Video Transmission - Essay Example

Running head: Video Transmission - Wired Versus Wireless VIDEO TRANSMISSION - WIRED VERSUS WIRELESS Wired technology started gaining its rightful place in the 1900s with the advent of telephone networks. This involved the use of a metal conductor in the form of wires that carried electrical signals. The factors in favor of wired communication were high reliability, low price, durability, speed and quality of service. However, as in everything, this technology had some limitations in the form of the wires being affected by weather conditions, noise and electromagnetic radiations and their length being limited by the end points of communication. Wireless technology was used as early as 1950s for transmitting AM and FM signals to radio and television as well as for military intelligence. However its use had become widespread with the introduction of cellular phones to common people. The factors siding its use were the convenience of usage, possibility of high range of coverage and high life expectancy. However, the data transmission here was not very reliable, costs more, has lesser speed, can be easily intercepted and not very good quality of service. Video streaming can be done in different ways. The applications may be Video on Demand (VoD), real-time and near real-time video streaming and Multi Media Systems. The video streams can be multicast or unicast. The data in the video must be first encoded and the choice of encoding depends upon the codec used (i.e. MPEG-2, MPEG-4, H.264, AVI, WMV etc.), the target bit rate, the frame rate, quantization parameter, the resolution and so on. After being encoded the video is transmitted using a streaming server. The clients are expected to reply periodically, depending upon the application, to this server about how much data has been received, whether it has been interpreted correctly, whether there are any errors, whether the bit rate and frame rate is as required and many other details regarding the transmission. End-to-end delay plays a vital role. If a particular packet of the video stream arrives late at the client, it is bound to affect the quality. In some cases it is required that a frame should arrive within a particular time for the other frames to be correctly decoded. Intensive resource requirements are encountered due to the bursty nature of video traffic. If a downlink has been used and the packet size proves to be very large, it leads to saturation ultimately. Moreover if an 802.11b standard has been used in the implementation, it leads to contention problems since every client is expected to have equal opportunity in server access. The quality of the received video can be assessed only by comparing with the original video or through the QoS level parameters promised by the server. This is not so easy to be done in the case of real time transmissions. Also Carrier Sense Multiple Access / Collision Avoidance (CSMA/CA) used in many wireless standards may lead to slow transmissions that suffer numerous time lags. All the above can be taken as the issues to be addressed while using wireless technology in video transmissions. The wireless standards come under 802.xx category of IEEE standards. The 802.11 standard (for Local Area Networks) specifies the standard for interface between the client and the access point and also for the interface among the clients, in either case air being the medium. This standard is comparable with the 802.3 standard, which is for wired Ethernet LANs. This specification addresses both the physical layer and the MAC layer issues and also the compatibility problems. This entire standard specification makes use of CSMA/CA. A comparative study of wired network metrics and wireless network metrics based on the QoS is shown below. WIRED NETWORK METRICS WIRELESS NETWORK METRICS IMPACT OF AN UNSTABLE WIRELESS ENVIRONMENT ON MEASUREMENT Delay Delay High Jitter Jitter High Impact of rate adaptation on delay and jitter High Impact of roaming on delay and jitter High Impact of power management on delay and jitter High Impact of encryption on delay and jitter High Impact of MAC layer fragmentation on delay and jitter High The 802.11 standard supports a maximum bandwidth of 2 Mbps, which is not adequate for the present day applications. This has led to its expansion to 802.11b that supports up to 11 Mbps. It uses an unregulated range and hence experiences interference from various sources such as telephones, microwave ovens, baby monitors etc. Even though it might be affected due to interference from other electro magnetic transmissions, it can be overcome by maintaining the distance from such sources. Hence the two main advantages of this expansion are very low cost and difficulty in obstruction. However there are the limitations of low speed and hindrance to and from other appliances. Another expansion 802.11a developed in parallel to 802.11b supports a bandwidth of 54 Mbps. It used a regulated range of frequencies and hence experienced lesser interference. However it was 802.11b that became popular soon and because of this reason it caters to the needs of the business market. 802.11a has the limitation that it finds penetration of walls and other obstructions a difficult job and hence serves the home market. It is also expensive. The expansion 802.11g that came later tried to combine the features of both 802.11a and 802.11b. This too supports a bandwidth of 54 Mbps and it is backward compatible with 802.11b. This means that it makes use of the access point that will work with the network adapters of the latter too. This idea works if set up properly giving reasonable speed performance. Hence it has the advantages of good speed and good signal range but it is expensive than 802.11b and also suffers the problem of interference. The newest standard designed to be an improvement over 802.11g is 802.11n. It uses MIMO technology (i.e.) multiple signals and antennas instead of one to increase the amount of bandwidth supported but these multiple signals may interfere with other signals. It is expected to support over 100 Mbps and to be backward compatible. Hence it has best signal strength and more resistance to interference and obstruction. However it has not yet been finalized and costs more than its predecessors. It may also interfere with other-standard-based networks. Other standards such as 802.11h and 802.11j are radicals for specific purposes. All the above technologies are collectively termed as Wi-Fi technologies. Best effort data control and admission methods are available to ensure quality of service in 802.11e standard. The nodes listen to the available bandwidth in the link from the access point and based on this, decide whether or not to accept a video stream. Adaptive QoS can be provided by using priority queuing at the network layer level and establishing a retry adaptation limit at the link layer level. The IEEE 802.15 standard, designed for Personal Area Networks, specifies for low complexity and low power consumption. The substandard 802.15.1 is about Bluetooth whereas 802.15.2 is recommended for the coexistence of unlicensed bands. There are also sub standards in this category that allow the choice of higher rate alternative, low rate alternative and multimedia and digital imaging content. 802.15.3 is for the use of multimedia and digital imaging. 802.15.3a is to provide a higher rate alternative. 802.15.4 addresses the automation needs with 802.15.4a that allows a low rate alternative and 802.15.4b that provides for enhancements and clarifications. The IEEE 802.16 standard is a specification for the access of fixed broadband using wireless technology. This allows the use of innovative products by multiple vendors in the field of broadband. The IEEE 802.20 standard is a specification for Mobile Broadband Wireless Access (MBWA) with an interface that is optimized for IP based transport service. The Working Group endeavors to make it an always-on, affordable technology which satisfies both industrial as well as individual requirements when used in a multiple vendor broadband product. CSMA/CA is a modified version of CSMA. In CSMA, to use the common channel, a station has to first check it for a specific amount of time. If the channel is not used by anyone else during this period, then it goes for transmission of data in that channel. If not, it defers its transmission. CSMA/CD, where CD stands for Collision Detection, a jam signal is used to notify a collision when one station detects another signal other than its own in the channel. After this signal, the collision is avoided further by using exponential back-off algorithm that involves the use of random time intervals. CSMA/CD cannot be used in all cases. One of the reasons is that it is a media-dependent mechanism. Also in case of wireless media, it is not possible to monitor the channel as well as use it for data transmission simultaneously and this is he second reason. The next reason is the hidden terminal problem. Suppose there is a node A, which is in the range of receiver R but not in the range of the sender S. Node A cannot know that S is sending information to R and hence may, at the same instant of time, try to send its data too. In those cases where such a situation arises, CSMA/CA is used. CA stands for Collision Avoidance. This mechanism involves deferring transmission, when the channel is sensed to be busy, by a random interval. In case of video transmission, since it is a real time application that uses the data in most cases, it is not advisable to just defer sending the packet of data to avoid collision. This might lead to some problems related to QoS. The use of random intervals and exponential back off poses certain problems in video transmission. Whenever it is sensed in CSMA/CA that the channel is busy, the senders double their range of random interval. As a result, a successfully transmitting host just resets this range whereas all the other contenders just double their range. Hence, the channel utilization is taken care off but the applications waiting for the data suffer serious problems due to the resulting delays and delay variations. This is known as the 'capture effect'. A solution to the problem explained above is the use of Distributed Coordination Function (DCF). The idea is to sense the channel during back off. Once another transmission is detected, the sensing host freezes its back off timer. Once the channel is free, the back off timer is resumed. As a result, it is made sure that the host, which has been waiting for the longest time, is sure to get a chance to transmit. This idea of sensing the channel during back off requires extra power, which is minimum, and hence, precious in the cases of wireless devices. However, the information received as a result of spending this power is put to good use and hence this forms a feasible solution. DCF can be refined into Logarithmic Arbitration - DCF, where an idle counter is set during the doubling and back off. If no activity is sensed during back off, the range of random interval used is too large and hence the present range is halved. The process continues till a transmission is scheduled. Another problem, despite the use of CSMA/CA is the error control, which is left to the end-to-end protocol. It might be too late before the loss of frame due to collision has been detected and the corrective actions to be taken. This leads to inefficient use of the channel. DCF makes use of acknowledgements to contend with the loss of frames due to collision. However, this increases delay variations. LA-DCF overcomes the problem without retransmissions. Forward Error Correction (FEC) is one of the techniques to handle losses of frames. It enables the error detection and correction by making use of redundant information in the packets sent. It also has the advantage that there is no need for the sender to be contacted about the loss. Suppose there are k packets to be sent. These are coded at the sender using an encoder into n packets. The n packets are decoded at the receiver. The n packets are such that any k packets of these n packets are adequate to reconstruct the original k packets. The first k packets of the n packets are the original k packets and the rest n - k packets are the parity packets that contain the redundant information that are requisite to reconstruct the data. In a wireless environment, a proxy that implements FEC can be used between the sender and the receiver, its main function being error control. It can be provided with extra power and adequate buffer space. As a result, localization of retransmissions and localization of overhead of FEC packets are obtained as added benefits. This requires the placement of the proxy as close as possible to the wireless hosts. Another problem of using wireless for video transmissions is its unreliable nature. The reasons are the special characteristics of this medium. The Bit Error Rate (BER) of the wireless medium is more than that of the wired medium. Also, the loss of packets in this medium, which may be due to various reasons (collision and its avoidance being one of them), is known as 'fading'. The unreliability may arise from the incompetence of the traditional transport layer protocols such as TCP and UDP. The problem with the usage of TCP is that it assumes that the primary reason for the loss of packets is congestion and does its congestion control mechanism - exponential back off which in turn contributes to the delay rather than avoiding it. This unnecessarily affects the end-to-end throughput. The high BER and fading bring about the increase in packet loss in a UDP connection. Some solutions in TCP include the use of local retransmissions using base stations, methods to find out the reason of packet loss and so on. As for UDP, partial checksum method is used where only the checksum of the header is calculated and for fading, retransmission is used for lost packets after the fading occurs. It is clear that a method to correctly predict the reason of packet losses is required. Many methods have been proposed in this field. Proper congestion control and error control methods have been included in them. Errors in links in a wireless environment mainly depend on the location of the node. Nodes may share the same access point. In such cases, the link may appear to be in a good state (i.e.) stable to some nodes and not so to the other. Retransmissions in case of wireless links have two major drawbacks - one is that packets sent over an unstable link have even lesser chance of reaching their destination and the other is that these retransmitted packets may induce delay into the other packets. A retransmission queue can be used as the solution. When a packet has been lost, instead of immediate retransmission, the packet is put at the end of the retransmission queue. When the next packet to the same destination can be transmitted, the packet in the queue is sent and the other packet is put into the queue. This kind of scheduling of transmissions can stop as soon as the link becomes stable. If not, the node, which now has the stable link, will suffer unnecessary delays in receiving packets since each of the packets destined to this node has to pass through the registration queue. The link state can also be determined from the number f packets that have been sent successfully from the retransmission queue. When the number reaches a threshold, the count is reset. The process continues until the retransmission queue becomes empty. One of the methods of error control is using partial checksum. Packets with even single bit errors in this checksum are discarded but packets that have a chance of having errors in the payload part are delivered. In such cases, the dealing of such corrupted packets is left to the application at the receiver end. The packet loss ratio can be decreased effectively using partial checksum method. In another method, error control is done at application level as well as the receiver is expected to give a BER report to the server, when BER is found to be high. Error dependencies are taken into account here. Handling of the corrupted packets can be done according to the nature of the application. If the application is not very demanding, a request for retransmission can be sent through a negative acknowledgement and the sequence number of the corrupted packet so that the server will know which packet is to be sent again. If the application tends to be delay sensitive, a request for retransmission cannot be sent. The sender cannot know when the packets are lost and it continues sending the packets. In case of the receiver, it is possible to know when the packets are lost. Hence, the receiver may inform the server, when it is again in a position to receive, with an end of fading signal. Now the server may know that some packets have been lost. The sequence number of the last packet received correctly is sent along with this signal so that the server might know the point from which retransmission is required. Congestion control can be done at the server by using calculations of available bandwidth, round trip times, retransmission time outs and packet loss ratio. These calculations are made accurate by using the feedback from the receiver. All these calculations are similar to those used in TCP except that these have to be made adaptable to the wireless environment. For the quality analysis problem, certain indicators can be used. These indicators are the quality of the video sent and the characteristics of the wireless link used. Depending on these indicators, decisions regarding the nature of the video that can be sent through a particular network with minimal error rate and packet loss are made. Another problem is due to the limited bandwidth available in wireless and hence limited throughput. In 802.11 standard, the access point is the base station. Each node is given a timeslot during which it can enjoy the services of the base station. At the end of this timeslot, a node that gets the access point for itself now can announce to all other contenders. This scheme does not address the problem of priority handling between the nodes. Some applications may be very demanding and thus, cannot afford to wait until it gets its chance to the access point. Use of SDMA, TDMA (Space Division Multiple Access, Time Division Multiple Access) or a hybrid of these techniques along with some other methods may ensure fair use of the bandwidth available. The prediction of performance of video streaming over a wireless connection was experimentally done in Worcester Polytechnic Institute (MA, USA). Various metrics were used across the layers stack. These include the physical layer Received Signal Strength Indicator (RSSI), the link capacity, MAC layer retry fraction, loss rate in IP, round trip time and throughput. These six factors were adequate for predicting the quality of the streamed video at the receiver's end. The results were graphically shown as below. Thus, the numerous reasons for the erratic and slow transmission of video stream over wireless link such as high error rate, congestion, use of CSMA/CA, inefficiency of traditional transport protocols etc. are given and solutions to most f these problems have been discussed. REFERENCES (2008). Overview. Retrieved March 5, 2008, from http://standards.ieee.org/wireless/overview.html Rodriquez, E. (2002). Wired Vs. Wireless. Retrieved March 5, 2008, from http://www.skullbox.net/wiredvswireless.php (2006). Video Streaming Over Wireless Networks. Retrieved March 5, 2008, from http://www.cnri.dit.ie/pres_vid_streaming.html Ding, G., Ghafoor, H., & Bhargava, B. (2000). LA-DCF: A New Multiple Access Protocol for Ad-Hoc Wireless Networks. Retrieved March 5, 2008, from http://209.85.175.104/searchq=cache:IHanSR2c6l0J:raidlab.cs.purdue.edu/papers/VideoWireless.pdf Li, M., Li, F., Claypool, M., & Kinicki, R. (June 2005). Weather Forecasting Predicting Performance for Streaming Video over Wireless LANs. Retrieved March 5, 2008, from http://web.cs.wpi.edu/claypool/papers/weather/wlan_paper.pdf (January 29, 2008). Carrier sense multiple access with collision detection. Retrieved March 5, 2008, from http://en.wikipedia.org/wiki/Carrier_sense_multiple_access_with_collision_detection. (February 14, 2008). Carrier sense multiple access with collision avoidance. Retrieved March 5, 2008, from http://en.wikipedia.org/wiki/Carrier_sense_multiple_access_with_collision_avoidance Mitchell, B. (2008). Wireless Standards - 802.11b 802.11a 802.11g and 802.11n. Retrieved March 5, 2008, from http://compnetworking.about.com/od/wirelessfaqs/f/mixing80211bg.htm. Mitchell, B. (2008). "802.11a". Retrieved March 5, 2008, from http://compnetworking.about.com/cs/wireless80211/g/bldef_80211a.htm Mitchell, B. (2008). "802.11b". Retrieved March 5, 2008, from http://compnetworking.about.com/cs/wireless80211/g/bldef_80211b.htm (2008). 802.11 Wireless Standards. Retrieved March 5, 2008, from http://compnetworking.about.com/od/wireless80211/80211_Wireless_Standards.htm (2008). Wi-Fi Wireless Networks and Technology. Retrieved March 5, 2008, from http://compnetworking.about.com/od/wireless/WiFi_Wireless_Networks_and_Technology.htm (2008). Wireless Design and Development. Retrieved March 5, 2008, from http://www.wirelessdesignmag.com/ Ge, P., & McKinley, P. K. (n.d.). Comparisons of Error Control Techniques for Wireless Video Multicasting. Retrieved March 5, 2008, from ftp://ftp.cse.msu.edu/pub/crg/PAPERS/ipccc-02.ps.gz. Yang, F., Zhang, Q., Zhu, W., & Zhang, Y. (n.d.). AN EFFICIENT TRANSPORT SCHEME FOR MULTIMEDIA OVERWIRELESS INTERNET. Retrieved March 5, 2008, from http://research.microsoft.com/fanyang/papers/3GWireless_1125.pdf. Read More