Help
Options
Didn't know it?
click below
click below
Knew it?
click below
click below
Don't Know
Remaining cards (0)
Know
retry
shuffle
restart
0:00
Embed Code - If you would like this activity on your web page, copy the script below and paste it into your web page.
Normal Size Small Size show me how
Normal Size Small Size show me how
CCIE R&S - QoS
CCIE R&S - Written - QoS
| Question | Answer |
|---|---|
| What are the three steps of configuring the QOS MQC? | class-map, policy-map, service-policy |
| What are the two functions of NBAR? | protocol analysis and traffic classification for QOS |
| What is the name of NBAR definitions? | PDLMs |
| What benefit does NBAR have over access-lists? | Deep packet inspection to the application layer |
| Command to enable NBAR monitoring? | ip nbar protocol-discovery |
| Layer 2 QoS marking? How many bits? | CoS, 3 |
| Layer 3 QoS markings? How many bits? | ToS - IP Precedence (First 3 bits of byte) and DSCP (First 6 bits of byte) |
| MPLS QoS marking? How many bits? | Experimental Bits or Traffic Class, 3 |
| Frame-Relay QoS marking? How many bits? | DE, 1 |
| ATM QoS marking? How many bits? | CLP, 1 |
| What is the CoS 0 Application? | Best Effort (Routine) |
| What is the CoS 01Application? | Medium Priority Data (Priority) |
| What is the CoS 2 Application? | High Priority Data (Immediate) |
| What is the CoS 3 Application? | Call Signaling (Flash) |
| What is the CoS 4 Application? | Video (Flash-override) |
| What is the CoS 5 Application? | Voice Bearer (Critical) |
| What is the CoS 6 Application? | Reserved (Internet) |
| What is the CoS 7 Application? | Reserved (Network) |
| What are the last two bits of the IP ToS byte used for? | Flow Control |
| What is the DSCP Class Selector PHB used for? | Pure backward compatibility with IP Precedence |
| What is the DSCP Default PHB and bit value? | Best effort service (000 000) |
| What is the DSCP Assured Forwarding PHB and initial bit values? | Guaranteed bandwidth services - 001, 010, 011, or 100 |
| Describe the DSCP AF drop probabilities. | 010 - Low probability, 100 - Medium probability, 110 - High probability |
| What is the DSCP EF PHB and bit value? | Voice, 101 110 |
| How do the drop probability values differ between AF and EF? | In AF, 110 means high probability of drop - In EF, 110 means don't drop |
| In what three ways can QoS Policy Propagation through BGP classify packets? | Access lists, BGP Community lists, and BGP AS Paths |
| What problem does QoS Priority Queuing introduce? | Starvation |
| What problem does QoS Custom Queuing introduce? | Latency for voice |
| What problem does Weighted Fair Queuing introduce? | Can't accommodate bandwidth reservations |
| What is the default behavior for full queues in CBWFQ? | Tail Drop |
| What is the default processing behavior in each CBWFQ queue? | FIFO |
| What three types of weighting can be used in CBWFQ? | Bandwidth, Bandwidth Percent, Remaining Bandwidth Percent |
| How many types of weighting can you use in one QoS service policy? | 1 |
| What percentage of interface bandwidth should not be assigned by CBWFQ? | 25% |
| How is Low Latency Queuing different from CBWFQ? | It adds a priority queue |
| How does Low Latency Queuing stop the priority queue from starving other queues? | the prioritized bandwidth is policed |
| What is QoS Modified Deficit Round Robin? | Deficit Round Robin with a priority queue added |
| What are the two modes of Modified Deficient Round Robin and how do they work? | Alternate - there will never be starvation, Strict priority - services the priority queue no matter what |
| This congestion avoidance technique prevents global synchronization? | Random Early Detection (RED) |
| What is the problem with Random Early Detection (RED)? | No way to differentiate what traffic to drop |
| What is the default Weight Random Early Detection denominator? | 10 |
| Explain the WRED min threshold. | Specifies at what level to start random drops |
| Explain the WRED max threshold | Specifies at what level to start full drop |
| Explain the WRED max drop probability. | Specifies what percentage of packets will be dropped (1/N) |
| Command to enable WRED and in what MQC mode? | random-detect, policy configuration mode |
| What is the default ToS type that WRED uses to determine a traffic profile? | IP Precedence |
| Explain 3 ways that policing and shaping are different? | Shaping can only be outbound, Shaping buffers excess traffic, Policing drops or re-marks traffic |
| What is the policing token formula? | CIR = Bc / Tc |
| What is Tc? | timing interval |
| What is Bc? | normal burst size |
| What are the three states of a Dual-Token, Single Rate Bucket? | Conforming, Exceeding, Violating |
| What two rates are enforced on a Dual-Rate Token Bucket? | CIR and PIR |
| What interfaces are supported for AutoQoS? | Physical Serial (HDLC and PPP), Frame-Relay P2P subinterfaces, and ATM P2P subinterfaces |
| What is the interface speed limit for ATM subinterfaces? | 768k |
| What is a requirement of AutoQoS for slow speed links? | It must be configured on both sides |
| What is a requirement of AutoQoS for an ATM PVC? | a bit rate has to be set |
| What 5 signaling protocols is AutoQoS built to identify? | H.323, H.225 (Unicast Only), SIP, 'Skinny', and MGCP |
| "Where does the command ""auto qos voip"" go in a frame-relay environment?" | in dlci config mode |
| What are the two phases of AutoQoS for Enterprise? | Discovery Phase, Installation Phase |
| How is AutoQoS similar to AutoQoS for Enterprise? | Same set of restrictions |
| Command to enable Phase 1 of AutoQoS for Enterprise? | auto discovery qos |
| Command to enable Phase 2 of AutoQoS for Enterprise | auto qos |
| What is the limit of CoS, IP Precedence, and DSCP values that can be listed on a single match command inside the MQC? | CoS: 4, IP Precedence: 4, DSCP: 8 |
| In what two ways can the bandwidth be set for a priority queue in the MQC? | Bandwidth, Bandwidth Percent |
| What does the priority command's option burst value default to? | 20% of configured priority bandwidth |
| What is the requirement in a QoS service policy to classify or mark based on CoS? | The interface must be a 802.1q trunk |
| If a QoS policy map defines more than 100% of the link's bandwidth, what action does IOS take when the command is typed in? | It rejects the command |
| What is required to perform QoS marking? | CEF enabled |
| Where should QoS marking be done? | As close to the ingress network edge as possible by a trusted device |
| What happens to packets that don't match a defined QoS class? | They are put in the class-default class |
| What is the limit of number of services classes that Cisco Recommends? | 4 or 5 |
| What action is taken on packets that are in an MQC policy with no 'set' command? | They are not marked |
| What is the restriction of classification and marking for non-IP headers? | They can only be classified inbound and marked outbound |
| What is the requirement for Layer 2 Ethernet marking and classification? | 802.1q or ISL trunks |
| What field are the QoS values stored in on a 802.1 tag? Size of this field? | Tag Control Field, 2 bytes |
| What are the 3 QoS bits of the Tag Control field called? Which bits are they? | user-priority bits, 3 most significant bits |
| What field are the QoS values stored in on a ISL header? Size of this field? | User Field, 1 byte |
| What are the 3 QoS bits of the User field called? Which bits are they? | CoS (Class of Service), 3 least significant bits |
| What command is used to change interface statistics more frequently for better analysis? What is the default value? Lowest value? | load-interval, default is 50 minutes, lowest value is 30 seconds |
| What is QoS pre-classification? | Classification done on VPN endpoint routers to make egress decisions based on original traffic before encryption |
| How does QoS preclassification work? | Router stores unencrypted traffic in memory until egress QoS actions are taken |
| What three places can the 'qos pre-classify' command be applied? | tunnel interface, virtual-template, or crypto map config modes |
| How can traffic be marked in PBR? | Using the set command to mark IP Precedence or entire ToS byte |
| What are two other names for the hardware queue? | Transmit ring or transmit queue |
| What does IOS do to better utilize the queuing tool present? | Shrink the length of the hardware queue |
| Each class in the policy map defines a _____________. | single queue |
| How many queues does CBWFQ and LLQ support? | 64 |
| What types of queuing can be used inside of the class-default queue? | FIFO or WFQ |
| In CBWFQ, what happens if some queues are empty? | Unused bandwidth is shared proportionally across all other classes |
| How is CBWFQ bandwidth remaining percent calculated? | max reserved bandwidth (percentage) times interface bandwidth |
| What is the 'max-reserved-bandwidth' default value? | 75 |
| In QoS, what constitutes a 'time of congestion'? | Hardware queue is full |
| How does LLQ treat multiple priority queues? | Grouped into a single LLQ with packets serviced according time of arrival |
| What is different with bandwidth command in LLQ vs. CBWFQ? | LLQ can use both bandwidth and bandwidth percent in the same policy map. CBWFQ can only use one of the three available. |
| What is a restriction of the class-default queue? | It cannot be a priority queue |
| What is the queue structure of Modified Deficit Round Robin (MDRR)? | seven round-robin queues (0-6) plus a priority queue (7) |
| What is the downside of using the alternate mode in MDRR? | jitter and latency for priority queue |
| How does alternate mode MDRR service all queues when they are full? | Services priority queue in between every other queue (0, P, 1, P, 2, P…) |
| What is the MDRR Quantum Value (QV)? | Number that specifies how many bytes can be sent per queue, per round |
| What happens if more bytes are sent than allowed in MDRR? | That queue has a deficit that will be added to the QV for the next round |
| How many switch ingress queues are there? | 2 |
| How many switch ingress queues can be configured as priority queues? | 1 |
| What does switch ingress queuing using to set discarded thresholds for each queue? | Weight Tail Drop (WTD) |
| How many thresholds are created on switch ingress queuing? | 3 |
| Which switch ingress queues are configurable? | 1 and 2 |
| What is the behavior of switch ingress queue Threshold 3? | It will drop traffic when the queue gets to 100% |
| What method is used to control the packet send rate from ingress queue to internal switch fabric? | Shared Round-Robin (SRR) |
| What do the relative weights represent in SRR? | Ratio of bandwidth sharing |
| In what mode are the switch ingress queuing commands applied? | Global configuration mode |
| How can traffic be mapped for QoS in switch ingress queuing? | CoS or DSCP value |
| Which queue is CoS 5 traffic placed into by default in switch ingress queuing? | Queue 2 |
| Which queue is non CoS 5 traffic placed into by default in switch ingress queuing? | Queue 1 |
| What does the command 'mls qos srr-queue input priority-queue (queue-id) bandwidth (weight)' accomplish? What does the 'weight' variable specify? | enables priority queuing, weight specifies link bandwidth percentage to use |
| What does the command 'mls qos srr-queue input buffers (percentage1) (percentage2)' accomplish? What are the variables' default values? | Specifies how to divide ingress buffer to the two queues, default is 90% to queue 1 and 10% to queue 2 |
| What does the command 'mls qos srr-queue input bandwidth (weight1) (weight2)' accomplish? What are the variables' default values? | Specifies frequency of packets being taken from two buffers, Default values are 4 and 4 |
| How many switch egress queues are there? | 4 |
| What is the restriction on a priority egress queue? | Must be queue 1 |
| When a priority egress queue is configured, how are bandwidth commands pertaining to that queue in the share or shape commands treated? | Ignored |
| How does switch egress queuing determine which four queues to place packets in? | Frames DSCP value compared to DSCP-to-CoS map, then that is compared to CoS-to-queue map |
| Is the DSCP-to-CoS map on a global or interface level? | global |
| Is the CoS-to-queue map on a global or interface level? | interface |
| What are the two options for switch egress scheduler? | Shared, Shaped |
| What is the default weights of the four switch egress queues? | 25 percent each |
| What is the difference between shared and shaped mode for egress queuing? | Shared mode only polices during congestion, shaped polices at all times |
| Can a 3560 have shared and shaped mode on the same interface? | Yes |
| What can be configured for each egress queue-set? | buffer and WTD thresholds |
| How many queue-sets are there for switch egress queuing? | 2 |
| How are reservations made in RSVP? | On a flow-by-flow basis |
| How is an RSVP flow defined? | Destination IP address, protocol ID, and destination port |
| How many reservations are made for one host communication with another? | 2 |
| What type of RSVP message is sent from the source Router towards the destination? | PATH |
| What does a RSVP PATH message contain? | Previous Hop information, bandwidth, and QoS needs |
| What does a non-source, non-destination, RSVP-aware router do when it receives a PATH message? | Change PHOP and forward |
| What does the last router do when it receives the RSVP PATH message? | Replies with a RESV message to source router and initiates PATH message to source router |
| What does a non-source, non-destination, RSVP-aware router do when it receives a RSVP RESV message and cannot fulfill the requirements? | Send error message to originator and drops RESV packet |
| How does the source router know that reservations have been made in both directions? | It receives a RESV CONF message |
| What does RSVP use to be able to dynamically adjust to network changes? | periodic refresh messages |
| What values are uses if these variables are not set? 'ip rsvp bandwidth (total-kbps) (single-flow-kbps)' | total-kbps = 75% of interface bandwidth, single-flow-kbps = any size flow |
| What QoS method does RSVP use by default? How to change this behavior? | WFQ, ip rsvp resource-provider none |
| How does the command 'ip rsvp data-packet classification none' change RSVP behavior | It prevents RSVP from trying to classify every packet |
| What considerations need to be taken when setting the RSVP bandwidth using LLQ? | priority queue minus Layer 2 overhead |
| What is the average queue depth in WRED? | The measurement compared with minimum and maximum thresholds to determine when packets should be dropped |
| When the average queue depth is above the max threshold, what happens? | Full Drop |
| How is full drop different from tail drop? | full drop can discard packets even when the queue is not full |
| What three places can WRED be configured? | physical interface with FIFO, non-LLQ class in CBWFQ policy map, and ATM Virtual Circuit |
| This WRED value is the calculation of the rolling average queue depth. | exponential weighting constant |
| What does a low Exponential Weighting Constant mean? | Old average is a small part of the average calculation, this makes the average change more quickly |
| When does AutoQoS trust the values it receives? | When the link is a uplink port, trunk port, or Cisco phone |
| What does a switch do if the access port is not connected to a Cisco phone? | Mark all traffic down to DSCP 0 |
| If the switch trusts QoS markings, which 5 traffic types are placed into the priority queue? | Voice and Video control, Video, Voice, routing protocol, and STP traffic |
| When a switch trusts the QoS markings which traffic goes into the normal input and output queues? | Input: non Voice and Video control, non Video, no Voice, non routing protocol, and non STP traffic; Output: non Voice traffic |
| What is enabled on links with speeds 768k or less for AutoQoS? | Compression and Fragmentation |
| In AutoQoS, what will happen to non-frame relay serial interfaces that are 768k and less? Why? | Their encapsulation will be changes to PPP, to enable link fragmentation and interleaving |
| If the 'trust' keyword is used, what action does the router take on that AutoQoS interface? | Class maps are created based on DSCP values |
| If the 'trust' keyword is omitted, what action does the router take on that AutoQoS interface? | Access lists are created to match voice, video, and call control ports |
| What technology does CB Policing replace? | CAR |
| What 7 actions can be taken in CB policing? | transmit, set dscp and transmit, set precedence and transmit, set qos and transmit, set clp and transmit, set frde and transmit, drop |
| What does 1 token represent in policing? | 1 byte |
| How often does a policer refill a bucket? | Whenever a packet arrives |
| How does a policer refill the bucket? | Prorated number of tokens [(CurrPktArvTime - PrevPktArvTime) * PolRate] / 8 |
| How are dual-rate tokens bucket filled and drained? | Bc filled based on CIR, Be filled based on PIR, tokens drained from both bucks for packets conforming to CIR |
| What configuration constitutes three colors in CB Policing? | violate action or Be higher than 0 |
| What configuration constitutes dual rate? | CIR and PIR configured |
| What command is used to get into multi-action policing (sub configuration) mode? | omit action parameters from the 'police' command |
| In what two ways can a policing rate be set? | bps or bandwidth percentage |
| What is the only flavor of policing that CAR supports? | Singe rate, two color |
| What are the 4 main differences that CAR has in comparison to CB Policing? | non-MQC 'rate-limit' command, can used nested (cascading) rate-limit commands, supports Be but not violate action, different internal logic to determine if a packet conforms or exceeds |
| What is the purpose of using nested (or cascading) rate-limit commands in CAR? | Allows multiple commands to process the same packet |
| What are the two ways to classify traffic in CAR? | Normal IP ACLs or Rate-Limit ACLs |
| What can a CAR Rate-Limit ACL match? | MPLS Exp Bits, IPP, or MAC address |
| What are the three typical causes of QoS problems? | No proper planning, outdated configuration, no use of documentation |
| What is egress blocking? | SP frame-relay or ATM switch has to queue data before it can send it out the other end of the VC |
| What is the difference between the terms CIR and shaping rate? | CIR is the contractual rate, shaping rate is the configured rate |
| What does 1 token represent in shaping? | 1 bit |
| When is the token bucket filled in traffic shaping? | At the beginning of every Tc interval |
| In the single token bucket model, what happens to extra tokens that do not fit in the bucket? | They are wasted |
| In the single token bucket model, how is Be implemented? | The bucket capacity is enlarged |
| When congestion occurs, the shaper can lower the shaping rate until __________________ or _______________. | It reaches the minimum information rate or congestion stops |
| How is adaptation triggered in traffic shaping? | Router receives a frame with the BECN bit set or a Cisco ForeSight congestion message |
| Each time the router receives a congestion notification, how does the shaping rate change? | Bc and Be decreased by 25% of maximum |
| When does the shaping rate grow after being slowed down due to congestion messages? | Grows after 16 consecutive Tc values without a congestion message by 1/16 until maximum rate reached |
| In Generic Traffic Shaping (GTS) what is the only value required? | shaped-rate |
| In Generic Traffic Shaping (GTS), if Bc or Be is not required, what are their values? | Both set to 1/4 of the shape-rate |
| What command is added in GTS to configured adaptive shaping on a Frame Relay link? | traffic-shape adaptive bit-rate |
| What method of shaping allows the creation of a settings group to be applied on multiple interfaces? | Class-Based Shaping |
| What two ways can a shaping rate be specified? | bps or bandwidth percentage |
| What is the default shaping rate for a subinterface with no bandwidth command? | 1544 |
| In Class-Based Shaping, how is Bc and Be measured if configured? | Number of milliseconds |
| If shape rate <= 320 kbps in MQC, what are the values for Bc, Be, and Tc? | Bc = Be = 8000 bits, Tc = Bc / shape rate |
| If shape rate > 320 kbps in MQC, what are the values for Bc, Be, and Tc? | Tc = 25 ms, Bc = Be = shaping rate * Tc |
| What type of queuing does software queues default to? | WFQ |
| What type of queuing does shaping queuing default to? | FIFO |
| What is the benefit of small Tc values? | Low latency |
| What does shaping become active? Inactive? | When a single packet exceeds the policy, Inactive when all shaping queues are empty |
| What can FRTS learn? | CIR, Tc, Bc, Be from frame-relay switch |
| What restriction does FRTS have on a physical interface? | Can't apply special queuing tools with FRTS |
| What is structure is used to assign a shaping policy in FRTS? | Map Class |
| What is the preference of assigning a map class to an interface in FRTS? | First looks under the interface DLCI. If not there, looks under the sub-interface. If not there, it looks under the physical interface. If not there, it applies the defaults |
| If a map class is not found on an interface, what default values are applied? | Shaping rate = 56 kbps, Bc = 7000 bits, Tc = 125 ms |
| How are frame-relay values explicitly set in FRTS? Why is this used? | under the 'map-class frame-relay NAME' command, used to achieve small Tc |
| In a non map-class configuration with FRTS, what is the value of Tc? | 125 ms |
| In a non map-class configuration with FRTS, what is needed to use a Be? | peak rate must be configured |
| What must be used in FRTS to enable adaptive sharing? | map-class with commands frame-relay adaptive shaping {becn | foresight} |
| What is the default mincir rate for adaptive sharing in FRTS? | 50% of shaping rate |
| What must be created to use the MQC with FRTS? | default-class |
| If FRTS and fragmentation are both applied to a PVC with the MQC, what type of queue will be used? | dual FIFO queue, 1 for voice and control traffic, the other for everything else |
| When using nested policy maps in FRTS, what is the total rate available to the child? | Parent shaping rate |
| If both the shape average and shape adaptive command is configured in the MQC for FRTS, which bandwidth is used? | shape adaptive bandwidth |
| What replaces the 'frame-relay ip rtp priority' command for FRTS in the MQC? | LLQ |
Created by:
ajballer15
Popular Computers sets