Constant bit rate has the best QoS, with low cell jitter. It can be used for broadcast video contribution circuits, possibly a 34 Mbit/s data stream. The variable bit rate real-time can be used for standard voice circuits. IP traffic usually is allocated the unspecified bandwidth. Some characteristics that make ATM particularly attractive for television distribution are the choice of uni- or bidirectional links; the two-way links can be asymmetric. As network traffic increases, the bandwidths grow to the point where optical switching becomes the only cost-effective way to handle the very high data rates of the network backbones. So the electronic processing of SONET and SDH will fall by the wayside. The move now is to an all-optical infrastructure called photonic networking. Photonic networking As IP becomes the standard for data exchange, and voice-over IP becomes more used, the overhead of the ATM traffic engineering becomes more of a hindrance. Proposals to run IP directly over the Dense-Wave Division Multiplexed (DWDM) photonic network would greatly simplify the routing of traffic by stripping out two layers, ATM and SONET. To introduce this new concept, the capability of IP routing would have to be extended. IP networks and telecommunications 29 real time variable bit rate available bit rate time bandwidth best effort constant bit rate non real time VBR link bandwidth Figure 2.6 ATM traffic classes.
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Adult Web Hosting services
Comments Off
share a physical link; the link bandwidth is shared dynamically, giving much more efficient use of the link. This is called opportunistic bandwidth. Because there is now a queue of data awaiting available bandwidth, quality of service becomes an issue. The advantage is that the toll rates can be lower than the reserved bandwidth of a voice circuit. ATM Asynchronous Transfer Mode (ATM) was developed for high-speed packet transport over optical networks like SONET. ATM uses small fixed-size data cells rather than the large packets of frame relay. Each cell is 53 bytes long, with a payload of 48 bytes. These small cells are more suited to voice and multimedia traffic, where low latencies are demanded. Both permanent and switched (PVC and SVC) virtual circuits can be set up. The cell header carries the virtual channel and virtual path identifier that are used to identify connections within the network. ATM is linked to other network layers by the ATM Adaptation Layer (AAL). The packets are given one of five categories of priority by traffic class: 1. Constant bit rate 2. Real-time variable bit rate 3. Non-real-time bit rate 4. Available bit rate 5. Unspecified bandwidth or best effort 28 The Technology of Video and Audio Streaming Table 2.6 Synchronous Digital Hierarchies Data rate SDH SONET Signal Capacity Signal Capacity 51.84 Mbit/s STM-0 21 E1 STS-1, OC-1 28 DS1 or 1 DS3 155 Mbit/s STM-1 63 E1 STS-3, OC-3 84 DS1 or 1 E4 or 3 DS3 622 Mbit/s STM-4 252 E1 STS-12, OC-12 336 DS1 or 4 E4 or 12 DS3 2.48 Gbit/s STM-16 1008 E1 STS-48, OC-48 1344 DS1 or 16 E4 or 48 DS3
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost PHP Web Hosting services
Comments Off
The Synchronous Optical Network (SONET) is a subset of the Synchronous Digital Hierarchy (SDH), an ITU-T standard. The SDH standard can accommodate both ITU and ANSI PDH signals. Frame relay So far I have been describing voice circuits. When a voice circuit is set up you reserve a bandwidth slot for the duration of the call. If none is available you get a busy tone. The requirements for data are different. The reserved bandwidth is not as important as ensuring the delivery. Data can use the spare capacity as voice traffic changes up and down; data packets can be dispatched as capacity is available. Frame relay is a standard for packet-switched networks that operate at layer 2 the data link layer of the OSI model. A bidirectional virtual circuit is set up over the network between the two communicating devices. Variable-length data packets are then routed over this virtual circuit. A number of virtual circuits can IP networks and telecommunications 27 Table 2.5 Plesiochronous Digital Hierarchies ITU-T standard ANSI standard Signal Data rate Channels Signal Data rate Channels DS0 64 kbit/s E1 2.048 Mbit/s DS1 1.544 Mbit/s 24 DS0 E2 8.45 Mbit/s 4 E1 DS2 6.3 Mbit/s 96 DS0 E3 34 Mbit/s 16 E1 DS3 45 Mbit/s 28 DS1 E4 144 Mbit/s 64 E1 1 24 28 2 24 x DS0 (64 kb/s) 27 x DS1 1 x DS3 (45 Mb/s) DS1 (1.5 Mb/s) MUX MUX Figure 2.5 Voice circuit multiplexing.
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost PHP Web Hosting services
Comments Off
at 1.5 Mbit/s in the DS-1 (digital signal) format. The European equivalent is E-1 at 2 Mbit/s. U.S. and international standards There are two main telecommunications standards: ANSI, used in North America and parts of the Pacific Rim, and the ITU-T standards, used in the rest of the world. The ANSI hierarchy is based on a digital signal (DS0) of 64 kbit/s. Plesiochronous Digital Hierarchy (PDH) The early digital trunk circuits multiplexed a large number of voice circuits into a single high data-rate channel. The systems at the remote ends were not absolutely locked together; instead, each runs off a local reference clock. These clocks were classed as plesiochronous; plesio is a Latin term derived from the Greek meaning near, so plesiochronous refers to clocks that are in near synchronism. The early data circuits were asynchronous; the clocks were derived from simple crystal oscillators, which could vary from the nominal by a few parts per million. Large receive buffers are used to manage the data flows. In PDH networks, to cope with terminal equipment running on slightly different clocks, extra bits are stuffed into the data stream. This bit stuffing ensures that a slower receiver can keep up with the real payload rate by simply dropping the extra bits. To extract a single voice circuit from a DS3, the channel has to be demultiplexed back to DS1 channels. To build trunk circuits in rings around a country, each city passed would have to demultiplex and remultiplex the data stream to extract a few voice circuits. Synchronous networks (SONET) To avoid the multiplexing issues and the overheads of bit stuffing, highly synchronous networks were developed. By referencing terminal equipment to a single cesium standard clock, the synchronism could be ensured to a high degree of accuracy. The standard uses a byte-interleaved multiplexing scheme. The payload data is held in a fixed structure of frames. At a network terminal the signals can be added or dropped from the data stream, without the need to process the other traffic. It is rather like a conveyor belt carrying fixed size containers at a regular spacing. As the belt passes a city, you take away the containers you want, and drop new ones into gaps. The other containers pass unhindered. 26 The Technology of Video and Audio Streaming
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Adult Web Hosting services
Comments Off
area of the network. In such a case the dense-mode flooding would cause unnecessary congestion on the network. The core-based tree (CBT) uses a core router to construct a distribution tree. Edge routers send requests to join the tree, and then a branch is set up. Network traffic will concentrate around the core, which can cause problems with congestion. MBone The multicast-enabled backbone project (MBone) was set up in 1992 to enable the IETF (Internet Engineering Task Force) meetings to set up audioconferencing to communicate with remote delegates. In 1994, the MBone was used to multicast a Rolling Stones concert to the public. The term is used more now to refer to the general multicast-enabled backbone. This piggybacks onto the general unicast Internet backbone. The multicast datagrams are encapsulated as unicast packets and tunnel through unicast networks. Telecommunications Telecommunications networks originally were set up for telephony, but more than half the traffic is now data. The packet-switched networks used for data and telephony traffic also can be used to carry the Internet. The circuits are constructed in a hierarchy of bit rates designed to carry multiple voice circuits, with the basic unit being 64 kbit/s. Data is carried in a compatible form. T-1 and E-1 If you ever have tried encoding multimedia content, you will have seen T-1 on the menu. T-1 is the basic digital carrier used in North America. It transmits data IP networks and telecommunications 25 Table 2.4 Time-to-Live Initial Values Scope TTL Local area 16 National high-bandwidth sites 32 National sites 48 Continental sites 64 Intercontinental high-bandwidth sites 128 Worldwide coverage 192
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Cheap Web Hosting services
Comments Off
and it is not unknown for an address not to be unique. Hopefully the upgrades to IP version 6 will help to solve that problem by making far more addresses available. IGMP Most of the complexity of a multicast lies with the routing, not with the server and the client. Instead of the server sending individual packets to each client, a single packet is transmitted to a multicast group. This group is allocated a single IP address. The primary mechanism for controlling the delivery of the datagram to multiple destinations is the Internet Group-Membership Protocol (IGMP). It is a session-layer protocol used by the client to join and leave a multicast. A multicast- enabled router uses this session information to set up the route from the server to the client. The router will forward multicast datagrams only if regular IGMP messages are received from downstream clients (typically at intervals of about 60 seconds). Several routing options have been developed for multicast routing: Protocol Independent Multicast (PIM) Distance-Vector Multicast Routing Protocol (DVMRP) Core-based tree (CBT) Multicast Open Shortest Path First (MOSPF) There are two ways of multicast routing: dense mode and sparse mode. Sparse and dense routing Dense mode floods the network then prunes back the unused branches. This assumes that the viewers of the multicast are densely distributed through the network, which could be the case for corporate communications over an intranet. It requires a generous bandwidth. DVMRP, MOSPF, and PIM dense mode are all such protocols. The reach of dense routing trees is limited by the time-to-live parameter (TTL). The value of TTL is decreased by one each time a datagram passes through a router; once it reaches zero, the router will discard the packets. This can be used to restrict the range of the multicast that potentially could propagate through the entire Internet. TTL is measured in seconds and usually is set to a default value of 64. The other type of multicast routing is sparse mode, which is used for applications where the clients are dispersed, possibly geographically, over a wide 24 The Technology of Video and Audio Streaming
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Cheap Web Hosting services
Comments Off
multicast event. There is a proposal to dynamically allocate the host group addresses, much like the dynamic allocation of client IP addresses. The permanent addresses have to be registered with the Internet Assigned Numbers Authority (IANA), or the designated proxy organization for your country (like the RIPENCC in Europe). There has been a certain amount of chaos in this area, IP networks and telecommunications 23 LAN 500 clients East coast office LAN 1,000 clients South coast plant LAN 500 clients West coast HQ media server VPN VPN 1 stream 1 stream encoder live presentation 1 stream mcast router mcast router mcast router Media players Figure 2.4 Multicast presentation.
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Adult Web Hosting services
Comments Off
connection for each media player client. This is called unicasting. In this example you would have to transmit 500 + 1,000 + 500 (= 2,000) separate content streams. The webcaster can look with envy at the television broadcaster. With one transmitter and one tower the broadcaster can reach every resident living within his or her service area. In a metropolitan area he or she can reach an audience of several million people. As a webcaster you have to provide server resource for each viewer, plus the bandwidth of the Internet has to be sufficient to carry all the streams that you want to serve. Multicasting offers an alternative to conventional streaming or unicasting. A single stream is served to the Internet as a multicast. All the viewers then can attach to the same stream. The client initiates a multicast; the server just delivers the stream to the network. Further viewers just attach to the same stream. The server has no knowledge of where the stream is going, unlike the normal TCP client server handshaking interactions of an Internet connection. A client will be made aware of a multicast by some out-of-band channel; it could be by e-mail or through publicity on a web site. The viewer then requests the multicast at the appropriate date and time. An alternative is to use the session announcement protocol. Note that you can broadcast to a network, but it is not like a television broadcast. It is used by network administrators for control messages, and does not propagate beyond the local subnet. Multicasting sounds like a very efficient solution to the resource problems of delivering a webcast to very large audiences. But there are catches. First, it can be used only for live or simulated live webcasting. You lose the interactivity of on-demand streaming. The second drawback is that many older network routers do not support multicasting. There are ways around this: Multicast streams can be tunneled through legacy plant, and the multicast enabled backbone (MBone) can be used. Many of the problems have restricted its use to corporate networks (intranets). Large public webcasts have had to resort to conventional splitting and caching to guarantee delivery to all potential clients. Note that multicasting is not limited to streaming; it also can be used for general data delivery (like database upgrades across a dispersed enterprise, or for video conferencing). Multicast address allocation Most IP addresses that are classless (CIDR) fall into Class C. If you work for a very large corporation or government department, then you may use the Class A and B address spaces. Multicasting uses a reserved set of IP addresses in Class D, ranging from 224.0.0.0 to 239.255.255.255. To make public Internet multicasts you will need a unique address. Although some addresses are permanently allocated to hosts, they are usually transient and allocated for a single 22 The Technology of Video and Audio Streaming
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Cheap Web Hosting services
Comments Off
Multicasting Suppose the CEO of an enterprise wants to stream an address to all the staff. Let us say there are 500 staff at headquarters on the West coast, 1,000 personnel work at the South coast plant, and another 500 at the East coast offices. The normal way to transmit an Internet presentation is to set up a one-to-one IP networks and telecommunications 21 LAN 500 clients East coast office router LAN 1,000 clients South coast plant router LAN 500 clients West coast HQ router media server VPN VPN 500 streams 1,000 streams encoder live presentation 2,000 streams Media players Figure 2.3 A unicast presentation.
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost PHP Web Hosting services
Comments Off
changed to counter network congestion. Some RTCP messages relate to control of a video conference with multiple participants. Session Description Protocol (SDP) SDP is a media description format intended for describing multimedia sessions, including video-conferencing. It includes session announcement and session invitation. Real-Time Streaming Protocol (RTSP) The Real-Time Streaming Protocol is an application-level protocol for the control of real-time multimedia data. RTSP provides an extensible framework rather than a protocol. It allows interactive, VCR-like control of the playback: Play, Pause, and so on. A streaming server also can react to network congestion, changing the media bandwidth to suit the available capacity. RTSP was developed intentionally to be similar in syntax and operation to HTTP version 1.1. It does differ in several important aspects, however. With RTSP both client and server can issue requests during interaction with HTTP the client always issues the requests (for documents). RTSP has to retain the state of a session, whereas HTTP is stateless. RTSP supports the use of RTP as the underlying data delivery protocol. The protocol is intended to give a means of choosing the optimum delivery channel to a client. Some corporate firewalls will not pass UDP. The streaming server has to offer a choice of delivery protocols UDP, multicast UDP, and TCP to suit different clients. RTSP is not the only streaming control protocol. Real s precursor, Progressive Networks, used a proprietary protocol before RTSP was developed. MMS The Microsoft Media Server (MMS) is Microsoft s proprietary control protocol. MMS handles client interaction the controls like Play or Stop. MMS uses TCP as the delivery layer. The media data can be transmitted separately over UDP or TCP. SMPTE time code RTSP uses Society of Motion Picture and Television Engineers (SMPTE) time code as a time reference for video frames. Note that RTP uses a different time reference, the Network Time Protocol (NTP), which is based on universal time (UTC). RTP uses the middle 32 bits of the NTP 64-bit fixed-point number to represent the time. The high 16 bits of the 32-bit NTP fraction are used to represent subsecond timing this gives a resolution of about 15 ms, or about one quarter of a television line. 20 The Technology of Video and Audio Streaming
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost PHP Web Hosting services
Comments Off