Application of Polling Models to Computer Networks - Assignment Example

PROBLEM SOLVING REPORT 2 Student’s Name Class Instructor Institution of Affiliation City Date Answer #1(a) Polling can be used in routers where the router (controlling device) repeatedly checks the state of the external device (low-level hardware) its forwarding packets to whether it’s ready by reading the memory addresses through the use of memory mapped I/O. Polling can also be used with busy-wait polling where when the receiving device is ready, the router then forwards packets to it otherwise the router performs other tasks. Polling in networks has its own advantage in that it is efficient in applications where the inbound data is a lot as in the case of a router receiving more traffic than it is able to forward. The disadvantages of polling in networks is that the network becomes unpredictable and undependable causing polling to occur at all times. Secondly, in case the rate of polling is low trouble with latency might happen. Lastly, polling can lead to wastage of CPU cycles if no data is present when polling (Takagi, H. 1991, 193). Answer #1(b) In a Token Bus Network every device has knowledge of its succeeding devices network address. The devices are interconnected to the mainframe by a copper coaxial cable (communication bus) where the Token Bus protocol creates a token that manages a devices access to the communication bus. Any device that has the token packet is granted permission to transmit data. When done transmitting or when the transmission time specified by the network expires, the device sends the token packet to the successor device in the logical ring. This ensures that no two devices are transmitting at the same time and also gives each device time to transmit (Park, H. Ahn, S. and Kwon, W. 1993, 412). In Token Bus networks as the transmission speed increases, the network efficiency drops. This is because the communication medium, a coaxial cable has a low transfer rate of 10 Mbps which leads to a reduced data transfer rate thus requiring more time for transmission compared to the Ethernet cable with a transfer speed of 100Mbps. Answer #1(c) (i) The nodes employ technology for media access control in order to be able to transmit simultaneously on the same bus. The bus has only one channel for receiving and sending signals. However it does not support receiving and sending of signals at the same time. Therefore, if two nodes try to simultaneously send data collision happens since the shared media is half-duplex. Increased collision results to a network of low quality. A bus topology is considered a broadcast network because when one node transmits a signal, the signal travels in both directions of the bus to every node in the topology. The bus must therefore handle all the signals coming from the various nodes connected to it. However, because only one signal can be sent or received at one time, it limits the data exchange rates by decreasing the available bandwidth. Answer #1(c) (ii) Switches help in reducing collisions where they are employed to divide big collision domain to small collision domains causing every port in the switch operates in a different collision domain. The switch also forwards broadcast packets to every port of the switch where each switch floods all ports with the broadcast packet. A bridge also assisting in reducing collision by dividing the one collision domain to many small collision domains, buffering and forwarding frames. However, devices connected to switches and bridges belong to one broadcast domain. A repeater addresses attenuation where upon receiving a signal it rebuilds it and sends it out. A repeater extends broadcast and collision domains. Hubs extend collision and broadcasts where upon receiving a data packet at one port, the hub transmits that packet to its other ports. Devices connected to a hub belong to one broadcast and collision domain. Thus, if two devices connected to the same network send data packets simultaneously, collision occurs. Answer #2(a) IPv4 checksum is an error-detection scheme that is used to protect IPv4 data packets header from data corruption. Only the header of the datagram is checked using this checksum. This is because, when sending a message, the sender calculates the datagrams checksum and places it on the header of the data packet and then sends the calculated value with the message. At the destination, the receiver calculates the received data packets header checksum again and cross checks to ensure that the accompanying value is identical. In case the numeric value is not the same, then the data packet has been corrupted. Cyclic redundancy checks (CRC) are employed at the link layer for the purpose of detecting more types of packet errors. Despite being a sophisticated error checking mechanism, CRC is not appropriate for protecting data packets from intentional alteration because CRC is a linear function and does not employ authentication thus an attacker can manipulate the message and recomputed the CRC making it impossible to detect the substitution. Secondly, the CRC function is easily reversible thus it’s unfit for use in digital signatures. However in the case of IPv4 checksum, any corruption of the data packets is identifiable. Answer #2(b) Subnet mask = 255.255.192.0 = 11111111 11111111 11000000 00000000 IP address = 165.131. 121. 111=10100101 100000011 01111001 01101111 10100101 10000011 01111001 01101111 (IP) 11111111 11111111 11000000 00000000 (mask) 10100101 10000011 01000000 00000000 165.131.64.0 (Network address) range= compliment of 11111111 11111111 11000000 00000000= 00000000 00000000 00111111 11111111 = 63.255 IP address range =165.131.64.0 to 165.131.127.255 Answer #2(c) The router segments the packets into transmittable fragments with the maximum fragment size been MTU (Maximum Transmission Unit) minus the IP header size. The router then puts every fragment with its header checksum recomputed and fragment offset field set into separate packets and then transmits them. At the destination, the receivers identifies a packet as a fragment if the flag for more fragments is set and the field for fragment offset is nonzero. The receiver using local and foreign internet address, identification field and protocol ID identifies the fragments and reassembles data from the fragments having the same ID employing both more fragments flag and fragment offset. Upon receiving all fragments the receiver calculates original data payload length, puts all the fragments in the appropriate order and passes the fragments data for further processing up the stack. IPv4 fragmentation contributes to congestion in that upon packet fragmentation, the many fragments leads to more packet headers overhead, fragments reassembly at the receiving node and additional routing decisions. This creates a multiple tasks from one task of transmitting only one packet to the destination which leads to congestion. Secondly, during reassembly of fragments, if one fragment is identified as missing the error checking mechanism requests the entire fragments to be sent again. This creates more traffic and leads to congestion. Also, in case the Internet gateways get congested, they will also drop fragments leading to missing fragments at the receiving node hence more retransmissions thus creating more congestion. Answer #2(d) (i) This protocol combines ARQ error-control and high-rate forward error-correcting coding. There is an addition of redundant bits to the data been transmitted occurs where error detecting code, in this case cyclic redundancy check is employed. At the receiving end, the receiving node upon detecting a corrupted datagram requests the sending node to retransmit a new datagram and the corrupt datagram is discarded. Answer #2(d) (ii) Datagram ACK Datagram Datagram corrupt NAK Retransmission Answer #2(d) (iii) The advantages of this protocol include: It is a faster way of transmission since as the corrupt datagram is discarded and the receiving node awaits a retransmission or the datagram is acknowledged, the sending node can send other data and the receiving node can receive data in the same time interval. Secondly, upon arrival of a corrupt datagram the receiving node sends a retransmission request for the same datagram automatically until an error free datagram is retransmitted. The disadvantages of this protocol is that in order for the datagram transmission to pass error detection any given communication should be received with no errors. Secondly, an increased number of packet retransmissions lowers the data flow performance. Answer #3(a) RIPv2 is usually used in small networks while OSPFv2 is employed by larger networks. The information contained in RIPv2 updates includes regularly updated information from the routers describing each routers routing database to all the routers it is connected to directly and an update algorithm (Malkin, G. 1994, 16). The OSPFv2 update contains a topological database containing all connections in a local network and an algorithm that enables the routers to update their link state regularly that explains each routers routing database to other routers (Baker, F. and Coltun, R. 1995, 21). RIPv2 decisions are said to come from second hand information because it keeps track of the router that is closest to each destination address while OSPFv2 decisions are said to come from first-hand information because it keeps complete track of all the connections topological database of a local network. Answer #3(b) M 0 {C} 1 ∞ 0 ∞ 2 ∞ 1 {C,A} 1 ∞ 0 ∞ 2 2 2 {C,A,F} 1 8 0 4 2 2 3 {C,A,F,D} 1 8 0 4 2 2 Routing table Destination Next hop A 1 F 2 D 2 Answer #3(c) (i) Header length (HLEN) is the internet header length (32-bit words number). The IPv4 header length field specifies the header size that also corresponds with the data offset. The total length field is a 16-bit field that determines the whole packet size in bytes. In IPv6 payload length indicates the IPv6 payload length and the field’s size is 16 bits. All these fields are related in that the IPv6 is a version of IPv4 header where rarely used fields in the IPv4 header were eliminated and fields providing better support were added for better real-time traffic support. Answer #3(c) (ii) In IPv4 the protocol field determines the protocols used in the IP datagram data portion while the IPv6 next header field betokens either the upper layer PDU protocol (TCP, UDP) or the extension header with no predecessor. The two are related in that in the IPv6 indication of upper layer protocol, similar values employed in IPv4 (when indicating the upper layer protocol) field are also employed here. This protocols used in IPv4 and IPv6 are the same and have not changed. Answer #3(c) (iii) In IPv4, Time To Live (TTL) field is 8 bits and assists prevent datagrams from going in circles (persisting) on an internet. It limits the lifetime of a datagram. In IPv6, the 8 bits hop limit field betokens the IPv6 packets maximum number of connections beyond which it can travel before it is discarded. The two fields are similar except the lack of historic relation to the time in which the packet is lined at the routers queue. References Baker, F. and Coltun, R., 1995. OSPF version 2 management information base (No. RFC 1850). Malkin, G., 1994. Routing Information Protocol (RIP) Version 2 Carrying Additional Information. Request for comments (RFC 1723) . Park, H.S., Ahn, S.C. and Kwon, W.H., 1993. Performance and parameter region for real-time use in IEEE 802.4 token bus network. IEEE Transactions on Industrial Electronics, 40(4), pp.412-420 Takagi, H., 1991. Application of polling models to computer networks. Computer Networks and ISDN systems , 22(3), pp.193-211. Read More