Internet Technology NETW902 Final Exam 11th Jan. 2011 Please read carefully before proceeding: 1. This is an Open Book exam 2. The Exam duration is 2 hours 3. The exam booklet contains Pages including the cover page. Viel Erfolg; Problem 1 Question Max. Pt. Grade Q1 Part I Q2 Q3 Q4 Q5 Q1 12 30 20 6 20 6 Part II Q2 Q3 19 27 Problem 2 Total Q4 Q1 Q2 Q3 Q4 Q5 8 20 18 34 25 28 273 Problem 1 Internet Routing Note: Parts 1 and 2 can be solved independently of each other. Part 1 Routing Algorithms In Figure 1, the topology is sketched of an Internet Service Provider (ISP) network. In the following tasks, various routing algorithms will be studied. Question 1 Adjacency Matrix a) Give the cardinality |V| and the size |E| of the graph G in Figure 1. b) Construct the adjacency matrix A = (aij) of the graph G. c) Interprete the connectivity of the nodes and their reflection within A. Question 2 Distance Vector Routing Algorithm a) Construct the routing table for the Distance Vector Algorithm hosted in a1) node 1 a2) node 8 b) Consider the case that link number 15 is interrupted due to an outage of the transmission facilities. Construct the Routing Table for node 8. c) Describe shortly how the Routing Tables are kept up to date due to events like equipment outages (links, nodes). d) In the basic Distance Vector Routing Algorithm, nodes have only a partial knowledge of the network state. Describe shortly how the source-to-destination route is found. e) How could the Distance Vector Routing Algorithm be upgraded to follow estimates for the source-to-destination packet transfer delay? Question 3 Link State Routing Algorithm a) Construct the routing table for the Link State Algorithm. A1) Which information contains the Link State Algorithm? A2) Construct the Routing Table. b) Describe how the source-to-destination route is found. Question 4 Policy-Driven Routing Algorithms Describe possibilities by which the source-to-destination path is found by policies rather than routing tables. Question 5 Autonomous Systems Consider now the case where the network given by the graph of Figure 1 is subdivided into 3 Autonomous Systems (AS) containing the following nodes: AS 1: nodes 2, 3, 4 AS 2: nodes 1, 5 AS 3: nodes 6, 7, 8 a) Subdivide the graph of Figure 1 such that the 3 Ass are completely separated where the links within each of the Ass are maintained from the graph G. b) Interconnect the Ass such that no AS remains isolated and that there exist at least two different paths to interconnect any two of the three Ass. Define Edge Routers for each AS accordingly. c) Which routing protocol is adequate for the routing between Ass? Give reasons for the idea of this routing protocol. d) Which routing protocols can be applied? D1) within an AS? D2) between Ass? e) Which information is exchanged between Ass and how is this accomplished? Part 2 Interworking between IPv6- and IPv4-Networks As the IPv4 address space will soon be exhausted a strategy has to be developed for the interworking between IPv6 islands across the IPv4-network. In the following the case outlined in Figure 2 will be studied: N3 IPv4 N1 IPv6 ER 1 H1 N2 IPv6 ER 2 Figure 2: Interworking between IPv4 and IPv6 Hosts H1 and H2 are connected to IPv6-networks N1 and N2 which are interconnected by the IPv4-network N3 through Edge Routers ER1 and ER2, respectively. The Edge Routers perform functions of a Gateway between the different network types. ER1 and ER2 host the complete protocol stacks for IPv4 and IPv6. H1 and H2 should be able to communicate between each other exclusively through IPv6-packets and should not even be aware of trans-passing the IPv4-network N3. Question 1 Protocol Architectures Sketch the principal layered protocol architectures of N1(N2) and N3 by a) their protocol stacks b) their packet formats. Three principal methods are considered for the exchange of IPv6-packets between H1 and H2 across N3: - Tunnelling - Packet Header Translation - Dual Protocol Stacks. H2 Question 2 Tunnelling a) Describe the method of Tunnelling by a sketch of the packet structure used within N3 for the considered communication between H1 and H2. b) Are there any changes required in the Routing Tables of the Edge Routers ? Give reasons for your answer. c) Are there any effects on the transport layer protocols UDP and TCP running within H1 and H2 ? Give reasons for your answer shortly. Question 3 Packet Header Translation By 'Packet Header Translation' IPv6-packets will be translated packet-by packet into IPv4-packets to be routed and forwarded through N3. a) Identify the principal problems arising with this method. Hint: Consider the main elements of the packet headers for IPv4 and IPv6. Discuss which functions are candidates for a translation and, as a result of it, whether there exists a full equivalence at all. b) Discuss shortly the problem of routing and forwarding of packets. Question 4 Dual Protocol Stacks By 'Dual Protocol Stacks' all routers in the networks would have to be upgraded such that they can handle both types of packets and protocols. With the increasing degree of maturity of the change from IPv4 to IPv6 the IPv4 stacks could be taken out of operation. A simple looking, but more expensive, solution would be a complete duplication of all routers in hardware and software, i.e. a one-to-one overlay of IPv4 and IPv6 networks. a) Discuss this solution with respect to the assignment of IP addresses. b) At which layers can both networks be integrated and at which layer do they have to be separated ? Problem 2 Flow, Congestion, Traffic and QoS Control In this problem main control aspects will be studied which are used in the internet and other services integrated networks. Note: Questions 2,3,4 and 5 can largely be solved independently of each other. Question 1 Purposes of Control Actions a) Define shortly the purpose of a1) Flow Control (FC) a2) Congestion Control (CC) a3) Traffic Control (TC) a4) Quality of Service Control (QoSC) b) Identify b1) the instances (such as source host, destination host, network) being involved in the respective control actions b2) the criteria upon which the control action is initiated. Question 2 Mechanisms for Flow Control a) Name principally different mechanisms used for FC actions. b) Which FC mechanisms are used in TCP ? Identify the parameters used in TCP for FC and illustrate the control action by a generic temporal signalling diagram. c) To utilize network resources more effectively ISP's use Load-Balancing mechanisms by which the packet load is more evenly distributed across various links. Discuss the effect of Load Balancing on Flow Control. Question 3 Mechanisms for Congestion Control a) Name principally different mechanisms used for CC actions and classify them with respect to the identified instances (c.f. Question 1, b1). b) Which CC mechanisms is used for TCP ? Identify the parameters used in TCP for CC and illustrate the control action by a generic state transition diagram and a temporal signalling diagram. c) Which other CC mechanisms are used where the network is involved? In which sense does 'Load Balancing' belong to CC? Question 4 Mechanisms for Traffic Control a) Name principally different mechanisms for TC actions. b) Which TC mechanisms are used in: b1) ATM networks b2) the internet ? Identify the parameters which are applied in these TC methods. c) Example: A traffic stream of cells (ATM network) or of packets (internet) shall be controlled with respect to the conformance with parameters of a "Traffic Contract" (i.e., quantitive metrics for the peak cell or peak packet rate, and a limited tolerance with respect to some "burstiness"). Which algorithms are used: c1) for ATM? c2) for the Internet? Give the name and the parameters of both algorithms. Question 5 Mechanisms for Quality of Service Control a) Which parameters characterize Quality of Service on the network layer in a1) ATM networks? a2) The Internet? b) Name principally different mechanisms/protocols to enhance QoS and which are applied in a1) ATM networks a2) The Internet. Describe these mechanisms shortly.