Electrical computers and digital data processing systems: input/ – Access arbitrating
Reexamination Certificate
2000-09-25
2003-11-25
Ray, Gopal C. (Department: 2181)
Electrical computers and digital data processing systems: input/
Access arbitrating
C710S107000
Reexamination Certificate
active
06654838
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to methods and apparatus used in systems for communicating data (e.g., voice, video and alphanumeric data), including but not limited to telecommunications systems, computer systems, etc. More particularly, the invention relates to methods and apparatus used to switch telephony and data signals without wasting bandwidth and without compromising the quality of telephony.
2. Brief Description of the Prior Art
Telephony switching was originally based on a system known as time division multiplexing (TDM). Although the actual implementation of TDM is quite complex, the concept of TDM is easy to understand.
Several relatively low frequency signals are interleaved to form a single relatively high frequency signal. The signals are mapped into what is referred to as a frame. Individual telephone connections are assigned a slot in the frame. Each slot corresponds to a destination (or a switch connection).
By nature, TDM provides a constant bandwidth allocation to each telephone connection. Although originally designed for telephony, TDM is also used in data switching. However, when used for data switching, TDM is inefficient. Bandwidth is wasted because most data communication does not require constant bandwidth.
Efficient data switching was initially provided through the use of packets. Packet switching techniques require an arbitration system whereby bandwidth is arbitrated among users. Some packet switching techniques utilize fixed length packets and some use variable length packets. Switch connections (or packet destinations) are specified in a part of the packet called the header.
Packet switching allows for the dynamic allocation of bandwidth to wherever it is needed and allows for “bursty” traffic, i.e. traffic which requires a large amount of bandwidth for a short amount of time and then requires little bandwidth.
As the popularity of data communication has spread, many different methods have been proposed for integrating telephony and data. Some of these proposals include ISDN (Integrated Services Digital Network), voice over ATM (Asynchronous Transfer Mode), ATM over SONET (Synchronous optical Network), and voice over IP (Internet Protocol).
The challenge in each of these proposals is to fairly allocate bandwidth without wasting bandwidth, while maintaining quality of service. However, each of these proposals is ultimately based on either TDM (ISDN and SONET) or packet switching (ATM and IP).
Those based on TDM continue to waste bandwidth and those based on packet switching provide poor quality of service during periods of congestion.
In conjunction with the methods proposed for integrating telephony and data, different apparatus have been developed. Although TDM traffic can be packetized and packet traffic can be provisioned over a TDM connection, different switches must be used for each type of traffic.
All digital electronic communication consists of payload data and control data. The payload data can not be readily distinguished from noise without some type of control information by which to interpret it.
To better appreciate the background of the invention it should be understood that there are two types of control information: element synchronization and transmission source synchronization.
Element synchronization provides a means of delineating the logical elements of the data stream, so bits, bytes, frames, etc. can be delineated. Transmission source synchronization, on the other hand, is necessary where more than one source can be simultaneously using the same transmission medium. The sources must be synchronized in time. If two sources attempt to send different data at the same time, the resulting ambiguity renders the data useless.
Both types of control information can be supplied in many forms, from a very simple time synchronization format built into the data stream, as exemplified by the RS-232 serial protocol, to a completely separate stream of data, complex in its own right, as in the PCI bus architecture.
Element synchronization can be achieved by one or more clock signals or by built-in (or on-line) timing.
An example of built-in timing is the RS-232 serial data signal. The RS-232 serial stream, like most digital data, is composed of bits and bytes. A bit can be either logical one or zero; it can have no other value. This can be represented electrically by two voltage levels, two frequencies, presence or absence of a voltage, etc. Eight bits compose a byte. Most digital data is composed of strings of bytes.
An RS-232 stream delineates the beginning of each byte of data by a rise of voltage to a predetermined MARK level (the START bit), followed by eight bits, each being present for a certain length of time, and terminated by a STOP bit of a certain length. This element synchronization makes it possible to extract the intelligence from the data stream by taking samples of the line voltage at periodic intervals following the leading edge of the START bit. This sampling interval is determined by the baud rate of the transmitting device. There is no need to provide a separate clock signal to mark the individual bits.
In contrast, in the PCI bus architecture, discrete time periods are marked off by a separate CLOCK signal. The timing of the clock pulses allows the receiving station to discern the individual bits and bytes of the stream of data on the data transmission bus. In a parallel bus architecture of this type, multiple channels carry data simultaneously, with the bit timing of all channels synchronized to the same external clock.
Transmission source synchronization mechanisms can also be either built-in (“in band”) or external (“out of band”). As mentioned above, if more than one source transmits data into the medium at the same time, unless the data is exactly the same for all sources, the result will be useless noise. Therefore, some of the resources used in the data transmission process must be used for controlling the flow of data.
In other words, there must be some way to determine who can transmit at any particular time.
Resources that are used only for the control of the data flow are referred to as “overhead”. These resources are of no value to the ultimate users of the system other than as conveyors of the payload data. The source synchronization overhead, necessary to prevent or compensate for simultaneous transmission by more than one device, can become a significant portion of the available bandwidth; that is, the overall data carrying capacity of a particular transmission medium, such as a fiber optic link or a radio transmission frequency spectrum. Although necessary, the synchronization control overhead is not desirable, since the overhead contributes to the expense of the system.
In some transmission systems, the bandwidth is divided into channels, wherein the data is carried in several parallel “pipes”. Parallel means that the data is carried simultaneously in all channels.
In other system, data is carried in a single “pipe”, in which only one basic element (“bit”) is transmitted at a time.
When the medium used carries only a single bit at a time, as in a high-speed serial system (e.g. Ethernet), source synchronization involving several originating sources is a serious problem. If several sources begin transmitting messages at the same time, no other station in the system knows where the message is originating.
In addition, the data of the various transmitting sources is combined in a completely random fashion so that all intelligence is lost. Of necessity, a system of this type must build the source synchronization mechanism into the transmission protocol because there is no external method of controlling access.
The controls for source synchronization that are built-in involve either some method of rotating control of the medium to each source or, alternatively, some method of seizing control of the medium without interfering with or being interfered with by another source.
The former method is embodied in a token ring system, in which a logical “token” is passed
Ray Gopal C.
Siemens Information & Communication Networks, Inc.
LandOfFree
Methods for performing bit sensitive parallel bus peer... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Methods for performing bit sensitive parallel bus peer..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Methods for performing bit sensitive parallel bus peer... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3138220