Mid-Term Exam Solutions Organizational Communications and Technologies Course Number: 90702 Fall 1999 Question 1 ========== a) Increasing importance of wide area networking Combining WAN and LAN technologies Support for multi-protocol based networks Multiple topology networks Application Interface Requirements b) Mainframe Networks ------------------ Mainframe based networks were centralized and had few computer to computer integration requirements. Terminal devices that were connected were passive devices (dumb terminals). Conformed to a star topology. PC Networks ----------- First networks supported simple file and resource sharing (printers). Move towards sharing of files, printers and other peripherals. Support for distributed processing. Support for bus or ring topologies. Intelligent workstations. 2) a) Digital encoding Sampling is the basis for digital encoding. Main advantages of digital encoding is the increased immunity to noise and the ability to support more flexible equipment. b) Three kinds of digital encoding are: Manchester encoding Diffferential manchester encoding Non return to zero Non return to zero inverted Alternate mark inversion c) Handshaking methods are: Synchronous and Asynchronous Synchronous handshaking uses a clock to coordinate the movement of bits through the system and so no start or stop bits are required. Asynchronous handshaking uses a frame to indicate the beginning and end of each piece of data. d) Simplex defines one way communication from sender to receiver. Half Duplex defines bi-directional communication but only one way at a time. Full duplex defines bi-directional communication simultaneously. e) (i) Parity is one method of error checking. Relies on a basic calculation of the number of '0's and '1's in a transmission unit. CRC is another method of error checking. It involves dividing the total transmission by a pre-determined prime number and the remainder is the CRC value. CRC can be used with a larger units of data (blocks or frames). Parity checking can be ambiguous. Parity checking checks one byte at a time. CRC is more reliable. (ii) Retransmission is a universal error recovery mechanism. Retransmission occurs if the receiver does not send an ACK. 3) OSI Model (a) and (b) Application - provides interface between end users. Presentation - performs protocol conversion, data encryption and decryption. Session - establishes and terminates data streams between nodes. Transport - ensures that session connections are transparent and handles details of data transfer. Network - provides routing mechanisms between nodes on a network. DataLink - defines the access method for connection with network. Physical - responsible for bit stream error checking and transmission. c) ARP is reponsible for mapping an IP address to a hardware address. d) Client/Server Strengths Central file System. Communication can be managed by central server. Client communication software is simple. Client/Server Weaknesses Single point of failure. Communication overhead because all communication goes through central server. Network OS could be overly complex. Peer-to-peer Strengts No single point of failure. More flexibility for storing distributed data and sharing remote peripherals. Availability of more API for development of distributed applications. Peer-to-peer Weaknesses Host software must be more self sufficient requiring it to be more complex than client software. Absence of central monitor to manage network communication. Lack of authority mechanism to resolve contention issues that may arise during operation. 5) Token Ring (i) Fundamental to the operation of Token Ring is data exchange. A given node can only transmit when it has possession of the token. A token is created and travels throughout the network. Each node gets a chance to transmit and does so if there is no data currently being transmitted. If there is nothing to transmit by a given node then the token is passed to the next node. The node seizing the token, sets the token bit and expands the token to add the frame to the end. The token then is transmitted and data is copied at the destination node. The destination node does not remove the frame but copies the token. The token circles till it reaches the source and the source can reconstruct the token with a new data frame. If there is no more data to transmit then the available bit is set. (i) Domain Name Service provides a mapping between a machine's hostname and its IP address. It avoids the need to have static host tables containing this information. Static tables are inefficient with large numbers of machines on a network. (ii) Internet Control and Messaging Protocol is used with "ping". c) SAPs or Service Access Points serve as entry points to the protocol layers above and/or below a given layer in the seven layer TCP/IP model. (Diagram can be found in class notes) 5) (a) Bridges act as address filters and they operate at the MAC layer. They connect networks of the same type. Routers are used to connect dissimilar networks. They operate at the network layer. Repeaters permit multiple lan segments to be connected at the physical layer. Gateways permit the inclusion of OSI and non-OSI protocol models and use all seven layers of the protocol stack. (Diagram can be found in the class notes Lecture8). (b) Operational steps of DNS can be found in the lecture notes (Lecture9).