CSE 126/228F: Second Homework

CSE 126: Second Homework Solutions

1. (3 pts.) List 3 distinctive characteristics of multimedia data as compared to conventional text data, with regard to storage and retrieval.

each of the following is 1 point. 0.5 for the keyword and 0.5 for the description.

Continuity. Multimedia data is time-based media. Thus, audio and video make sense when sent a displayed at a certain rate. Most file servers save parts of a file in several different locations that may be far apart on the disk itself. This causes many head movements (seeks) that slow down transfer rates and the servers can not continuously send a multimedia stream. While text does not have to flow continuously, this is imperative for multimedia.
Most existing systems are optimized for small files such as word processor documents. Multimedia files tend to be very large in comparison and therefore require special treatment. POINTERS ARE USED TO EDIT/MANIPULATE THEM SINCE THEY ARE LARGE FILES.

Multiple media streams. A multimedia stream is made up of elementary streams such as audio and video. A video server needs to be able to distinguish video from audio from text, etc. and treat each of these different media in a special way. Each of these streams have to be synchronized in time during display.

2. (2 pts.) What is the need for buffering in multimedia data retrieval?

1 point for the correct answer(jitter). 1 point for the description.

Buffering is necessary in order to eliminate jitter, which is caused by the widely varying amount of space (OR ARRIVALTIME) between blocks. Prefetch or read-ahead is done before decoding begins. In other words, to retrieve enough data in advance AND FILL UP THE BUFFER so the decoder DOES NOT STARVE FOR LACK OF FRAMES TO DISPLAY.

3. [5 points] Describe the three levels of media integration in a network?

If any of the following is missing, deduct 2 points, upto 5 points. 1 for the keyword and 1 for the description.

The three levels of media integration in a network are as follows:

Integrated Access - ADSL and ISDN both provide integrated access. This is a type of system that provides a single access (e.g. a box on the wall) to several different networks. However, behind this box each network is transmitted independently and also switched independently.

Integrated Transmission - HDSL provides Integrated Transmission, the next step in multimedia integration. It provides integrated ac cess, and in additon transmits each of the various network streams along the same line. However, at some point the networks must be spearated and switched independently.

Integrated Switching - ATM provides Integrated Switching which is the highest level of integration. Integrated switching provides integrated access, transmission, and it switches all networks in the same manner.

THEY DON"T HAVE TO MENTION ADSL, HDSL.

4. [6 points] Compare the advantages and disadvantages of Cable Television, Telephone, and computer internet networks for delivering multimedia traffic.

each of the following is 2 point. 1 point for the advantage and 1 point for the disadvantage.

Cable Television networks have the advantage of being able to handle large amounts of bandwidth in one direction to the user. This is also an inherent disadvantage, as data often needs to be transmitted from the user rather then just to him THIS SENTENCE IS VERY CONFUSING. IT IS JUST THAT THE DISADVANTAGE IT IS BROADCAST - SO INDIVIDUAL USERS CANNOT BE ADDRESSED AND SENT ANY EMAIL. Thus, switching is a drawback in the sense that the network is designed for one-way BROADCAST communication as well as very low bandwidth on the back channel to the head end. Furthermore, since homes do not have individual network addresses, any type of signal must be broadcast to the entire network.

Telephone networks have the advantage of already being in just about every home. They already have individual addressing in the form of telephone numbers and they are a direct connection network that works equally well in both directions. Unfortunately the plain old telephone system (POTS) does not allow for a very high bandwidth and is mostly analog. As it was originally designed for voice (which is fairly low bandwidth) it does not support very well the high volume of data necessary for multimedia.

Computer internet networks come in many varieties. The two most commonly used today are Ethernet and ATM. Both networks use a form of packet switching, as opposed to the circuit switching employed by the telephone networks. Packet switching has the advantage that if the bandwidth required to send a packet to the next switch is not available, the packet is stored and forwarded to the switch as soon as possible. The problem with packet switching is that there are no performance guarantees available, packets can take highly variable amounts of time to be delivered. This is a major problem for multimedia. ATM uses statistical guarantees on performance to help solve this problem.

5. [4 points] The packet size in ATM networks is chosen to be a small constant value of 53 bytes. How does this help or hurt multimedia traffic?

The telephone companies wanted ATM to use a relatively small packet size since most of their transmissions would be low bandwidth voice with very strict bounds on delay in order to keep the transmission interactive. Because humans require a delay of less than around 150ms to feel that a conversation is interactive, by keeping to a small packet size they could minimize the bandwidth wasted on less then full packets. The advantage of small packet/cell size is that it takes less computational overhead to create and decode the packets, providing less end-to-end delay.

The computer industry generally wanted a larger size because computer data tends to be larger. Since each packet must use some space for the header, the computer industry argued that since most transmissions would necessarily span multiple packets, a larger packet size would minimize the wasted bandwidth of the header. Multimedia traffic tends to be extremely large, even larger than the computer industry was arguing about when the ATM standards were established. Since almost all multimedia will span many packets, it could be argued that a larger packet size would be of benefit. This would minimize the wasted bandwidth used for headers on so many small packets.

2 point for the answer. 2 point for the reason.

If the students have mentioned the following ideas:

"Arguments for small or big packet size:

small packet size is better since delay is smaller (for packetization) -e.g., voice.Plus no wastage of space when small amounts of data are transmitted (e.g.,in short messages/email, chat, etc.). disadvantage: packet header overhead.

large packet size is better since packet overhead is small, but delay in packetization, plus wastage of space if actually data size to be transmitted is small, are both disadvantages.so an in-between value was chosen. "

This should get full points.

Also the following answer should be fine.

" it hurt the multimedia traffic since Multimedia traffic tends to be extremely large and larger packet size would be of benefit. This would minimize the wasted bandwidth used for headers on many small packets.".