Network Protocol Services - Term Paper Example

Network Protocol Services Abstract Computer networks constitute a major aspect of the global economy in terms of communication, operation, efficiency and reliability. The world has become a global village because of computer interconnectivity. The interconnection of computers to form both local area networks (LANs) and Wide Area Networks (WANs) is an intricate process involving network hardware, network software, operating systems, and user applications. The numerous differences in computers, network devices, operating systems, languages, and so on provide a major challenge to an effective networking environment. The Javvin Company (2013) describes the open systems interconnection (OSI) model as the standard benchmark for communication between different hardware and software architectures. The OSI model layers network communication into a seven-fold process, where each layer attends to a specific task. These layers are application, presentation, session, transport, network, data and physical layers. Layers 1 to 3 control communication between network devices while layers 4 to 7 attend to data source and destination communication. According to Oracle (2010), communication within a network involves the movement of data packets or frames from the source device to the receiving device. A network packet consists of both the source and destination address and a payload. Citap (2013) explains that a frame consists of five fields i.e. a start indicator, the source address, the destination address, control, data, and error control. The devices communicating within a network adopt the name node, device, or station. Princeton University (2013) informs the data link layer performs local delivery of packets between nodes. The process results in several challenges such as collision of frames, loss of packets, discarding of packets, and so on, mostly without feedback to the sending device, thus leading to unreliable communication. A functional feedback system is thus crucial within a network. According to Princeton University (2013), the second layer of the OSI model named the data link layer serves both LAN and WAN nodes by transferring data between them. It hosts protocols such as Ethernet for a multi-node LAN, point-to-point protocol (PPP), HDLC and ADCCP for dual-node PPP connections. Princeton University (2013) explains that the data link layer has the mechanism to detect and correct errors that occur within the physical layer. This layer focuses on delivery of frames between the devices on the LAN, local addressing and device prioritization and allocation, while leaving routers and other devices to perform global addressing. The requests from different network devices to use the same medium simultaneously results to frame collisions. Data-link-layer protocols detect and resolve such collisions while at the same time laying down the mechanisms to prevent or reduce them in the future. Princeton University (2013) informs the data link layer transmits packets by applying specific hardware addressing to frames. The headers of these frames indicate the sending device and the intended recipient. However, despite the good intentions of this layer, some data link protocols lack ability to give feedback on the successful transmission and acceptance of a packet. The use of higher-level protocols helps to solve this problem by providing flow control, error checking, acknowledgement of well-received packets, and attempt or request re-transmission of failed packets. Examples of such high-level protocols include Transmisson Control Protocol (TCP), Trivial File Transfer Protocol (TFTP), High Level Link Control (HDLC) ABM, and so on (Fairhurst, 2001). Flow control gives the receiving node the ability to control the packet transmission speed, allowing both slow and fast senders’ packets to reach the recipient smoothly without interfering with each other or clogging the system. Humphrys (2013) explains that this is especially a problem if the receiver is much slower than the sender is. There are two types of flow control according to Humphrys (2013) namely feedback based and rate-based flow control. Feedback-based flow control takes place in the data link layer where the receiving devices dictates to the sender the speed and the duration that the sender can transmit. The protocols that have in-built data-rate controls in higher layers of the OSI model Rate-based flow control is managed by. According to Fairhurst (2001), the HDLC protocol operates within the data link layer but uses the services of the physical layer to provide a reliable transmission path that provides acknowledgement for successfully received frames. He informs HDLC encapsulates or encloses each protocol data unit (PDU) due for transmission within a header and a trailer. The header comprises of a HDLC address and control field, while the trailer contains the CRC. Fairhurst (2012) describes the CRC as checksum for checking the integrity of frames. The data-link layer protocol computes it and appends it to a frame before transmission. The receiver reverses this operation and compares the two values. Inconsistency results in a CRC error. This way the HDLC trailer manages to detect errors in transmission. Microsoft (2013) emphasizes that the data link layer ensures the error-free transmission of packets between nodes through the following techniques; 1) It establishing the link between nodes with imminent communication and severs the same when communication ends. 2) It forces transmitting nodes to pause or look for alternative when frame buffers are unavailable. 3) It enables sequential transmission and receipt of packets thus ensuring priority. 4) It provides frame acknowledgement from successfully transmitted frames, and expects the same from other nodes. When none is forthcoming, it retransmits the failed frame. 5) It checks integrity of received frames, especially using check-sums. 6) It performs media management by prioritizing node access to the media. Fairhurst (2001) describes four parameters which best indicate that frames reached their destination successfully. They include; a) The packet must not be lost in part or in full. b) The node must send only one copy of the packet. c) The priority of sending should reflect in the receipt i.e. first in first out (FIFO). d) Recipients receive the frame within a reasonable period. To effective address these issue, each protocol must include an error-detection and recovery procedure. Fairhurst (2001) recommends the Stop-and-Wait ARQ. In this technique, a stop-and-wait protocol begins by sending a protocol data unit (PDU) or data block, and then pauses until it receives confirmation that the PDU reached the destination successfully. When the receiver gets the PDU successfully, it sends a positive acknowledgement (ACK), but if it does not receive the PDU, it sends a negative acknowledgement (NACK). The sender remains idle until ACK arrives, after which it resumes transmission of the next PDU. There are situations where the receiver is unable to determine that a PDU arrived successfully, and therefore it delays in sending back an ACK or it does not send it at all. To correct this, Fairhurst (2001) explains that the stop-and-wait ARQ protocols implement a timer that checks that ACK arrives within reasonable time. Delay leads to a timeout and the sender begins to retransmit the frame after the timeout. The following illustrations extracted from Fairhurst (2001) depict the ACK-NACK process more clearly; The other system that produces reliable communication within a LAN is the Ethernet. This is one f the standards produced by the Institute of Electrical and Electronic Engineers (IEEE 802) in 1985 (Fairhurst, 2009), for the purposes of LAN communications. He explains that in an Ethernet, the Medium Access Control (MAC) protocol forms data packets or frames while exploting Manchester Line Encoding (MLE). MAC provides the data link layer in a LAN. MAC provides more reliability in a LAN by encapsulating the payload data by pre-fixing a header called Protocol Control Information (PCI) and appending a trailer holding an integrity checksum or CRC data. According to Fairhurst (2009), the structure of a MAC consists of a preamble, header CRC, and an inter-frame gap (IFG). Besides the MAC, an Ethernet also exploits the Carrier-Sense Multiple Access Protocol with Collision Detection (CSMA/CD) which ensures that computers that share a communication medium do not transmit at the same time resulting in frame collision. In the event that this happens, CSMA/CD retransmits the corrupted frames. Again, Fairhurst (2004) explains that to avoid collisions, the ALOHA protocol determines which NIC will transmit at any given time. Aloha directs each NIC to append a CRC or checksum to its transmission sequence thus giving the receiver the ability to verify the integrity of the received frame. This is very important in enabling effective communication, although it is really no guarantee that destination nodes receive packet correctly. ALOHA depends on the ARQ protocols to retransmit corrupted frames. CSMA is an improved version of ALOHA (Fairhurst, 2004). He explains that with CSMA, a NIC first “listens” to the transmission medium (cable) using a transceiver to determine if the cable is in use. The presence of a current confirms this. The NIC begins transmission only if the cable is idle. Idle NICs on the other hand “listen” to determine if any frames are destined for them. This method works only if two NICs do not attempt transmission simultaneously, an event that leads to collisions. Collision Detection (CD) comes in to correct this situation. According to Fairhurst (2004), by checking for excessive current, a NIC determines a collision has occurred, since is current combined with the current of the other transmitting NIC will produce excess. It ceases further transmission and instead transmits a 32-bit jam sequence causing the other NICs to receive this sequence in place of the corrupted frame. The sequence does not have the MAC CRC and therefore the NICs discard it due to a CRC error. The other transmitting NICs respond in similar fashion discarding the corrupted frames thus paving way for correct retransmission. According to Fairhurst (2004), in a shared medium, all nodes can receive the frame; however, the MAC header holds an address for the destined node. Only this node can forward the packet towards the destinations. The network interface card (NIC) whose address exists on the MAC header of a frame recognizes the address as valid and safe for forwarding. All the other nodes that receive the frame on the shared medium simply discard it. TechFest (2007), an IT technical information website suggests Frame Relay as an error-free system designed to provide reliability and speed. The Frame Relay service overrides the network layer of the OSI model while performing both routing and multiplexing. TechFest (2007) explains that Frame Relay is a data link protocol for handling large volumes of data with phenomenal speed, a critical factor of modern day networks. It performs three functions; 1. It provides multiple accesses to the network. 2. It structures frames, prioritizes them, and delivers them in the correct order. 3. It recognizes transmission errors using the standard CRC. Interestingly it ignores some traditional routines such as frame acknowledgement, re-transmission and so on. In fact, it eliminates the use of error-handling and flow-control routines commonly used by many protocols, although it utilizes other higher-level protocols in higher layers of the OSI model to recover from errors and congestion-related issues. Frames Relay does not deliver faulty packets; instead, it discards them if a subscriber exceeds the network load, if a CRC indicates a physical transmission error, or if there is network congestion. The higher-layer protocols will detect that frames were discarded using rotating sequence numbers. According to TechFest (2007), these protocols also use throttling and flow-control techniques to detect and control network congestion. When a device receives a frame with a sequence number that is not in order, it requests the sender to re-transmit all the frames in that block. Another very important and powerful aspect of Frame Relay is that it has the capability to inform subscribers of network congestion. TechFest (2007) informs it does this by attaching two Explicit Congestion Notification bits in the frame header. A value of ‘1’ in the Forward Explicit Congestion Notification (FECN) header bit indicates congestion on the path of the sent frame, while a value of ‘1’ in the Backward Explicit Congestion Notification (BECN) header bit indicates that the opposite-direction path of the sent frame is congested. FECN and BECN direct the subscriber either to reduce transmission rate drastically or to withhold traffic completely until the congestion clears. The use of Transmission Control Protocol also provides a reliable method of local delivery. The functionality of TCP provides facilities that ensure this. According to Kristoff (2000), TCP is connection-oriented meaning that the two devices that wish to communicate must agree on the type, time, channel and method of communication first before any communication takes place. It does this through a technique known as the three-way handshake. This preempts errors such as collision, congestion, and addressing. The following illustration extracted from Kristoff (2000, Fig. 2) demonstrates the three-way handshake; He also explains that TCP offers reliability because each TCP frame must carry a checksum, which the receiver evaluates to detect any error. Again, TCP keeps a record of the received steams of bits and therefore it is able to discard duplicate frames. Kristoff (2000) also informs TCP has the capability to arrange frames sequentially in the correct order even if the packets were not in order at receipt, and then it forwards them in the correct sequence. Finally, TCP corrects loss of packets with positive acknowledgements by making requests fro retransmission from the sender. Even in the absence of positive acknowledgements, it applies timers so that if a positive acknowledgement does not suffice within a reasonable period, TP makes a request for retransmission to the sender. References Citap. (2013). The Seven Layer Open Interconnection Reference Model. Retrieved September 19, 2013, from http://www.citap.com/documents/tcp-ip/tcpip006.htm Fairhurst, G. University of Aberdeen – School of Engineering. (2009a). Carrier Sense Multiple Access with Collision Detection (CSMA/CD). Retrieved September 23, 2013, from http://www.erg.abdn.ac.uk/~gorry/eg3567/lan-pages/csma-cd.html Fairhurst, G. University of Aberdeen – School of Engineering. (2009b). Ethernet. Retrieved September 21, 2013, from http://www.erg.abdn.ac.uk/~gorry/eg3567/lan-pages/enet.html Fairhurst, G. University of Aberdeen – School of Engineering. (2001c). Reliability. Retrieved September 21, 2013, from http://www.erg.abdn.ac.uk/~gorry/eg3567/arq- pages/reliability.html Fairhurst, G. University of Aberdeen – School of Engineering. (2012d). High Level Link Control (HDLC) Protocol. Retrieved September 23, 2013, from http://www.erg.abdn.ac.uk/~gorry/eg3567/dl-pages/crc.html Javvin Company. (2013). OSI 7-layer Model for Network Communication. Retrieved September 19, 2013, from http://www.javvin.com/osimodel.html Kristoff, J. (2000, April 24). The Transmission Control Protocol. Tech Notes. DePaul University. Retrieved September 24, 2013, from, http://condor.depaul.edu/jkristof/technotes/tcp.html Oracle Corporation. (2010). System Administration Guide: IP Services. Retrieved September 19, 2013, from http://docs.oracle.com/cd/E19253-01/816-4554/ipov-29/index.html Princeton University. (2013). Data Link Layer. Retrieved September 19, 2013, from http://www.princeton.edu/~achaney/tmve/wiki100k/docs/Data_Link_Layer.html TechFest. (2007). Frame Relay Overview. Retrieved September 24, 2013, from, http://www.techfest.com/networking/wan/frrel.htm Read More