Abstract: Technologies for controlling jitter at network packet egress at a source computing device include determining a switch time delta as a difference between a present switch time and a previously captured switch time upon receipt of a network packet scheduled for transmission to a target computing device and determining a host scheduler time delta as a difference between a host scheduler timestamp associated with the received network packet and a previously captured host scheduler timestamp. The source computing device is additionally configured to determine an amount of previously captured tokens present in a token bucket, determine whether there are a sufficient number of tokens available in the token bucket to transmit the received packet as a function of the switch time delta, the host scheduler time delta, and the amount of previously captured tokens present in the token bucket, and schedule the received network packet for transmission upon a determination that sufficient tokens in the token bucket.

Abstract: Technologies for controlling jitter at network packet egress at a source computing device include determining a switch time delta as a difference between a present switch time and a previously captured switch time upon receipt of a network packet scheduled for transmission to a target computing device and determining a host scheduler time delta as a difference between a host scheduler timestamp associated with the received network packet and a previously captured host scheduler timestamp. The source computing device is additionally configured to determine an amount of previously captured tokens present in a token bucket, determine whether there are a sufficient number of tokens available in the token bucket to transmit the received packet as a function of the switch time delta, the host scheduler time delta, and the amount of previously captured tokens present in the token bucket, and schedule the received network packet for transmission upon a determination that sufficient tokens in the token bucket.

Abstract: In described embodiments, a timestamp generator includes a fixed clock domain driven by a fixed frequency clock, a core clock domain, coupled to the fixed clock domain, which is driven by a core clock whose frequency is adjustable during an operation of the timestamp generator. A timestamp logic operating in the core clock domain generates a timestamping output of the timestamp generator. A rate generator operating in both the fixed clock domain and the core clock domain generates per clock cycle increments in the fixed clock domain and transfers carry units from the fixed clock domain into the core clock domain, and a timestamp increment generation of the timestamp logic is clocked by the fixed frequency clock provided by the rate generator. A method for enabling timestamp in an ASIC to be accurate with system clock changes is also described.

Abstract: In one embodiment, queues associated with a first traffic class (FTC) are selected for service. Each FTC queue having at least one enqueued cell is identified as an occupied FTC queue, Where at least one FTC queue is provisioned for burst scheduling of multiple cells when serviced. An occupied FTC queue provisioned for burst scheduling is identified as a super-occupied FTC queue when the number of cells enqueued is greater than a specified number. Each occupied FTC queue is set as eligible for service based on a FTC scheduling algorithm. An eligible FTC queue is selected for service based on a corresponding sub-priority of each eligible FTC queue. Each FTC queue is assigned a sub-priority based on a service level of a connection associated with enqueued cells. When the super-occupied queue is serviced, the number of cells dequeued is based on a burst size.

Abstract: A scheduler allocates service to enqueued cells of connections provisioned for guaranteed service levels employing a structure with one or more of the following features to achieve efficient high-speed packet switching (cell relay). For very high-speed switching of connections, such switches operating up to 10, or even 100, Tbps, burst scheduling of cells is employed in which a number of cells, termed a burst, are serviced when a queue is eligible for service. When a queue has less than the number of cells in the burst (termed a short burst), the scheduler still schedules service, but accounts for saved service time (or bandwidth) of the short burst via queue length when the queue's eligibility is next considered for service. In addition, high and low bandwidth connections of a queue may be allocated into two sub-queues, with priority assigned to the two queues and delay-sensitive traffic (high bandwidth connections) assigned to the higher priority sub-queue.