Abstract: Embodiments include methods to manage caches in a network device. In one embodiment, a method is performed by a network device and the method comprises determining a first cache of the network device is overloaded, where the first cache is within a cache hierarchy including a plurality of caches, each cache including a set of forwarding table entries for the network device to forward packets of traffic flows, and where one cache lower in the cache hierarchy is to be checked earlier in flow lookup than one higher in the cache hierarchy. The method further comprises upon determining that the first cache is overloaded, disabling the first cache from the flow lookup, where the flow lookup skips the first cache and is performed at a second cache higher in the cache hierarchy and disabling the first cache from forwarding table entry insertion when the forwarding table entry insertion is performed in one or more caches of the plurality of caches.

Abstract: A software switch and a method performed by the software switch are disclosed. The software switch receives, from a node deploying a virtual machine, a request for a virtual port to be polled by the virtual machine. The request includes a Central Processing Unit “CPU” identity identifying a CPU on which the virtual machine executes. The request includes an indication of a clock frequency at which the CPU is set to operate. The software switch determines a number of packets in a queue associated with the virtual port. The software switch adjusts the clock frequency of the CPU based on the number of packets in the queue. A corresponding computer program and a computer program carrier are also disclosed.

Abstract: A method is implemented by a network device for orchestrating execution of a polling thread of a software-based switching program on a heterogeneous multicore processor. The method includes causing the polling thread to be executed on a first processing core in a first cluster of a plurality of clusters of processing cores, determining a value indicative of a number of active processing cycles used by the polling thread, determining whether the value is higher than a high threshold associated with the first processing core or lower than a low threshold associated with the first processing core, and if so causing the polling thread to be moved to a second processing core in a second cluster of the plurality of clusters, where the second processing core has a different processing capacity than the first processing core.

Abstract: A method is implemented by a network device for orchestrating execution of a polling thread of a software-based switching program on a heterogeneous multicore processor. The method includes causing the polling thread to be executed on a first processing core in a first cluster of a plurality of clusters of processing cores, determining a value indicative of a number of active processing cycles used by the polling thread, determining whether the value is higher than a high threshold associated with the first processing core or lower than a low threshold associated with the first processing core, and if so causing the polling thread to be moved to a second processing core in a second cluster of the plurality of clusters, where the second processing core has a different processing capacity than the first processing core.

Abstract: A software switch and a method performed by the software switch are disclosed. The software switch receives, from a node deploying a virtual machine, a request for a virtual port to be polled by the virtual machine. The request includes a Central Processing Unit “CPU” identity identifying a CPU on which the virtual machine executes. The request includes an indication of a clock frequency at which the CPU is set to operate. The software switch determines a number of packets in a queue associated with the virtual port. The software switch adjusts the clock frequency of the CPU based on the number of packets in the queue. A corresponding computer program and a computer program carrier are also disclosed.

Abstract: A method in a first network device of performing a software update of a line card of a second network device without disruption to data traffic. The method includes causing a redundant control plane component of the second network device to be updated according to the software update. The method continues with causing the second network device to instantiate, based on the software update, a line card virtual machine (LC VM) as a redundant data plane component for the line card. The method further includes causing a third network device to forward data traffic to both the line card and the LC VM of the second network device, and causing the second network device to update the line card according to the software update while processing the received data traffic using the LC VM.

Abstract: A method in a first network device of performing a software update of a line card of a second network device without disruption to data traffic. The method includes causing a redundant control plane component of the second network device to be updated according to the software update. The method continues with causing the second network device to instantiate, based on the software update, a line card virtual machine (LC VM) as a redundant data plane component for the line card. The method further includes causing a third network device to forward data traffic to both the line card and the LC VM of the second network device, and causing the second network device to update the line card according to the software update while processing the received data traffic using the LC VM.

Abstract: Exemplary methods for upgrading route processor virtual machines (RP VMs) and line card virtual machines (LC VMs) include upgrading a second RP VM serving as a standby RPVM with a new RP VM image. The methods further include spawning, using a new LC VM image, a second set of LC VMs to serve as standby LC VMs. The methods also include performing incubation of standby VMs, by configuring a switch to forward data traffic to both the first and second set of LC VMs, thereby allowing system states of the standby VMs to be auto-populated by the data traffic. The methods further include, after the incubation has completed, swapping roles of VMs at the virtual router, such that data traffic are exchanged only between the switch and the upgraded VMs.

Abstract: Exemplary methods for upgrading route processor virtual machines (RP VMs) and line card virtual machines (LC VMs) include upgrading a second RP VM serving as a standby RPVM with a new RP VM image. The methods further include spawning, using a new LC VM image, a second set of LC VMs to serve as standby LC VMs. The methods also include performing incubation of standby VMs, by configuring a switch to forward data traffic to both the first and second set of LC VMs, thereby allowing system states of the standby VMs to be auto-populated by the data traffic. The methods further include, after the incubation has completed, swapping roles of VMs at the virtual router, such that data traffic are exchanged only between the switch and the upgraded VMs.