GoConqr hat ein neues Design! Sag uns, wie es dir gefällt.

Kostenlos registrieren

Karteikarten

Chapter 5 Network layer: The Control Plane

Dylan Boyd 364 Englisch 136 Karten

Karten in diesem Stapel (136)

  1. 1In which plane does forwarding occur?The data plane
  2. 2In which plane does routing occur?The control plane
  3. 3Which network-layer function involves moving packets from a router's input to the appropriate output?Forwarding
  4. 4Which network-layer function involves determining the route taken by packets from a source to a destination?Routing
  5. 5What is another name for the traditional approach to structuring in the network control plane?Per-router control
  6. 6What's another name for software-defined networking when structuring in the network control plane?Logically centralised control
  7. 7In a logically centralised control plane, what does a distinct (typically remote) controller interact with to compute forwarding tables?Local control agents (CAs)
  8. 8In a per-router control plane, what is the purpose of individual routers interacting with each other?To compute forwarding tables
  9. 9What is the goal of a routing protocol?To compute "good" paths or routes from a sending host to a receiving one
  10. 10What traverses a sequence of routers in a path?Packets
  11. 11In a network graph abstraction G = (N,E), what are N and E?N = set of routers/peers
    E = set of links/TCP connections
  12. 12What is the goal of a routing algorithm?To find the least cost path
  13. 13In a global routing algorithm, what do all routers have?Complete topology and link cost information
  14. 14What type of algorithm is used in global routing?Link state algorithm
  15. 15In a decentralised routing algorithm, what two pieces of information do routers have?1. Physically-connected neighbours
    2. Link costs to neighbours
  16. 16In decentralised routing, what is the nature of the process of computation that occurs to exchange information?Iterative
  17. 17What type of algorithm is used in decentralised routing?Distance vector algorithm
  18. 18In what type of routing do routes change slowly over time?Static
  19. 19In what type of routing do routes change more quickly?Dynamic
  20. 20In what type of routing do periodic updates occur in response to link cost changes?Dynamic
  21. 21What does load-sensitivity indicate about a link?Whether link costs vary based on congestion levels
  22. 22Using Dijkstra's algorithm, which nodes know of the topology and link costs, and have the same information?All nodes
  23. 23By what method does Dijkstra's algorithm share information between nodes?Link state broadcast
  24. 24What does Dijkstra's algorithm compute?The least cost paths from the source node to others
  25. 25What does the result of Dijkstra's algorithm allow us to compute?The forwarding table for a node
  26. 26After k iterations of Dijkstra's algorithm, what will we know?The least cost path to k destination nodes
  27. 27Using Dijkstra's algorithm, what does c(x, y) mean?The link cost from node x to node y
  28. 28Using Dijkstra's algorithm, what is the initial link cost from node x to y if they are not direct neighbours?Infinity
  29. 29Using Dijkstra's algorithm, what does D(v) mean?The current cost of least-cost path from source to destination v
  30. 30Using Dijkstra's algorithm, what does p(v) mean?The predecessor node along current least-cost path from source to v
  31. 31Using Dijkstra's algorithm, what is N'?The set of nodes whose least cost path is definitively known
  32. 32What is the complexity of Dijkstra's algorithm?O(n^2)
  33. 33In what algorithm is the Bellman-Ford equation used?Distance vector algorithm
  34. 34Using the Bellman-Ford equation, how is the minimum of the result used?The minimum node is the next hop in the shortest path of the routing table
  35. 35In the distance vector algorithm, what is Dx(y)?Estimate of least cost from x to y
  36. 36Using the distance vector algorithm, what is c(x, v)?The cost to a neighbour v
  37. 37What three pieces of information does a node maintain using the distance vector algorithm?1. Cost to each neighbour
    2. Distance vector of cost estimates to all destination nodes
    3. Distance vectors of each of its neighbours
  38. 38Using the distance vector algorithm, what does a node do periodically?Send its own distance vector estimate to neighbours
  39. 39What does a node do when it receives a distance vector update from a neighbour?It updates its own distance vector estimate using the Bellman-Ford equation
  40. 40What does a node do if its own distance vector changes?Communicate the distance vector to neighbours
  41. 41What condition is required for cost estimates to converge to the real value?Nodes exchange distance vectors in an asynchronous manner
  42. 42Which is communicated slower: a distance vector decrease or increase? Briefly explain why.Increase - count to infinity problem
  43. 43In the link state algorithm, how many messages are sent given N nodes and E links?O(NE)
  44. 44Between distance vector and link state, one can advertise an incorrect path cost, and one an incorrect link cost. Which is which?Distance vector: Incorrect path cost
    Link state: Incorrect link cost
  45. 45To support scale, routers are aggregated into regions known under which two terms?Autonomous systems (AS)
    Domains
  46. 46What must be the same among routers in an intra-AS network?The intra-domain protocol
  47. 47What is the type of node at the edge of its own AS?Gateway router
  48. 48What algorithm(s) determine the forwarding table entries for external destinations?Inter-AS and intra-AS
  49. 49What is another name for intra-AS routing?Interior gateway protocols (IGP)
  50. 50What are the three most common intra-AS routing protocols?1. RIP (routing information protocol)
    2. OSPF (open path shortest first)
    3. IGRP (interior gateway routing protocol)
  51. 51Which algorithm does OSPF use?Link-state algorithm
  52. 52Briefly describe the communication method used in OSPFRouter floods OSPF link-state advertisements to all other routers in AS
  53. 53What other protocol is nearly identical to OSPF?IS-IS routing
  54. 54Which protocol allows for multiple same-cost paths?OSPF
  55. 55Which protocol allows for multiple cost metrics for different ToS?OSPF
  56. 56List the four types of routers in hierarchical OSPF1. Boundary router
    2. Backbone router
    3. Area border router
    4. Internal router
  57. 57What is the purpose of a boundary router in OSPF?Connect to other AS's
  58. 58In OSPF, what are the two purposes of an area border router?1. Summarise distances to networks in its own area
    2. Advertise existence to other area border routers
  59. 59In hierarchical OSPF, what entity are link-state advertisements limited to?An area
  60. 60What protocol is the "glue holding the Internet together"?BGP (border gateway protocol)
  61. 61An AS uses an eBGP connection to do what?Obtain prefix/subnet reachability information from neighbouring AS's/domains
  62. 62An AS uses an iBGP connection to do what?Propagate reachability information to all AS-internal routers
  63. 63BGP determines the "best" route to other networks based on what two factors?1. Reachability information
    2. Policy
  64. 64Are BGP messages exchanged over TCP or UDP?TCP
  65. 65What are the four types of BGP messages?1. OPEN
    2. UPDATE
    3. KEEPALIVE
    4. NOTIFICATION
  66. 66What is the purpose of an the BGP message OPEN?To open a TCP connection to a remote BGP peer and authenticate the sending BGP peer
  67. 67What is the purpose of an the BGP message UPDATE?To advertise a new path (or remove an old one)
  68. 68What are the two purposes of an the BGP message KEEPALIVE?1. Keeps connection alive in the absence of UPDATEs
    2. ACKs OPEN request
  69. 69What are the two purposes of an the BGP message NOTIFICATION?1. Reports errors in previous message
    2. Close connection
  70. 70What entities run on both iBGP and eBGP protocols?Gateway routers
  71. 71Over what kind of a TCP connection do peers exchange BGP messages?Semi-permanent
  72. 72In BGP, what does one AS promise another?That it will forward packets to a given destination.
  73. 73What two things make up a BGP route?1. Prefix
    2. Attributes
  74. 74What does the BGP attribute AS-PATH describe?The list of AS's through which prefix advertisement has passed
  75. 75What does the BGP attribute NEXT-HOP describe?Indicates specific internal-AS router to next-hop AS
  76. 76In BGP policy-based routing, a gateway receiving route advertisements uses what to accept/decline a path?Import policy
  77. 77An AS policy determines whether to advertise a path to what?Neighbouring AS's
  78. 78What entity in BGP can learn about multiple paths to a destination?Gateway router
  79. 79In BGP, what four criteria are considered when choosing a route?1. Policy decision
    2. Shortest AS-PATH
    3. Closest NEXT-HOP router: Hot Potato
    4. Additional criteria
  80. 80Briefly describe hot potato routingChoose local gateway that has least intra-domain cost to a NEXT-HOP router
  81. 81What cost(s) does hot potato routing take into account out of intra-domain costs, and inter-domain costs?Intra-domain costs
  82. 82What method could an ISP use with regards to BGP to restrict traffic from other ISPs?Restrict advertisements
  83. 83What can hierarchical routing reduce the size of?Routing tables
  84. 84How was the Internet network layer historically implemented?Distributed, per-router approach
  85. 85When was there a renewed interest in rethinking network control?2005
  86. 86In which two ways does a logically centralised control plane allow for easier network management?Avoid router misconfigurations, greater flexibility of traffic flows
  87. 87With regards to table-based forwarding, which is more difficult: centralised or distributed?Distributed
  88. 88What type of control plane allows for an open (non-proprietary) implementation?Logically centralised
  89. 89What are the three components of SDN?SDN controller, data plane switches, control applications
  90. 90In SDN, what is the purpose of data plane switches?To implement generalised data-plane forwarding in hardware
  91. 91In SDN, what part of a data plane switch allows for defining what is controllable and what is not?API, such as OpenFlow
  92. 92What is the purpose of an SDN controller?Maintain network state information
  93. 93In SDN, what is the target of interaction through the northbound API?Network control applications
  94. 94In SDN, what is the target of communication through the southbound API?Network switches
  95. 95An SDN controller is implemented as what type of system for performance, scalability, fault-tolerance, and robustness?Distributed
  96. 96What is the purpose of network-control apps in SDN?The brains of control
  97. 97Which component of an SDN has common standards for easy use/reuse?Network control apps
  98. 98In SDN, what does the communication layer allow interaction between?SDN controller and controlled switches
  99. 99An OpenFlow controller operates between what two entities?Controller and switches
  100. 100What protocol does OpenFlow use to exchange messages?TCP
  101. 101What are the three classes of OpenFlow message?Controller-to-switch, switch-to-controller, symmetric
  102. 102What is the purpose of the OpenFlow controller-to-switch message Features?Controller queries switch features, switch replies.
  103. 103What is the purpose of the OpenFlow controller-to-switch message Configure?Controller queries/sets switch configuration parameters
  104. 104What is the purpose of the OpenFlow controller-to-switch message Modify-state?Add, delete, modify flow entries in the OpenFlow tables of switch
  105. 105What is the purpose of the OpenFlow controller-to-switch message Packet-out?Controller can send specific packet out of specified port of a switch. Message contains packet in payload.
  106. 106List the four controller-to-switch message typesFeatures, Configure, Modify-state, Packet-out
  107. 107What is the purpose of the OpenFlow switch-to-controller message Packet-in?Transfer packet (and its control) to controller. Processing packet does not match any flow-table entry; additional processing for matched packets
  108. 108What is the purpose of the OpenFlow switch-to-controller message Flow-removed?Flow table entry deleted at a switch
  109. 109What is the purpose of the OpenFlow switch-to-controller message Port status?Inform controller of a change on a specific port
  110. 110List the three switch-to-controller message typesPacket-in, Flow-removed, Port status
  111. 111What is the name of the SDN controller including the term "Lithium"?OpenDaylight (ODL)
  112. 112Using an ODL controller, where are network apps, choosing out of within/external to SDN controller?Either - application designer's choice
  113. 113How do external apps communicate with an ODL controller?Using controller modules through REST API over HTTP
  114. 114What is the purpose of the service abstraction layer in ODL?To interconnect internal apps, external apps, and services
  115. 115In an ONOS controller, are the control apps bound to the controller or separate from it?Separate
  116. 116Briefly describe the intent framework of an ONOS controllerHigh-level specification of service; what rather than how
  117. 117Where is emphasis placed in an ONOS controller?Distributed core
  118. 118What are the three main objectives of an ONOS controller's design?Service reliability, scaling of service replication, device/protocol agnostic
  119. 119List three mission-specific requirements of SDNReal-time, ultra-reliable, ultra-secure
  120. 120Where are ICMP messages carried?In IP datagrams
  121. 121Describe "where" ICMP fits in layer-wiseNetwork-layer "above" IP
  122. 122What three parts are there to an ICMP message?Type, code, first eight bytes of IP datagram causing error
  123. 123What type of messages are used in traceroute?ICMP
  124. 124When performing a traceroute, what value is decremented to ensure the desired host is reached?TTL (or time to live)
  125. 125Managed devices contain managed objects whose data is gathered into what?Management Information Base (MIB)
  126. 126What protocol is used for managing a network?Simple network management protocol (SNMP)
  127. 127What two modes are used to convey MIB information?Request/response mode, trap mode
  128. 128What are the two types of manager-to-agent SNMP messages?GetRequest and SetRequest
  129. 129What is the only type of manager-to-manager message?InformRequest
  130. 130What are the two types of agent-to-manager message?Response and Trap
  131. 131What is the purpose of an SNMP GetRequest/GetNextRequest/GetBulkRequest message?"Get me data"
  132. 132What is the purpose of an SNMP InformRequest message?"Here's MIB value"
  133. 133What is the purpose of an SNMP SetRequest message?Set MIB value
  134. 134What is the purpose of an SNMP Response message?Response to request
  135. 135What is the purpose of an SNMP Trap message?Inform manager of exceptional event
  136. 136What are the three fields in the get/set header of an SNMP message?Request ID, error status, error index

Informationen

Beschreibung
364 Karteikarten am Chapter 5 Network layer: The Control Plane, erstellt von Dylan Boyd am 16/05/2017.
Erstellt von
Dylan Boyd
Aufrufe
0
Bewertung

Feedback

Worum geht es?

Wie gefällt dir das neue Design?

Dein Eindruck