Nine:
1. Week List 2 reasons why QoS is needed for VoIP. Method of measuring level of service to end user. Method of assessing assurancethat a particular application can meet its service
requirements. 2. List 5 technical components of QoS Availability Throughput Connection Setup Time % of successful transmissions Speed of fault detection and correction 3. List 4 ways in which QoS could be measured in an IP network. Bandwidth Packet Loss Delay Jitter 4. Define a `level' of QoS. For a given session, the network must provide assurance that the measurement of the QoS parameters will fall within certain bounds. 5. What assumptions are made if UDP is used to transport data? Use of UDP assumes that network is not congested, and packet loss is relatively low. 6. List 2 effects of excessive data loss in a VoIP implementation. Some CODECS can compensate for a small amount of data loss. Excess data loss causes `drop-outs', and lowers users perceptions of QoS. 7. List 5 `non-network' QoS expectations. Rapid service provisioning Excellent Customer Service Accurate Billing Competitive Rates Superior Marketing 8. Briefly describe 3 network approaches to solving QoS. Ensure all required resources are dedicated prior to the session. Fairly share resources depending on priority. Overprovision the network. 9. What 2 levels of service does RSVP offer? Guaranteed Controlled Load 10. Explain the RSVP terms "Guaranteed" and "Controlled Load" Guaranteed - Close to Circuit Switching level of service. Controlled Load - Equivalent to best-effort under no-load. 11. Explain briefly how RSVP works RSVP is a signaling protocol that, in the context of Intserv, is used for admission control and resource reservation. There are two main control messages in RSVP:
Path message this message is sent by the sender to the receiver and is used to (1) carry the sender's traffic specifications, and (2) set up the path. Upon receiving this message each node stores the previous node along the path. Resv message this message is sent by the receiver to sender and is used to perform admission control and resource reservation. Upon receiving such a message a node checks whether it has enough resources to satisfy the receiver requirements. If yes, it reserves resources; if not, it marks the message to indicate that the reservation is denied. In both cases the node forwards the message to the previous node (set up by the Path message) towards the source.
12. TOS bits that are set to 1 basically help speed up the packet flow. 13. List the 8 TOS precedence's, and their binary values. 000 (0) - Routine 001 (1) - Priority 010 (2) - Immediate 011 (3) - Flash 100 (4) - Flash Override 101 (5) - Critical 110 (6) - Internetwork Control 111 (7) - Network Control 14. What 2 PHB types does DiffServ define?. Expedited Forwarding (EF) Assured Forwarding (AF) 15. AF has four classes Each class is allocated a specific amount of resources within the router. Each class has 3 drop rates (making 12 combinations in all) Packets with the highest drop rate are discarded first. Depends on packets having a proper AF; if the same AF is used for all packets, then
nothing is gained. 16. Explain what happens to traffic when a packet enters an MPLS domain. Traffic is marked by a router as it enters an MPLS domain 17. All packets with the same FEC are given the same forwarding treatment. List the 4 elements of the forwarding treatment. All packets traveling from `A' to `B' take the same path if they have the same FEC. They all have the same pre-allocated bandwidth. They all arrive in the same order. The above results in a higher QoS. 18. Define "Traffic Aggregate". Traffic Aggregate is a collection of packets with a code point that maps to the same PHB, usually in a DS domain or some subset of a DS domain.
Week Ten:
DSCP value is set by the ___________________ ____or_______ ______________. 2. Describe 4 aspects of traffic conditioning. Source Address Destination Address Protocol ID Time of Day 3. Define Traffic Policing Traffic policing is a traffic regulation mechanism that is used to limit the rate of traffic streams. 4. A traffic `stream' is said to be conforming when While waiting to be serialized at the physical interface. 5. Define `Traffic Shaping' Traffic shaping is a tool used to manage network traffic by shaping the traffic to a specified rate. 6. List 7 items that traffic shaping allows you to do. Control access to available bandwidth Ensure that traffic conforms to the policies established for it. Regulate the flow of traffic to avoid congestion Ensure that a packet, or data source. Determine the QoS to apply to the packet. Enables you to control the traffic leaving an interface Eliminate bottlenecks in topologies with data-rate mismatches. 7. Exp|ain `Class--based' Traffic Shaping. Class-based traffic shaping allows you to control the traffic going out of an interface in order to match its transmission to the speed of the remote target interface and to ensure that the traffic conforms to policies contracted for it. 8. Define `Percentage-based' Traffic Shaping. Percentage-based shaping allows you to configure traffic shaping on the basis of a percentage of bandwidth available on the interface. 9. List 2 pairs of differences between policing and shaping Criteria Shaping Policing Primary Function Buffers and queues excess Drops or remarks excess packets above the committed packets above the committed rates. rates. Does not buffer. Bits per second (bps) No Yes Bytes Yes Yes
1. The
Token Value Applicable on Inbound Applicable on Outbound
10. Define `Per hop Behaviour' Defined as the packet applied treatment by a router based on the specific DSCP value. 11. All flows between two nodes, with the same DHCP value, are said to comprise an aggregate. 12. List 2 PHBs
Expedited Forwarding (EF) Assured Forwarding (AF) 13. What does the PHB EF attempt to provide? EF PHB tries to provide low loss, low delay, low jitter service. 14. Describe the characteristics of priority queuing as it applies to EF. EF traffic will always get pre-emptive priority over non-EF traffic. Potentially cripples performance of non EF traffic. 15. What risk is there to AF traffic? If it is `out of profile' all packets, then _____________is gained. 16. Explain why AF traffic that is out-of-profile could receive a best-effort service that is worse than if all traffic were best effort. In a non-DiffServ network.
Week Eleven:
1. Which device labels and strips labels when traffic enters and leaves an MPLS
domain? Routers 2. Explain the term "MPLS Shim Header" MPLS shim header is a special header placed between layer two and layer 3 of the OSI model. 3. Routers not at the edge of an MPLS domain are called? LSR (Label Switch Routers) 4. An MPLS `Shim Header'is a special header inserted between layer two and layer 3 of the OSI model. 5. Explain the purpose of an LSR. The purpose of LSR is to examine incoming packets. 6. Describe an LSP. Paths are established between the LER and the LSR these paths are called LSP (Label Switch Paths). 7. T/F: MPLS and IP can coexist in the same MPLS domain without interfering with the operation of each other. True 8. List 5 functions of an LER.
9. The traffic classification process on an MPLS network is called the FEC (Forward
Equivalence Class).
10. Describe "triggered mode"
One decision-making method is called "triggered mode." 11. The forwarding table in an LSR is called the Label Information Base [LIB] or a connectivity table. 12. Compare RSVP, DiffServ, and MPLS. RSVPVery powerful at the session levelPath state must be maintained in each router (scaling problems) DiffServSimpler approachPrioritizes, does not reserve resources
MPLSA good overall solutionRequires significant administration and software/firmware 13. When designing a VoIPnetwork, list 3 `dimensions'which must be balanced. Cost Capacity Quality 14. List 3 criteria when designing a `carrier grade'VoIPNetwork. `5 nines' Capacity to meet projected business needs Must cost less to build and operate than the revenue generated. 15. List 6 design criteria when considering an overall approach to a VoIP network. Build Ahead (Capacity Buffer) CODECS Network Technology choices Silence Suppression decision Resilience and Redundancy Blocking 16. Define `Build Ahead' Essentially, you predict how much capacity will be needed at some point in the future, and you design the network to satisfy the future requirements NOW. 17. List the advantages and disadvantages to VAD. Use of VAD will reduce bandwidth requirements, but can cause clipping and lack of comfort noise may confuse subscribers. 18. About how much B/W does VAD conserve? 80% of the total bandwidth 19. What is a good compromise value for packetization interval? 20 40 ms is a good compromise value. 20. Define and give examples of one Erlang 1 Erlang corresponds to the use of a number of channels for total of 1 hour; 1 channel for 1 hour 21. Explain the term "N+1"redundancy
22. Calculate the MTBF of a system with 7 components, each having an MTBF of 5, 5, 10,
15, 15, 20, and 25 years respectively.
23. List 3 categories of node alarms. Critical Major Minor
Week Twelve:
1. There are 20 IP Soft phones at location `A'. Each Soft phone is using the G.729 CODEC,
which operates at 8Kbps. Assuming a 3 10ms samples per packet, PPP over a fractional
2. 3. 4. 5.
T1 to location `B', how many simultaneous calls could be made from A to B if the fractional T1 was operating at 768Kbps? As above, but assume a 10Mbps Ethernet link between A and B. As in 1, but assume G.723.1-5.3K CODEC. G.723.1-5.3K produces frame sizes of 30ms. Each frame has 3 control bits. Assume 1 frame per packet, and MPLS. As in 3, with G.723.1-6.3K CODEC. Frame size is 30ms and 3 control bits per frame. As in 3, but assuming a full T1 rate.
Week Thirteen:
1. How many DS0s are required to support 200 IPSoftphones using G.723-5.3K with 30ms
2. 3. 4. 5. 6.
7.
8.
frames, 2 frames per packet and PPP, MPLS, and RTCP with cRTP reducing the RTP, IP and UDP headers to 4 bytes? As above, without cRTP? What is the frame size of a CODEC which operates at 24000 Kbps and has a frame length of 50 bytes? Assuming that the calculation overhead on coding is 5 ms and decoding is 2 ms, how much delay would the use of the above CODEC introduce end-to-end? What is the frame length of a CODEC with operates at 8000 Kbps and generates 10ms packets? G.729 represents 10 ms with 10 bytes of data. 3 G.729 samples are transported within a packet. There are 75 IP phones using the above coding scheme, each generating .5 Erlangs of traffic during the busy hour. How much bandwidth will you need to provision on the WAN (PPP and MPLS) side, assuming also RTCP and cRTP if you want to provide a GoS of .005? How many DS0s are required to support 100 concurrent IPSoftphones using G.723-6.3K with 30ms frames and PPP, MPLS, and RTCP with cRTP reducing the RTP, IP and UDP headers to 4 bytes? How many inbound trunks are required to support a GoS of .001 with 100 IPSoftphones using G.723-6.3K with 30ms frames and PPP, MPLS, and RTCP with cRTP reducing the RTP, IP and UDP headers to 4 bytes? Each softphone generates .4 Erlangs of busy hour load. How much LAN side bandwidth is required from the router?