marks a QoS field before transmitting the packet.
Compare and contrast the effects that shaping and policing have on bandwidth, delay, jitter, and loss.
Shaping places packets into queues when the actual traffic rate exceeds the traffic contract, which causes more delay, and more jitter. Policing when making a simple decision to either discard or forward each packet causes more packet loss, but less delay and jitter for the packets that do make it through the network. Shaping and Policing both limit the amount of bandwidth allowed for a particuilar class of traffic.
Describe the typical locations to enable shaping and policing in an internetwork.
Shaping is typically performed before sending packets into a network that is under some other administrative control. For instance, shaping is typically performed before sending packets from an enterprise into a service provider’s Frame Relay network. Policing, although supported as both an input and output function, is typically performed at ingress points, once again at the edge between two administrative domains.
Describe the reasons behind egress blocking in a Frame Relay network with a T1 access link at the main site, 128-kbps access links at each of 20 remote sites, with 64-kbps CIR VCs from the main site to each remote site.
Egress blocking can occur for frames leaving the Frame Relay network going to the main site, because the sum of the access rates of the 20 sites exceeds the access rate at the main site. Egress blocking occurs for packets leaving the Frame Relay network going to an individual remote site, because the access rate at the main site exceeds the access rate at each remote site.
If a router has CB Shaping configured, with a shaping rate of 256 kbps, and a Bc of 16,000 bits, what Tc value does the shaping tool use?
Because Tc = Bc/CIR, the answer is 16,000/256,000, or 62.5 ms.
If a router has CB Shaping configured, with a shaping rate of 512 kbps, and a Be of 16,000 bits, what Tc value does the shaping tool use?
Tc is not calculated based on Be. However, at rates higher than 320 kbps, CB Shaping uses a set 25 ms Tc.
Time interval, measured in milliseconds, over which the committed burst (Bc) can be sent.
committed burst size, measured in bits. This is the amount of traffic that can be sent during every interval Tc. Typically also defined in the traffic contract.
Excess burst size, in bits. This is the number of bits beyond Bc that can be sent in the first Tc after a period of inactivity.
committed information rate, in bits per second, defines the amount of bandwidth that the provider has agree to provide as defined in the traffic contract.
Describe the concept of traffic-shaping adaption.
Adaption causes the shaper to reduce the shaping rate during congestion. Shapingreacts to frames with the BECN bit set, or to Foresight congestion messages.
Describe the difference between interface output queues and shaping queues, and explain where the queues could exist on a router with 1 physical interface and 20 subinterfaces.
Output queues exist on the physical interface, and can be controlled with queuing tools such as CBWFQ and WFQ. Shaping queues exist when traffic shaping is enabled; the shaping queue is associated with the particular instance of shaping. If shaping has been enabled on 20 subinterfaces on a single physical interface, for instance, 20 sets of shaping queues exist, all feeding into the single set of physical interface software queues.
How many token buckets are used by the CB Shaping internal processes with Be = 0? How big are the buckets?
Only 1 token bucket is used. The size is equal to Bc bits.
How many token buckets are used by the CB Shaping internal processes with Be = 8000? How big are the buckets?
Only 1 token bucket is used, but the size is Bc + Be bits.
How many token buckets are used by the CB Policing internal processes with Be = 0? How big are the buckets?
Only 1 token bucket is used. The size is equal to Bc bytes.
How many token buckets are used by CB Policing internal processes, configured for single-rate policing, with Be = 8000? How big are the buckets?
Two token buckets are used, with Bc bytes in one bucket, and Be bytes in the other
Imagine a CB Shaping configuration with a rate of 128000, Bc = 8000, and Be = 16000. What is the Tc value, and how many tokens are refilled into the first bucket during each Tc?
Tc = Bc/CIR, or in this case, 8000/128000, or 62.5ms. Each Tc (62.5ms), 8000 tokens (Bc tokens) are refilled into the bucket. CB Shaping only uses one bucket. (Author’s note: IOS actually rounds these numbers so there are no fractions in use.)
Imagine a CB Shaping configuration with a rate of 128000, Bc = 8000, and Be = 16000. At the beginning of the next time interval, the token bucket is full. If the physical clock rate of the interface on which shaping is enabled is 256 kbps, describe how much traffic that will be sent in this next Tc, and why.
At a physical link speed of 256 kbps, with a calculated Tc of 62.5 ms (see previous question’s answer for that math), the maximum number of bits that can be sent in 62.5 seconds at that rate is 256000 * .0625 = 16000. The bucket has a size of Bc + Be, or 24,000; because there are 24,000 bits worth of tokens are in the bucket at the beginning of the interval, all packets totalling 16,000 bits can be sent in this first interval, with the bucket containing 8000 more tokens.
If a policer is called a “two color” policer, what does that mean?
It means that the policer designates each policed packet as either conforming to the traffic contract, or exceeding the contract. The numer of colors is the number of categories, in terms of meeting the traffic contract, into which the policer can place a packet.
If a policer is called a “three color” policer, what does that mean?
It means the same general thing as a “two color” policer, but with three categories – conform, exceed, and violate.
With CB Policing, how are tokens refilled into the bucket associated with the CIR policing rate?
Unlike CB Shaping, CB policing replenishes tokens in the bucket in response to policing a packet, as opposed to every Tc seconds. Every time a packet is policed, CB policing puts some tokens back into the bucket. The number of tokens placed into a bucket is calculated as follows: (current arrival time - previous arrival time * cir) / 8
With a dual-rate policer, how are tokens refilled into the token bucket associated with PIR?
It fills the PIR bucket in the same general method as filling the CIR bucket, but with the formula using the PIR,
With a single-rate policer, with Be > 0, how are tokens refilled into the excess token bucket?
The tokens are filled into the first token bucket. Any that spill due to that bucket already being full of tokens spill into the excess token bucket.
With a single-rate policer, with Be = 0, what must be true for the policer to decide that a packet exceeds the traffic contract?
If there are fewer tokens in the single token bucket than the number of bytes in the packet, the packet is considered to exceed the contract.
With a single-rate policer, with Be > 0, what must be true for the policer to decide that a packet exceeds the traffic contract?
If there are fewer tokens in the first token bucket than the number of bytes in the packet, but at least that many tokens in the second bucket, the packet is considered to exceed the contract.
With a single-rate policer, with Be > 0 what must be true for the policer to decide that a packet violates the traffic contract?
If there are fewer tokens in the first token bucket than the number of bytes in the packet, plus fewer than that many tokens in the second bucket, the packet is considered to violate the contract.
With a single-rate policer, regardless of Be setting, what must be true for the policer to decide that a packet conforms to the traffic contract?
If there are at least as many tokens in the first token bucket than the number of bytes in the packet, the packet is considered to conform to the contract.
For policing configurations that use two buckets, a packet is classified as conforming, exceeding, or violating the traffic contract. When processing a new packet, in which of these three cases does the policer then also remove or spend the tokens?
When the packet either conforms to or exceeds the traffic contract. In order to conform or exceed, one or the other bucket must have had enough tokens to allow the policer to consider the packet either as conforming or exceeding. For packets that violate, the buckets are not decremented.
Comparing the logic used for a single-rate and dual-rate policer, when both use two token buckets, their logic differs slightly in terms of how the tokens are removed from the buckets when policing a packet. Explain that difference.
For a single-rate, two bucket policer, for packets that conform to the contract, the policer removes tokens from the first bucket only. With a dual-rate policer, for packets that conform to the contract, it removes tokens from both buckets.
Comparing the logic used for a single-rate and dual-rate policer, when both use two token buckets, their logic differs slightly in terms of how the tokens are added to the buckets before policing a newly-arrived packet. Explain that difference.
For a single-rate, two bucket policer, Bc bytes of tokens are added to the first bucket; spillage falls into the second bucket; and any spillage from the second bucket is wasted. For a dual-rate policer with two buckets, each pucket is replenished directly, based on the CIR and PIR, respectively. Tokens spilled from either bucket are wasted.
Along with the class-map, policy-map, and service-policy commands, CB shaping requires one specific command that actually sets values used for the shaping function. List the command, with the correct syntax, that sets a shaped rate of 128 kbps, a Bc of 8000, and a Be of 8000, when using CB shaping. Do not assume any defaults; explicitly set the values in the command.
shape average 128000 8000 8000
Explain the context inside the configuration mode under which the service-policy command can be used to enable LLQ on a CB shaping queue. (“Context” means what part of configuration mode—for instance, global-configuration mode, interface configuration mode, and so on.)
CB shaping requires a policy map, with class commands inside the policy map. Inside class configuration mode inside the CB shaping policy map, the service-policy command can refer to another policy map, which could enable LLQ for the class.
CB shaping has been configured under subinterface s0/0.1. What show command lists statistics for CB shaping behavior just for that subinterface?
show policy-map interface s0/0.1
Which of the traffic-shaping tools can be enabled on each VC on a Frame Relay multipoint subinterface?
At what rate would CB Shaping actually shape traffic when using the command shape peak 64000 8000 16000?
The formula to figure out the peak rate is Actual_rate = configured_rate (1 + Be/ Bc). In this case, the formula is 64000 (1 + 16000/8000), or 192,000 bits/second.
CB Policing has been configured under subinterface s0/0.1. What show command would list statistics for CB Policing behavior just for that subinterface?
show policy-map interface s0/0.1
List the command, with the correct syntax, that sets a Policed rate of 512 kbps, a Bc of 1 second’s worth of traffic, and a Be of an additional .5 seconds worth of traffic, when using CB Policer. Do not assume any defaults; explicitly set the values in the command. You can choose any other settings needed for the command.
police 512000 64000 32000 conform-action transmit exceed-action drop violate-action drop
Explain the concept behind re-marking policed packets versus discarding the packets.
By re-marking the packets, you can increase the packet’s likelihood of being dropped later. For instance, WRED reacts to the precedence or DSCP value, discarding certain marked values more aggressively. By re-marking, if no congestion occurs, the packet may still get through the network. If congestion does occur, the packet that the policer marked down has a greater chance of being dropped.
Describe the function of the congestion window in TCP, and how it is changed as a result of packet loss.
The TCP congestion window, or CWND, is one of two windowing mechanisms that limit TCP senders. CWND can be split in half as a result of packet loss, slowing the sending rate. CWND can also be slammed shut to the size of a single segment in some cases.
Identify the two TCP windowing mechanisms, and describe when each is used.
The TCP congestion window, or CWND, and the TCP receiver window, are the two windowing mechanisms. The lower of the two values is used at all times.
Describe the process of TCP slow start, and when it occurs.
TCP slow start governs the growth of the TCP congestion window after the window has been lowered in reaction to a packet drop. Slow start increases the window by one segment size for each positively acknowledged packet received.
Describe the meaning of the term “global synchronization,” and discuss what causes it.
Global synchronization describes a condition in which many TCP connections have their congestion windows lowered due to unacknowledged or lost segments at around the same instant in time. The connections all grow CWND at about the same rate, re-creating the same congestion levels again, causing more drops, which in turn reduces again the TCP congestion windows. Global synchronization is caused by a large number of packet drops in a very short period, typically the result of tail drops.
Define the meaning of the term “tail drop.”
When a queue fills, and a new packet must be placed into the queue, the packet is dropped. Because the packet would be placed into the end, or tail, of the queue, it is called tail drop.
Define the meaning of the term “TCP starvation.”
When packets are dropped, TCP connections slow down, but UDP flows do not slow down. UDP packets can consume a disproportionate amount of queue space as a result, which could get to the point that the TCP connections simply get little or no queue space; this is called TCP starvation.
Does RED compare the actual queue depth or the average queue depth to queue thresholds when deciding whether it should discard a packet? Why this one, and not the other?
RED uses average queue depth. By using the average, rather than the actual queue depth, RED behaves more consistently, rather than more erratically, which helps prevent synchronization of TCP flows.
Describe how RED uses actual queue depth to calculate average queue depth. Do not list the formula, but just describe the general idea.
RED calculates the average by adjusting the previously calculated average a small amount based on the current actual queue depth. By default, the current queue depth is weighted at about .2 percent in the formula.
Assume the RED minimum threshold is 20, the maximum threshold is 40, and the mark probability denominator is 10. What must be true for RED to discard all new packets?
The average queue depth must be above 40.
Assume the RED minimum threshold is 20, the maximum threshold is 40, and the mark probability denominator is 10. What must be true for RED to discard 5 percent of all new packets?
The average queue depth must be at 30. Because the discard percentage grows linearly from 0 percent to 10 percent (in this case), between average queue depth of 20 through 40, average queue depth of 30 would mean that the discard percentage had grown to 5 percent.
Define how RED uses the mark probability denominator. Give one example.
RED calculates the discard percentage based on the formula 1/MPD. For instance, with an MPD of 20, the discard percentage is 1/20, or 5 percent.
Define the term “exponential weighting constant.” If the value is lowered compared to the default setting of 9, how does RED behave differently?
The exponential weighting constant defines how quickly the average queue depth changes, by determining how much the actual queue depth affects the rolling average queue depth. If EWC is lowered, the average changes more quickly, because the formula weights the current actual queue depth more than before. Therefore, a larger constant provides more handling of bursty traffic, but too large and congestion avoidance will be ineffective.
Define the term “WRED Profile.”
A WRED profile is a collection of WRED parameters applied to a single IP Precedence or DSCP value. The parameters include the minimum threshold, the maximum threshold, and the Mark Probability Denominator (MPD).
Explain how you can tune how fast or slow that WRED changes the calculated average queue depth over time.
WRED calculates a new average based on the old average and the current queue depth. You can tell WRED to count the current queue depth as a larger or smaller part of the calculation by tuning the exponential weighting constant. The formula is:
New average = (Old_average * (1 – 2–n)) + (Current_Q_depth * 2–n)
Where “n” is the exponential weighting constant.
Spell out the words represented by the initials RED, WRED, and FRED.
Random Early Detection (RED), Weighted Random Early Detection (WRED), Flow-Based Weighted Random Early Detection (FRED).
List the three WRED terms that name the separate states in which WRED discards no packets, a percentage of packets, and all packets.
No Discard, Random Discard, and Full Discard, respectively.
List the queuing tools that can be concurrently supported on an interface when WRED has been enabled directly on a serial interface, assuming no retrictions on the particular model of router.
FIFO Queuing only.
Identify the most important difference between RED operation and WRED operation.
WRED weights its discard decisions based on precedence or DSCP, whereas RED ignores precedence and DSCP.
Describe how WRED “weights” packets.
WRED weights packets based on precedence or DSCP by assigning different minimum threshold, maximum threshold, and mark probability denominator values for each precedence or DSCP.
List the queuing tools that can enable WRED for use with some or all of their queues, effectively enabling WRED concurrently with the queuing tool, assuming no retrictions on the particular model of router.
CBWFQ and LLQ.
What command enables you to look at WRED drop statistics when WRED is configured inside an MQC class?
show policy-map interface
Taking as many defaults as possible, list the configuration commands needed to configure precedence-based WRED on interface S1/1.
interface serial 1/1
Taking as many defaults as possible, list the configuration commands needed to configure DSCP-based WRED on interface S1/1.
interface serial 1/1
aking as many defaults as possible, list the configuration commands needed to configure DSCP-based WRED inside class class1, inside policy map my-policy. (You can assume that the CBWFQ configuration has already been completed, and you just entered global configuration mode. Assume that you need just to enable WRED in class class1.)
List the command needed to set the minimum threshold to 25, the maximum threshold to 50, and the mark probability denominator to 4, for precedence 2.
random-detect precedence 2 25 50 4
What show command lists detailed statistics about random drops on interface S1/1?
show queueing interface s1/1
For a single WRED profile, WRED can be either dropping no packets, randomly choosing packets to discard, or dropping all packets. For which of these three states does ECN impact WRED’s discard actions? How does it change what WRED does to the packets?
For Random Discard only. WRED forwards the packets instead of discarding them, but only after setting the ECN bits to “11”.
Identify the bits in the IP header used with ECN, by name and location.
The low-order 2 bits of the DSCP byte are called the ECN field. The first bit is called the ECN Capable Transport (ECT) bit, and the second one is the Congestion Experienced (CE) bit.
Imagine a router on which WRED and ECN are enabled, and WRED decides to randomly discard a packet. What must be true in order for WRED to discard the packet, instead of using ECN logic to mark and forward the packet? Explain the role of any other devices besides the router.
With ECN enabled, it would set the ECN bits to “11”, unless the ECN field was set to 00. An ECN field of 00 means that the sender did not support ECN for that TCP connection.
Imagine a router on which WRED and ECN are enabled, and WRED decides to randomly discard a packet. What must be true in order for WRED to use ECN logic to mark and forward the packet, instead of discarding the packet? Explain the role of any other devices besides the router.
The ECN field must be set to something besides 00. The sender of the packet would choose to set ECN to one of those two values if it did support ECN for that TCP connection.
Imagine a policy map with WRED already configured for class class-web. What additional command is required to also enable ECN for the packets in that class?
The random-detect ecn command.
Describe what is compressed, and what is not compressed, when using payload compression. Be as specific as possible regarding headers and data.
Payload compression does not compress the data-link header and trailer, but it does compress all the higher-layer headers and data between the two. Specifically, the IP, TCP, UDP, RTP headers as appropriate, and the user data, are compressed.
Describe what is compressed, and what is not compressed, when using TCP header compression. Be as specific as possible regarding headers and data.
IP packets that also have TCP headers are compressed. The compression algorithm does not compress the data link header or trailer. It does compress both the IP and TCP headers. It does not compress any user data that follows the TCP header.
Describe what is compressed, and what is not compressed, when using RTP header compression. Be as specific as possible regarding headers and data.
IP packets that also have RTP headers are compressed. The compression algorithm does not compress the data-link header or trailer. It does compress the IP, UDP, and RTP headers. It does not compress any user data that follows the RTP header.
Suppose a packet is sent across a network with no compression. Later, a packet of the exact same size and contents crosses the network, but payload compression is used on the one serial link in the network. Describe the difference in bandwidth and delay in the network between these two packets.
The packet experiences longer processing delay as a result of the compression algorithm. However, the packet requires less time to be serialized onto the link, resulting in less serialization delay. Overall queuing delay should be decreased, because the shorter compressed packets take less time to serialize, thereby causing packets to exit the queues more quickly. The overall reduction in queue sizes can reduce delay and jitter.
How much bandwidth should a G.729 call require over Frame Relay, and how much should be required with cRTP?
A single G.729 call requires 28 kbps over Frame Relay, but it only needs 12.8 kbps using cRTP.
When TCP header compression is used, what is the range of sizes of the part of the frame that can be compressed, and what is the range of sizes for this field of the frame after compression?
TCP header compression compresses the 20-byte IP header and 20-byte TCP header, with the combined field size of 40 bytes. The compressed field will be between 3 and 5 bytes.
When RTP header compression is used, what is the range of sizes of the part of the frame that can be compressed, and what is the range of sizes for this field of the frame after compression?
RTP header compression compresses the 20-byte IP header, 8-byte UDP header, and 12-byte RTP header, with the combined field size of 40 bytes. The compressed field will be between 2 and 4 bytes.
List the words represented by the abbreviation LFI.
Link fragmentation and interleaving.
Describe the main motivation for LFI tools in relation to the support of data, voice, and video traffic.
LFI tools interleave some packets between the fragments of other packets. Voice and two-way video traffic are particularly sensitive to delay. LFI reduces the delay for voice and video packets by interleaving voice and video packets between fragments of the data packets
To achieve a 20-ms serialization delay on a 128-kbps link, how long can the fragments be?
The formula is max-delay * bandwidth, which is .02 * 128,000 = 2560 bits, or 320 bytes.
To achieve a 10-ms serialization delay on a 64-kbps link, how long can the fragments be?
The formula is max-delay * bandwidth, which is .01 * 64,000 = 640 bits, or 80 bytes
To achieve a 10-ms serialization delay on a 56-kbps link, how long can the fragments be?
The formula is max-delay * bandwidth, which is .01 * 56,000 = 560 bits, or 70 bytes.
To achieve a 30-ms serialization delay on a 128-kbps link, how long can the fragments be?
The formula is max-delay * bandwidth, which is .03 * 128,000 = 3840 bits, or 480 bytes.
Suppose that a 1500-byte packet exits a 56-kbps serial interface, and LFI is not used. How long is the serialization delay?
The formula is packet length/link speed, which is 1500 * 8/56,000, or .214 seconds. The units used in the formula are bits, bits per second, and seconds, respectively.
Which queuing tools can you enable directly on a serial interface when using multilink Point-to-Point Protocol with link fragmentation and interleaving (MLP LFI), in order to interleave packets?
PQ, LLQ and IP RTP Priority. CBWFQ can be configured, but because it does not have a PQ-like function, it does not interleave packets.
Which queuing tools can you enable with FRTS in order to actually interleave the traffic?
LLQ and IP RTP Priority actually interleave packets.
Explain the scheduling logic used by MLP LFI to determine which packets can be interleaved in front of fragments of other packets.
MLP LFI does not define scheduling logic. Instead, it relies on the scheduler of the queuing tool enabled on the interface to decide which packets to send next. If LLQ were used, for instance, packets from the low-latency queue would be interleaved in front of packets from other queues.
Suppose a 1500-byte packet arrives and needs to be sent over an MLP bundle that has two active links. LFI has not been configured. Which link does the packet flow across to achieve MLP load balancing?
MLP fragments the packet into two equal-sized fragments, and sends one over one link, and one over the other.
What command can you use to determine the fragment size used for MLP LFI? What is the only parameter of the command?
The ppp multilink fragment-delay command sets the maximum serialization delay in milliseconds. IOS calculates the fragment size using the formula max-delay * bandwidth.
What command enables the interleaving feature of MLP LFI?
The ppp multilink interleave command.
What commands list counters for the number of interleaved packets using MLP LFI?
The show queue and show interfaces commands.
What other QoS feature for Frame Relay must you enable when you also configure FRF.12?
Frame Relay Traffic Shaping (FRTS).
What command enables FRF and sets the fragment size?
The frame-relay fragment fragment_size command.
What command lists counters for the numbers of packets and bytes that were fragmented and unfragmented by FRF.12?
The show frame-relay fragment interface subcommand.
What command lists counters for the numbers of packets and bytes that would have been sent if FRF.12 fragmentation had not been performed?
The show frame-relay fragment interface subcommand.
What command lists counters for the number of packets that end up in the High and Normal Dual-FIFO siftware queues, when using FRF.12?
The show queueing interface x/y command.
Why do you need QoS in the LAN?
Buffer Management as well as Classification and Marking as close to the sources as possible are the reasons you need QoS is the LAN.
What is buffer overflow and when does it occur?
The term buffer overflow indicates that a buffer on and interface has received more traffic that it can transmit. It occurs when an interface is oversubscribed.
What IOS types are available for the Catalyst 2950 and which one is preferred for QoS?
Standard and Enhanced. Enhanced is preferred for QoS
You have a Catalyst 2950 running the standard IOS image and need to migrate to a Catalyst 2950 running the enhanced IOS image. What are your options to migrate to a 2950 running the enhanced IOS image?
The Catalyst 2950 IOS version is hardware dependent. This means that you cannot upgrade a standard IOS image to an enhanced IOS image. You must order a new switch.
What methods can a Catalyst 2950 current use to classify traffic?
Trust, Port-based, CoS, DSCP and access lists
What map needs to be changed on the Catalyst 2950 to reflect the current markings of Cisco IP Phones?
What command is used to verify the CoS-to-DSCP map?
mls qos map cos-dscp
By default a Catalyst 2950 will map voice-media traffic and voice-signaling traffic to which DSCP values?
Voice-Media CS5 (decimal 40) and voice-signaling CS3 (decimal 24)
To keep the DSCP values of the voice-media traffic and the voice-signaling traffic consistent between the IP Phones and the Catalyst 2950, which DSCP values need to be configured on the Catalyst 2950?
Voice-Media EF (decimal 46) and voice-signaling AF31 (decimal 26)
By default, what values will the Catalyst 2950 trust?
Name two of the three markings and / or devices that the Catalyst 2950 can use to extend a trust boundary.
Trust of CoS, trust of DSCP or trust of Cisco IP Phone
Where is the trust command configured on a Catalyst 2950?
On an Ethernet interface
What does the switchport priority extend cos 0 command do?
This command causes any CoS value received from a PC attached to an IP phone to be overwritten with a CoS value of 0.
What does the mls qos trust cos pass-through dscp command do?
This command forces the Catalyst 2950 to trust the received CoS value and pass through the original DSCP value on a received packet rather than use the CoS to DSCP map to determine the DSCP value.
What command is used to enable the trust of a Cisco IP Phone on a Catalyst 2950?
mls qos trust device cisco-phone
How does the Catalyst 2950 determine if a Cisco IP Phone is connected to a port configured to trust device cisco-phone?
The switch will used CDP version 2 to discover the phone and apply the trust.
What happens if an interface is configured to trust device cisco-phone and a PC is connected to the interface?
If CDP version 2 does not detect a Cisco IP Phone, the interface will remain untrusted.
Which command is used to set the default CoS value of an interface to 2 for all untagged frames received?
mls qos cos 2
What does the command mls qos cos override do?
This command overrides the CoS values of packets received on an interface causing all tagged and untagged frames to use the default CoS value configured on that interface.
What command is used to view the current trust state of interface Fast Ethernet 0/1?
show mls qos interface fastethernet 0/1
A router that is not configured for 802.1Q is connected to interface 0/1 of the Catalyst 2950. The router interface forwards a voice-media packet to the switch. What CoS value will the switch receive?
None. If there is no 802.1Q trunk there is no CoS value.
A router that is not configured for 802.1Q is connected to interface 0/1 of the Catalyst 2950. The switch is configured to trust CoS on this interface. The router interface forwards a voice-media packet to the switch. What DSPC value will the switch use for this packet by default?
DSCP 0. Because the switch is configured to trust CoS and there is no 802.1Q trunk, the switch will use the default CoS for that interface. By default, this value is CoS 0 which will be mapped to DSCP 0. To allow the DSCP value to pass, enable trust of DSCP on interface 0/1.
What can MQC be used for on the Catalyst 2950?
Classification, Policing, and Marking.
What is a class map used for?
To identify or classify traffic.
What is a policy map used for?
To apply actions to the classified traffic.
How is a policy map applied to an interface?
Using the service-policy command
In which direction can a service policy be applied on a Catalyst 2950?
How many ingress queues are present on each Ethernet interface of the Catalyst 2950?
How many egress queues are present on each Ethernet interface of the Catalyst 2950?4
What scheduling methods are supported on the Catalyst 2950?
Strict priority scheduling, WRR scheduling and strict priority scheduling with WRR scheduling, also known as strict priority queuing.
What mark does the Catalyst 2950 use to determine the proper queue for a packet?
Which scheduling method is configured by default on a Catalyst 2950?
Strict priority scheduling
Which scheduling method is recommended for networks that transport IP Telephony traffic?
Strict priority queuing
How does strict priority scheduling work?
Each queue is services by order of priority with 4 being the highest and 1 being the lowest. A packet in a higher queue will always be transmitted before a packet in a lower queue.
What are the advantages and disadvantages of using strict priority scheduling?
Strict priority scheduling offers the ability to guarantee minimum delay to applications in the highest priority queue, but lower priority queues have the potential for queue starvation.
How does Weighted Round Robin scheduling work?
Each queue is allocated a minimum amount of bandwidth to be serviced in a weighted round robin fashion.
What are the advantages and disadvantages of WRR scheduling?
WRR scheduling offers each queue the capability to transmit packets eliminating queue starvation, but lacks a strict priority queue, which may cause voice quality issues with voice-media traffic due to added delay.
How does strict priority queuing work?
Strict priority queuing combines the benefits of strict priority scheduling and WRR scheduling by offering a single strict priority queue and 3 WRR serviced queues. Packets are transmitted from the strict priority queue if present. If no packets are in the priority queue, the WRR scheduler will divide that available bandwidth between the remaining 3 queues.
What queue is configured for the strict priority queue when using strict priority queuing?
What command is used to assign CoS value of 5 to queue 4?
wrr-queue cos-map 4 5
What command is used to assign WRR scheduling with queue 1 servicing 5 packets, queue 2 servicing 10 packets, queue 3 servicing 25 packets and queue 4 servicing 50 packets?
wrr-queue bandwidth 5 10 25 50
What command is used to enable the strict priority queue when using strict priority queuing?
wrr-queue bandwidth x x x 0 where x is the value of the WRR scheduler and the 0 indicates that queue 4 will be the priority queue.
What command is used to verify that queue 4 has been configured for the priority queue?
show wrr-queue bandwidth
What is a policer?
A policer defines the amount of acceptable traffic in a class and specifies what to do with traffic that exceeded the acceptable rate.
What is out of profile or nonconforming traffic?
Received traffic that exceeds the rate specified by the policer.
What can a policer do with nonconforming traffic on a Catalyst 2950?
Drop or remark the DSCP of the traffic.
How many policers can be applied to a single 10/100 Ethernet interface on a Catalyst 2950? How about a Gigabit interface on a Catalyst 2950?
What is the maximum burst size that a policer can use for a 10/100 interface on a Catalyst 2950?
What is the maximum burst size that a policer can use for a Gigabit interface interface on a Catalyst 2950?
What is the minimum traffic rate that a policer used for conforming traffic for a 10/100 interface on a Catalyst 2950?
1000000 (1 Meg).
What is the minimum traffic rate that a policer used for conforming traffic for a Gigabit interface on a Catalyst 2950?
8000000 (8 Meg).
What commands can be used to enable AutoQoS on the Catalyst 2950?
auto qos voip trust and auto qos voip cisco-phone
Where are the auto qos voip trust and auto qos voip cisco-phone commands applied?
On an Ethernet interface.
What does the command auto qos voip trust do?
It enables trust of CoS for the interface it was applied to, changes the CoS to DSCP map and configures strict priority queuing.
What doe the command auto qos voip cisco-phone do?
It enables trust of a Cisco IP Phone for the interface it was applied to, changes the CoS to DSCP map and configures strict priority queuing.
When is configuring trust manually preferred over Auto QoS?
When trust is based upon DSCP.
When Auto QoS is enabled, what CoS values are mapped to queue 1? What percentage of the WRR scheduler does queue 1 receive?
CoS values mapped to queue 1 include 0, 1, 2 and 4. Percentage of the WRR scheduler is 20 percent.
When Auto QoS is enabled, what CoS values are mapped to queue 2? What percentage of the WRR scheduler does queue 2 receive?
No CoS values are mapped to queue 2; however, the percentage of the WRR scheduler is 1 percent. Because there will never be traffic in this queue, it will never use the 1 percent.
When Auto QoS is enabled, what CoS values are mapped to queue 3? What percentage of the WRR scheduler does queue 3 receive?
CoS values mapped to queue 3 include 3, 6 and 7. Percentage of the WRR scheduler is 80 percent.
When Auto QoS is enabled, what CoS values are mapped to queue 4? What percentage of the WRR scheduler does queue 4 receive?
CoS 5 is the only CoS value mapped to queue 4. Queue 4 becomes the priority queue and will not use the WRR scheduler.
What command can you use to verify Auto QoS?
show auto qos
What is end-to-end QoS?
End-to-end QoS is a means of offering the same treatment to a packet on each node it traverses.
hat do service providers use to contractually guarantee specific delay, jitter, and packet loss to their customers?
A service-level agreement (SLA)
When would an SLA be offered on a per PVC basis?
When the physical interface services multiple customers.
In a Cisco IP Telephony environment, which codec is best suited for the LAN?
In a Cisco IP Telephony environment, which codec is best suited for the WAN?
By default, how many milliseconds of speech are placed into every packet in a voice media stream?
How much bandwidth is required to transport a G.729 VoIP call @ 50 packets per second over a Frame Relay circuit? Assume that cRTP and VAD are not enabled.
Enabling cRTP can reduce the overhead of the IP/UDP/RTP header from 40 bytes to how many bytes?
2 or 4
Cisco DSPs have the capability to compensate for how many milliseconds of lost speech?
A Cisco IP Phone can compensate for how much jitter?
Data applications should be separated into no more than how many distinct classes?
Four or five
In a converged network supporting voice, interactive video, streaming video, mission-critical data, transactional data, and best-effort traffic, which class or classes should use the low latency queue?
Voice and interactive video
What tool is recommended to discover applications on your network?
What DSCP value do Cisco IP Phones currently mark voice media traffic with?
What CoS value do Cisco IP Phones currently mark voice media traffic with?
What DSCP value do Cisco IP Phones currently mark voice signaling traffic with?
What CoS value do Cisco IP Phones currently mark voice media signaling with?
What is the recommended CoS value for IP routing protocols?
What is the recommended DSCP value for IP routing protocols?
What is the recommended DSCP class for transactional applications?
The AF2x class
Which type of queue should voice signaling traffic use in an enterprise network?
Which command is used to allocate 100 percent of the link to MQC?
What value should the Tc be configured for on a Frame Relay circuit that will transport voice media traffic?
What is needed in a campus LAN before QoS is implemented?
A hierarchical and redundant campus design
Where in the campus LAN is policing typically configured?
The access and distribution layers
What is the recommended CoS value for IP routing protocols?
CoS value of 6
What QoS mechanism should never be implemented in the core of the campus LAN?
Classification and marking
If Cisco IP Phones have the capability to mark voice media and voice signaling traffic, why is it necessary to use access lists to classify these types of traffic?
Cisco Softphone and certain CTI applications might not have the capability to mark this traffic, yet they have the same QoS requirements.
When is a CoS-to-DSCP map used?
When a switch receives a trusted frame with a CoS value, it will assign a DSCP value to the packet based upon this map.
When should the CoS-to-DSCP map be modified?
When the CoS value does not match the desired DSCP value by default. For example, a received packet with a CoS value of 5 will be marked with a DSCP value of decimal 40, which is not the same as DSCP EF (decimal 46). The CoS-to-DSCP map must be modified to map a CoS value of 5 to a DSCP value of EF (decimal 46).
Which two queuing methods should be used in the service provider’s backbone network?
LLQ/CBWFQ and MDRR
Define IP Precedence
A 3-bit field in the first 3 bits of the ToS byte in the IP header, used for QoS marking.
Define ToS byte
A 1-byte field in the IP header, originally defined by RFC 791 for QoS marking purposes.
Define Differentiated Services
A set of QoS RFCs that redefines the IP header’s ToS byte, and suggests specific settings of the DSCP field and the implied QoS actions based on those settings.
Define DS Field
The second byte of the IP header, formerly known as the ToS byte and redefined by DiffServ.
Define Per-Hop Behavior
The second byte of the IP header, formerly known as the ToS byte and redefined by DiffServ.
Define Assured Forwarding
A set of DiffServ PHBs that defines 12 DSCP values, with four queuing classes and three drop probabilities within each queuing class.
Define Expedited Forwarding
A DiffServ PHB, based on DSCP EF (decimal 46), that provides low-latency queuing behavior as well as policing protection to prevent EF traffic from starving queues for other types of traffic.
Define Class Selector
A DiffServ PHB that defines eight values that provide backward compatibility with IP Precedence.
Define Class of Service
A 3-bit field in an ISL header used for marking frames. Also, used generically to refer to either the ISL CoS field or the 802.1Q User Priority field.
Define Differentiated Services Code Point
The first 6 bits of the DS field, used for QoS marking.
Define User Priority
A 3-bit field in an 802.1Q header used for marking frames.
Define Discard Eligible
A bit in the Frame Relay header that, when set to 1, means that if a device needs to discard frames, it should discard the frames with DE 1 first.
Define Cell Loss Priority
A bit in the ATM cell header that, when set to 1, means that if a device needs to discard frames, it should discard the frames with DE 1 first.
Define MPLS Experimental Bits
A 3-bit field in an MPLS header used for marking frames.
Define Class map
A term referring to the MQC class-map command and its related subcommands, which are used for classifying packets.
Define Policy map
A term referring to the MQC policy-map command and its related subcommands, which are used to apply QoS actions to classes of packets.
Define Service policy
A term referring to the MQC service-policy command, which is used to enable a policy map on an interface.
Define Modular QoS CLI
The common set of IOS configuration commands that is used with each QoS feature whose name begins with “Class-Based.”
Define Class-Based Marking
An MQC-based feature of IOS that is used to classify and mark packets for QoS purposes.
Define Network Based Application Recognition
A Cisco IOS feature that performs deep packet inspection to classify packets based on application layer information.
Define Qos Preclassification
A process used in routers that are encrypting traffic to permit egress QoS actions to be taken on traffic that is being encrypted on that router. QoS pre-classification keeps a copy of each packet to be encrypted in memory long enough to take the appropriate egress QoS actions on that traffic as it leaves that router, because the encrypted traffic cannot be inspected for QoS actions.
AutoQoS is a macro that creates and applies quality of service configurations based on Cisco best-practice recommendations.
Define class-based weighted fair queuing
A Cisco IOS queuing tool that uses MQC configuration commands and reserves a minimum bandwidth for each queue.
Define low-latency queuing
A Cisco IOS queuing tool that uses MQC configuration commands, reserves a minimum bandwidth for some queues, provides high-priority scheduling for some queues, and polices those queues to prevent starvation of lower-priority queues during interface congestion.
Define weighted round-robin
A queuing scheduler concept, much like CQ’s scheduler, in which queues are given some service in sequence. This term is often used with queuing in Cisco LAN switches.
Define modified deficit round-robin
A Cisco 12000 series router feature that combines the key features of LLQ and CQ to provide similar congestion-management features.
Define shaped round-robin
A packet-scheduling algorithm used in Cisco switches that provides similar behavior to CBWFQ in shared mode and polices in shaped mode.
Define shared mode
The operating mode of shaped round-robin that provides behavior like CBWFQ with bandwidth allocated between different traffic classes by a relative amount rather than absolute percentage of the available bandwidth.
Define shaped mode
The operating mode of shaped round-robin that provides a low-latency queue with policing.
A method that creates three thresholds per egress queue in the Cisco 3560 switch. Traffic is divided into the three queues based on CoS value, and given different likelihoods (weight) for tail drop when congestion occurs based on which egress queue is involved.
Define quantum value
The number of bytes in a queue that are removed per cycle in MDRR. Similar to byte count in the custom queuing (CQ) scheduler.
Define alternate mode
One of the two modes of MDRR, in which the priority queue is serviced between each servicing of the non-priority queues.
Define tail drop
An event in which a new packet arrives, needing to be placed into a queue, and the queue is full—so the packet is discarded.
Define full drop
A WRED process by which WRED discards all newly arriving packets intended for a queue, based on whether the queue’s maximum threshold has been exceeded.
Define priority queue
Jargon referring to any queue that receives priority service, often used for queues in an LLQ configuration that have the priority command configured.
Define sequence number
A term used with WFQ for the number assigned to a packet as it is enqueued into a WFQ. WFQ schedules the currently lowest SN packet next.
Define finish time
A term used with WFQ for the number assigned to a packet as it is enqueued into a WFQ queue. WFQ schedules the currently lowest FT packet next.
Define modified tail drop
A WFQ term referring to its drop logic, which is similar to tail-drop behavior.
A queuing tool’s logic by which it selects the next packet to dequeue from its many queues.
Define queue starvation
A possible side effect of a scheduler that performs strict-priority scheduling of a queue, which can result in lower-priority queues getting little or no service.
Define strict priority
A queuing scheduler’s logic by which, if a particular queue has packets in it, those packets always get serviced next.
Define software queue
A queue created by Cisco IOS as a result of the configuration of a queuing tool.
Define hardware queue
A small FIFO queue associated with each router’s physical interface, for the purpose of making packets available to the interface hardware, removing the need for a CPU interrupt to start sending the next packet out the interface.
Define remaining bandwidth
A CBWFQ and LLQ term referring to the bandwidth on an interface that is neither reserved nor allocated via a priority command.
Define maximum reserved bandwidth
A Cisco IOS interface setting, as a percentage between 1 and 99, that defines how much of the interface’s bandwidth setting may be allocated by a queuing tool. The default value is 75 percent.
Define actual queue depth
The actual number of packets in a queue at a particular time.
Define average queue depth
Calculated measurement based on the actual queue depth and the previous average. Designed to allow WRED to adjust slowly to rapid changes of the actual queue depth.
Define minimum threshold
WRED compares this setting to the average queue depth to decide whether packets should be discarded. No packets are discarded if the average queue depth falls below this minimum threshold.
Define maximum threshold
WRED compares this setting to the average queue depth to decide whether packets should be discarded. All packets are discarded if the average queue depth rises above this maximum threshold.
Define mark probability denominator
Used by WRED to calculate the maximum percentage of packets discarded when the average queue depth falls between the minimum and maximum thresholds.
Define exponential weighting constant
Used by WRED to calculate the rate at which the average queue depth changes as compared with the current queue depth. The larger the number, the slower the change in the average queue depth.
Define expedite queue
A term used with Cisco LAN switches, referring to a queue treated with strict-priority scheduling.
Define DSCP-to-CoS map
A mapping between each DSCP value and a corresponding CoS value, often used in Cisco LAN switches when performing classification for egress queuing.
Define DSCP-to-threshold map
A mapping between each DSCP value and a WRED threshold, often used in Cisco LAN switches when performing WRED.
Define internal DSCP
A term used with Cisco LAN switches, referring to a DSCP value used when making QoS decisions about a frame. This value may not be the actual DSCP value in the IP header encapsulated inside the frame.
Define differentiated tail drop
A term relating to Cisco LAN switch tail-drop logic, in which multiple tail-drop thresholds may be assigned based on CoS or DSCP, resulting in some frames being discarded more aggressively than others.
sed to reserve network resources for a flow as it traverses the network. A device that creates an RSVP reservation guarantees that it can provide the bandwidth, latency, or other resources that are requested by RSVP.
Variable name for the time interval used by shapers and by CAR.
With shaping, the number of bits allowed to be sent every Tc. Also defines the size of the token bucket when Be = 0.
With shaping and policing, the number of additional bits that may be sent after a period of relative inactivity.
In shaping and policing, commonly used to refer to the shaping or policing rate. For WAN services, a common reference to the bit rate defined in the WAN service business contract for each VC.
Define GTS shaping rate
A basic form of traffic shaping that is applied to an interface or subinterface. By default, it shapes all traffic leaving the interface, but can be modified by using an access control list. The access list controls only what traffic is shaped; GTS cannot provide different levels of QoS for different types of traffic.
Define policing rate
The rate at which a policer limits the bits exiting or entering the policer.
Define token bucket
A conceptual model used by shapers and policers to represent their internal logic.
Define Bc bucket
Jargon used to refer to the first of two buckets in the dual token bucket model; its size is Bc.
Define Be bucket
Jargon used to refer to the second of two buckets in the dual token bucket model; its size is Be.
Define adaptive shaping
A Frame Relay traffic shaping feature during which the shaping rate is reduced when the shaper notices congestion through the receipt of BECN or ForeSight messages.
A bit inside the Frame Relay header that, when set, implies that congestion occurred in the direction opposite (or backward) as compared with the direction of the frame.
A Cisco-proprietary messaging protocol implemented in WAN switches that can be used to signal network status, including congestion, independent of end-user frames and cells.
A Cisco-proprietary LMI protocol, implemented in Cisco WAN switches and routers, through which the switch can inform the router about parameters for each VC, including CIR, Bc, and Be.
Jargon referring to the minimum value to which adaptive shaping will lower the shaping rate.
Define map class
An FRTS configuration construct, configured with the map-class frame-relay global configuration command.
Define marking down
Jargon referring to a policer action through which, instead of discarding an out-of-contract packet, the policer marks a different IPP or DSCP value, allowing the packet to continue on its way, but making the packet more likely to be discarded later.
Define single-rate two-color policer
Policing in which a single rate is metered, and packets are placed into one of two categories (conform or exceed).
Define single-rate three-color policer
Policing in which a single rate is metered, and packets are placed into one of three categories (conform, exceed, or violate).
Define dual-rate three-color policer,
Policing in which two rates are metered, and packets are placed into one of three categories (conform, exceed, or violate).
A category used by a policer to classify packets relative to the traffic contract. The bit rate implied by all conforming packets is within the traffic contract.
A category used by a policer to classify packets relative to the traffic contract. With two-color policers, these packets are considered to be above the contract; for three-color, these packets are above the Bc setting, but within the Be setting.
A category used by a policer to classify packets relative to the traffic contract. These packets are considered to be above the traffic contract in all cases.
Define traffic contract
In shaping and policing, the definition of parameters that together imply the allowed rate and bursts.
Define dual token bucket
A conceptual model used by CB Policing when using an excess burst.
In two-rate policing, the second and higher rate defined to the policer.
Define nested policy maps
An MQC configuration style by which one policy map calls a second policy map. For example, a shaping policy map can call an LLQ policy map to implement LLQ for packets shaped by CB Shaping.