Multiplex communications – Communication techniques for information carried in plural... – Combining or distributing information via time channels
Reexamination Certificate
1999-12-17
2003-09-16
Yao, Kwang Bin (Department: 2662)
Multiplex communications
Communication techniques for information carried in plural...
Combining or distributing information via time channels
C370S395610, C370S473000, C370S477000
Reexamination Certificate
active
06621833
ABSTRACT:
BACKGROUND OF THE INVENTION
a. Field of the Invention
The present invention relates generally to passing voice communications over a data communications network such as an asynchronous communications network or a synchronous communications network.
b. Background Information
Almost all customers of data traffic today have additional, separate links to carry voice. This is inefficient for the customer and the communications provider. Many are seeking techniques that place Ds
0
channels in data packets for transmission over a data link, so that they can remove their voice links.
A communications network serves to transport information among a number of locations. The information is usually presented to the network in the form of time-domain electrical signals and can represent any combination of voice, video, or computer data. A typical communications network consists of various physical sites called “nodes,” interconnected by conduits called “links.” Each link carries information from one site to another site. End user sites may contain data terminating equipment (DTE) for combining, separating, and transforming data with or without voice. Network provider sites usually include either edge switches, with user network interfaces (UNI), or backbone switches, which only connect to other backbone switches and edge switches and do not contain UNI.
Voice information is carried via a Ds
0
(or voice) channel that is a 64 kilobits per second (64 Kbps) channel and also the worldwide standard for digitizing voice conversation. The channel throughput is 64 Kbps because a digital data stream can adequately represent an analog voice signal if sampled at a rate of 8000 samples per second. If each voice sample is digitized using 8 bits, this results in a digital data stream of 64 Kbps. Since Ds
0
is a synchronous TDM link, once a channel connection has been setup between two users, that channel is dedicated until the connection is torn (or brought) down, and cannot be used by anything or anybody else even if there is silence in the line.
Data currently is transmitted between nodes either as synchronous or asynchronous. In a synchronous network using Synchronous Transfer Mode (STM), each timeslot is assigned a certain time when it is to arrive at each node. The time when the timeslot arrives determines where the timeslot goes. Thus, the individual timeslots do not need to have routing information within them.
Asynchronous Transfer Mode (ATM), Frame Relay (FR), and Internet Protocol (IP), collectively called data, are considered asynchronous because each node in the network does not know until after a data packet arrives where it is intended to go. The arrival of a particular data packet at a node, on the other hand, is not guaranteed to occur at a particular point in time. Only by analyzing the routing information in the header can the data switch know where to route the data packet.
Asynchronous Transfer Mode is designed to be carried over the emerging fiber optical network, called the Synchronous Optical NETwork (SONET), although it can be carried over almost any communications link. The basic unit of ATM is a data packet called the ATM cell. Each cell contains two parts, a header, which contains routing information, and a payload, which contains the data to be transported from one end node to another. The ATM cell is always the same size.
Frame Relay and Internet Protocol are two other asynchronous types of communications protocols. Each is similar to ATM in that they also consist of a data packet. However, they differ from ATM in that their packet size can vary from packet to packet, and both can be considerably larger than ATM. This allows them to make more efficient use of the bandwidth of the communications media they travel over, but it makes receiving them more difficult in that packet size must be calculated for each packet. Both the FR protocol and IP may be used in point to point connections, but IP may also be used when multiple ports are connected to a single transmission medium.
Data can consume as much or as little as is needed for carrying actual traffic, because data does not reserve a fixed amount of bandwidth per link. While voice will never overload, or oversubscribe, the capacity of its links, there are mechanisms in place to handle data overloads when more is available than a physical link can carry. It is these mechanisms that allow data network designers to specify more data demand than capacity to carry, which is a process called statistical multiplexing.
Statistical multiplexing is the concept of giving multiple customers, in sum total, more bandwidth through a physical connection than it can carry. This is also known as over-subscribing. Studies have shown that customers will not always use all of the bandwidth their carrier has set aside for them. It is during this period of non-use by a customer that spare bandwidth is available for the over-subscription. If sufficient numbers of customers are placed on a single physical connection then large quantities of spare bandwidth can be realized.
When traffic is isolated among two or more physical connections, less statistical multiplexing can occur, as customers on one connection cannot use spare bandwidth on another. By joining all customers into a single, large connection, better statistical multiplexing occurs and the carrier is able to sell more bandwidth on one high-speed physical connection than on several smaller connections whose sum is equal to the one high-speed connection.
There are different ways of handling overloads in the data network. In ATM, the network is designed with large buffers, which absorb the excess traffic, queuing it up until capacity is available to place the traffic through the network. The traffic that is delivered out of its buffers first is determined by the quality of service (QOS) the customer has paid the carrier to provide. Higher QOS traffic is removed from its buffers before lower QOS. This is important for real time applications such as voice or interactive TV services, which must get through the network with a minimum amount of delay.
In those instances where so much excess traffic is delivered that the network cannot queue it up in buffers, the lower QOS traffic is deleted, or dropped, from the buffers to make room for higher QOS traffic to be queued up. Ideally, customer end-to-end protocols will detect this loss of traffic and will re-transmit the lost information.
An emerging standard in the IP network uses a different approach to handling overloads. In IP, there is no QOS as in ATM. Once a data packet is injected into the IP network, it will be given equal priority with all other traffic and delivered to its destination with a minimum of delay.
In an IP network, the traffic density in a link is closely monitored. As it begins to approach the link capacity, the IP data switch send congestion notices back towards the data sources telling them to slow down the amount of data they send. Each notice will, for a limited length of time, force the data source to restrict what it sends out. As link traffic gets closer and closer to link capacity, more of these messages are sent backwards. When an IP switch receives congestion notices and reduces the rate of transmission, it may experience congestion as well and will send congestion notices back to its sources.
Eventually, the notices reach the traffic origins, customers. The customer equipment must then cut back on what is sent into the network, and must decide which traffic it puts out has the highest priority so that it goes out while the lower priority traffic has to wait. Thus, the IP network passes the job of determining traffic priority to the customer. If a customer has a great deal of high priority traffic, it may pay a premium to not receive as many congestion notices when congestion hits the network as another customer may pay, so that it will get more guaranteed traffic during congestion.
The IP data switches also usually maintain small buffers, but these are designed exclusively to handle the small, temporary overloads t
Nguyen Hanh
World Com, Inc.
Yao Kwang Bin
LandOfFree
Method and system for efficiently passing the silence or... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for efficiently passing the silence or..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for efficiently passing the silence or... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3014103