Abstract: A fair and efficient admission scheme enables sharing of a common resource among N traffic classes, such that each class is guaranteed (on a suitable long-term basis) a contracted minimum use of the resource, and each class can go beyond the contract when extra resource capacity becomes temporarily available. The scheme operates in an open loop mode, and thus does not require information describing the current status of the resource. For the purposes of description, one form of the invention is best described in terms of "tokens" and "token banks" with finite sizes. Our scheme uses one token bank per class (bank `i`, i= 1, . . . ,N), plus one spare bank. Class `i` is preassigned a rate, rate(i), i=1, . . . ,N, that represents the "guaranteed throughput" or contracted admission rate for class `i`. Tokens are sent to bank `i` at rate(i). Tokens that find a full bank are sent to the spare bank. When the spare bank is also full, the token is lost. Every admitted arrival consumes a token.

Abstract: A congestion control scheme for connectionless networks relieves congestion by routing a portion of traffic on a congested primary path onto a predefined alternate path constructed such that loop-freedom is guaranteed. Explicit care is taken to avoid spreading congestion onto alternate paths. The control actions are taken in a completely distributed manner, based on local measurements only and therefore no signaling messages need to be exchanged between nodes.If desired, lower loss priority may be assigned to alternate routed traffic. Congestion is monitored locally and thresholds defined to declare the onset and abatement of congestion. The present invention affords at least an order of magnitude improvement in end-to-end cell blocking under sustained focussed overload.

Abstract: A switching module comprises input buses (1--1 , . . . 1-n) and at least one output bus (2-ij), in which each input bus, for data packet transmission, is connected to the output bus by means of a crosspoint switch (4-i). Each crosspoint switch (4-ij) has a unique address and a request output (61) for sending out a request signal. The switching module further includes allocation means (3-j) which place empty packets having the unique address of a crosspoint switch (4-ij) onto the output bus (2-j). The allocation means (3-j) comprise recording means (105,201) for receiving the request signals and include a FIFO (107,206) for storing representatives of the unique crosspoint switch addresses that are derived from the request signals. The allocation means (3-j) further include a packet generator (103) for deriving addresses from the representations from the queue store (107,206) and allocating these addresses to the empty packets.

Abstract: Messages (such as telephone calls or data messages) are automatically forwarded to an individual who is away from his/her home or office based upon information obtained when the individual engages in a transaction or activity, such as making a credit card purchase or credit card call, that indicates his or her location, e.g., at a particular merchant or telephone. Information from the transaction that links the individual and his/her current location is forwarded to a database to update an electronic record indicating the individual's availability, and where the subscriber may be reached. A call thereafter placed to the individual, either using a personal telephone number or the ordinary home or office number, causes the individual's record to be accessed and current location retrieved, so that the message can be automatically forwarded to that location. In some instances, the caller is furnished with the caller's general location (i.e.

Abstract: Each potential user of a telecommunications network is assigned a single credit card containing at least first and second authorization codes, the first code being indicative of status as an authorized user of the telecommunications network and the second code being indicative of status as an authorized purchaser of goods and/or services, as in a conventional credit card. The authorization codes are preferably contained on the card in both human readable form, such as embossed lettering, and in machine readable form, such as magnetic encoding. The validity of the first authorization code is verified to permit access to the telecommunications network. After the telecommunications network has been used to complete a call, billing information including the first billing code that has been recorded locally is transmitted via the telecommunications network to a remote data base in which the second billing code associated with the first billing code is automatically determined in data processing equipment.

Abstract: An Inverse Multiplexer is disclosed which first demultiplexes a first data rate input signal into a plurality of second lower data rate subsectional signals, where each subsectional signal is provided with a periodic synchronization marker and includes a data rate which is less than the channel data rate used to transmit that subsectional signal to a remote terminal. Programmable Multiplexers (PMUXs) then operate to each take one or more subsectional signals that are (1) clock synchronized to a PMUX clock, and (2) a rational fraction of the channel data rate, and map contiguously assigned time slots in a capacity domain frame for each subsectional signal to time slots of a time domain frame format using a 2-step or 3-step digit reverse technique. The resultant time domain format has the input subsectional capacity domain time slots substantially uniformly distributed over the time domain frame.

Abstract: Method and apparatus are disclosed for a programmable multiplexer for converting each of one or more input data signals, with individual data rates, into a second higher data rate signal by mapping contiguously arbitarily assigned Capacity Time Slots (CTSs) in a capacity domain frame to Time Slots in a Time Domain frame (TDTSs) so the TDTSs for each input signal are substantially uniformly spread throughout the TD frame. The programmable multiplexer receives separate input data rate signals which are clock synchronized to the multiplexer clock from separate synchronizers, and maps the capacity domain of the input signals to the time frame format using a 2-step or 3-step digit reverse technique. Both techniques decompose the capacity domain address into predetermined digits from predetermined number bases and then combine the digits to perform a similar computation using the number bases in reverse order.

Abstract: In the present invention, a first agent terminal, of a group of multi-windowing agent terminals, answers an integrated voice/data call by (i) directly receiving and displaying call-related information from a communication system in a "Phone" window, and (ii) retrieving data information from a host database system, and then transferring the voice portion and either (a) predetermined received call-related information, or (b) a current data display status indication directly to a second agent terminal of a second group of agent terminals. The second agent terminal uses the received data to access a same or other host database system for automatically displaying the pertinent data at the second agent terminal to continue the voice and data call.

Abstract: The present invention enables multi-windowing agent computer terminals, that answer a variety of, e.g., telemarketing calls, to (i) receive call-related information from a communication system at an agent terminal for automatic display in a "telephony" window, and (ii) automatically send such information to a host computer system application software and retrieve caller desired information based on the calling or called number for display in a "host application" window without requiring a caller to provide verbal information. The present invention also enables an agent terminal to automatically transfer caller-related information from one window to another window via programmable function key commands which can be programmed by a user/system administrator regarding what information is to be copied and where and when to copy it.

Abstract: A video signal encoder uses three dimensional transform coding on blocks of intensity values and then selects the most significant coefficients for further processing by separating the coefficients into several groups have approximately the same energy and thus approximately the same significance. The intensity values can be error values computed by forming the difference between the actual intensity values and values computed using conditional motion compensated interpolation or motion compensated prediction. The encoder can use both two dimensional and three dimensional transformations, and if desired, adaptively switch between those processing techniques. A complementary decoder recovers the intensity values from encoded signals.

Abstract: A data receiver which includes a timing recovery circuit or periodically adjusting a timing signal used to control sampling of the received signal is arranged to reduce or eliminate phase jitter introduced as a result of the adjustments. Apparatus and a method are described for forming a correction factor which is a joint function of the timing adjustment increment and a previous correction factor. Typically, correction is accomplished by forming the complex product of the in-phase and quadrature-phase components of the received signal and the recursively updated correction factor.

Abstract: In accordance with the present invention, local loop testing in an echo canceller based modem is provided in a manner in which proper operation of both the receiver and echo canceller can be assessed simultaneously. Specifically, the modem includes means for generating data and interfering symbol sequences and for applying the interfering sequence to both the echo canceller and the modem transmitter, while applying the data sequence only to either the transmitter or the echo canceller. During closed loop testing, the modem connection to the transmission line is opened, and a portion of the output of the modem transmitter is applied to the modem receiver via a leakage path through the hybrid. The received symbol is subtracted from the echo canceller output, resulting in application to the receiver of only a version of the data symbol sequence. Advanced knowledge of the characteristics of the data symbol sequence allows proper receiver operation to be verified.

Abstract: An FSK demodulator of the differential type is arranged to provide two versions of an input FSK signal that, relative each other, (1) have a linear phase shift versus frequency characteristic and (2) are phase shifted by 90 degrees at a center frequency halfway between the upper and lower FSK frequencies. The linear phase shift is provided by an N sample delay element, while a constant phase shift circuit which includes a K factor gain element provides the 90 degree phase shift. By virtue of this arrangement, the demodulator characteristics can be adaptively varied and the demodulator tuned simply by changing the value of K. A similar adaptive technique is applied to a threshold detector arranged to position a threshold midway between the extremes of an input signal.

Abstract: An FSK demodulator of the differential type is arranged to provide two versions of an input FSK signal that, relative each other, (1) have a linear phase shift versus frequency characteristic and (2) are phase shifted by 90 degrees at a center frequency halfway between the upper and lower FSK frequencies. The linear phase shift is provided by an N sample delay element, while a constant phase shift circuit which includes a K factor gain element provides the 90 degree phase shift. By virtue of this arrangement, the demodulator characteristics can be adaptively varied and the demodulator tuned simply by changing the value of K. A similar adaptive technique is applied to a threshold detector arranged to position a threshold midway between the extremes of an input signal.

Abstract: In order to cool heat generating components on an electronic assembly without contamination by particles carried in a cooling gas stream, a cold plate is disposed adjacent the assembly, defining a gap region therebetween. The cooling gas is directed in proximity to the cold plate, dissipating heat transferred from the components to the plate. To enhance overall heat dissipation of the apparatus, an aperture is defined between the gap region and the gas stream, so that a negative pressure is induced in the gap by virtue of the Bernoulli principle. This pressure difference induces convection within the gap, increasing heat flow between the components and the plate, and resulting in overall improvement in heat dissipation. Embodiments in which the gas stream flow is parallel and transverse to the cold plate are described.

Abstract: Apparatus and a technique for equalizing non-linear distortion in a received modulated data signal by (1) forming tentative decisions as to the values of data symbols represented by the signal preferably using a receiver including a conventional linear equalizer, (2) constructing a replica of the non-linear distortion in response to the tentative decisions, and (3) forming a final decision as to the data symbol values in response to signals including the replica. If desired, the final decisions can also be stored and fed back to the processor which forms the replica, so that the replica is a joint function of past final decisions and future tentative decisions regarding the data symbols represented by the signal samples. The present invention provides increased accuracy by using tentative decisions rather than input samples to form the non-linear distortion replica.

Abstract: The occurrence of a transmission error which produces a multibit error burst in the decrypted version of a composite speech and synchronization encrypted signal using a one-bit cipher feedback or similar encryption scheme is detected by monitoring the decrypted signal to detect synchronization errors. Upon detection of an error, the decrypted speech output is muted or disabled to avoid the annoying audible click otherwise produced.