Data Link Control (Internet)

Society has become based solely on the ability to move large amounts of information across vast distances quickly. The natural evolution of computers and this need for ultra-fast communications has caused a global network of interconnected computers to develop. This global network allows a person to send E-mail across the world in mere fractions of a second, and enables even the common person to access information world-wide. With the new advancements in technology there must be a set of "rules" or better known as protocols that must be established and utilized at all times. In this short ten page paper I will be demonstrating the advancements in these protocols and there uses today.

To properly show the significant advancement, it will be best to show why Data Link Control was established. In the early 1970's, the U.S. Defense Advanced Research Projects Agency (DARPA) started a research program interlinking computer to share information. While sending information from one site to the other many problems arose with losing

data (Society). To decrease the amount of corrupted data being transmitted, protocols were established. These protocols were a drawn out process that was very slow but was able to transfer data all across the world.

By 1986, the US National Science Foundation Started the NSFNET which today provides one of the biggest backbones for the internet. This supercomputer was able to send packets on its 45 MBps trunk to different locations. Once this was in place the internet was born with TCP/IP Protocols of TCP/IP protocol suite became available in the 1980's. . By 1991-93 Home computers were starting to take advantage of the vast amount of information that is available. By this time the OSI protocol was created and by the end of 1991 the internet has grown to include 5,000 networks in three countries, serving over 700,000 host computers used by over 4,000,000 people. This was all possible due to strict sets of protocols that were followed (Society). By the mid to late 1990's society was using 56K modems in the residential areas and companies were purchasing faster dedicated connections. At this period of time flow control, error control, and High-level Data link Control (HDLC) were being implemented.

The control of the data being processed is referred to as flow control. Flow control was needed to be established to regulate the speed of data being transmitted. Regulating the speed of the transmission evens out the data so that very little errors will occur. There are two types of flow control, Stop-and-Wait and Sliding-Window. Just like any other advancements in technology we get better every day. Most information here-in are information found in Data & Computer Communications, our not so up-to-date book.

Stop-and-Wait flow control, the simplest form, will send one packet of data and wait for a response from the destination node that the data was received and then will send another packet of data until all packets are sent. The destination node can stop the flow of data by simply not responding that the packet was received (Williams 195). This has a very low error rate but unfortunately it comes at the costs of slow speeds. With the size of LAN's that we have now this will not be adequate to take up the resources of the network for and extended amount of time.

To best why to explain this process is to think about a slide at a park. There is a line of kids wanting to play on the slide but only one child should go down the slide at a time. When the kid is at the top of the slide they wait for the slide to be open and the previous child to move out of the way. The child once at the bottom moves out of the way signaling to the next child to go. This will go one usually with much success but every once in a while you will have a child not move out of the way at the bottom and the next child will have to wait for the other child to move. The same goes with Stop-and-Wait flow control. One a packet of data is sent by the sending station waits for an acknowledgment from the receiving station. Once the acknowledgement signal is received another packet is sent. This will carry on till all packets are sent. If the sending station does not receive a signal back the transmission is stopped.

The other method that the book talks about is Sliding-Window flow control. Sliding window flow control utilizes the efficiency by allowing multiple frames to be in transit at the same time. The best way to show this is to think about the slide again but to think about a longer slide. Kids are not spaced out like Stop-and-Wait. Kids are sent right after another. It becomes important that the kids move out of the way once at the bottom. Like the slide theory Sliding-Window flow control will work in the same way. If something happens in the transmission from one computer to the other, the rest of the packets are damaged till it gets resolved (Williams 198).

The packets being sent will have an identification telling the destination what has been sent and what still needs to be sent. Just like this is packet two out of twelve. The destination will only send an acknowledgement when the transmission is done or if a packet is missing. If a packet is missing then only that packet will be resent. With this advance in technology the speed of transmissions were greatly accelerated with accurate results. The numbers to assure all packets are received are called sequence numbers. Because the sequence numbers to be used occupies a field in the packet, it makes the data in the packets smaller. The sequence number may be in the eighth frame of the packet meaning the first seven frames are buffer frames of useless data. The downside to this is that the computer has to process all the frames in the packets. If a packet has a large amount of buffer data it can slow down a network. Other technologies have come about since this but we will talk about them later.

Now that we know how the data is being formatted to be transmitted between stations, let's talk about error detection. Error detection is a very complicated mathematical process that checks certain sections of the packets. The first type and simplest form of error-detection system is to use a parity bit at the end of every packet. There are two types of parity, even and odd parity. An example of an even parity is when this frame of 11100010. Since this frame has four ones, even number of ones, it has even parity. In this case the parity bit is the last bit of zero in the frame. If I switched the last bit to 1 making the frame 11100011, I increased the number of ones to five making it an odd parity. This process is not fool proof as noise can change one bit in a frame very easily.

The other type of error detection is Cyclic Redundancy Check (CRC). This is the most powerful error