Abstract: Embodiments herein provide more efficient, more flexible, and more cost-effective ways to provide additional and/or increased functionality to an information handling system. Presented herein are embodiments of an application acceleration port interface module (which embodiments may be referred to herein for convenience as “AAPIM”) into which pluggable I/O (input/output) modules may be inserted and the other ends inserted into ports of an information handling system to provide the information handling system with increase capabilities (e.g., increased resource, such as added processing, and increased services, such as new applications or accelerated services). AAPIM embodiments are versatile solutions to address application acceleration needs that can be quickly reprogrammed to address specific needs of a user.

Abstract: Described herein are systems and methods for host multihoming with no state synchronization between top-of-rack (ToR) switches coupled to multiple hosts. ToR switches of a multi-homing system share a virtual MAC address and respond to Address Resolution Protocol (ARP) requests and/or Neighbor Discovery (ND) solicitations for a default gateway IP address from the host with the virtual MAC address. Ports on a ToR switch may be configured either with a static ARP and/or ND entry, or be configured to learn via a discovery protocol. The lightweight host multihoming system may have modified flooding behavior, ARP/ND handling, and data path forwarding. ARP/ND traffic from a host is processed by a ToR switch, while other multicast traffic may be discarded. Embodiments of the host multihoming system provide a lightweight solution for software implementation with minimal changes imposed on the host or network design.

Abstract: Embodiments are directed to a process and system for avoiding temporary traffic loss in a network having first and second computers and provider edge devices, by retaining, in each provider edge device, path information including a MAC address for the first host computer in a control plane as a dynamic address, and maintaining, in each provider edge device, its own routing table listing network addresses of the other provider edge devices as multi-homed peers. A failure recovery process in each provider edge devices advertises all multi-homed peer network addresses as addresses that each respective provider edge device has dynamically learnt in the control plane such that when a provider edge device receives a route with a locally connected host, it re-advertises the host address on its own.

Abstract: Described herein are systems and methods for host multihoming with no state synchronization between top-of-rack (ToR) switches coupled to multiple hosts. ToR switches of a multi-homing system share a virtual MAC address and respond to Address Resolution Protocol (ARP) requests and/or Neighbor Discovery (ND) solicitations for a default gateway IP address from the host with the virtual MAC address. Ports on a ToR switch may be configured either with a static ARP and/or ND entry, or be configured to learn via a discovery protocol. The lightweight host multihoming system may have modified flooding behavior, ARP/ND handling, and data path forwarding. ARP/ND traffic from a host is processed by a ToR switch, while other multicast traffic may be discarded. Embodiments of the host multihoming system provide a lightweight solution for software implementation with minimal changes imposed on the host or network design.

Abstract: A hash collision mitigation system includes a source device coupled to a destination device by a networking device. The networking device receives a first packet from the source device, and uses first packet information included in the first packet to perform a hash operation on the first packet and generate a hash value. The networking device then determines that the hash value has caused a hash collision with a first hash value forwarding entry in a hash forwarding table in the networking device and, in response, generates a first Policy Based Routing (PBR) entry for the first packet. The first PBR entry is configured to provide for the forwarding of at least one second packet received from the destination device and directed to the source device, and that first PBR is stored in a Ternary Content Addressable Memory (TCAM) in the networking device.

Abstract: A method of identifying persistently congested queues among a plurality of ports on a switch may include polling a queue length associated with each port; determining that a queue associated with a first port is persistently congested; and initiating and maintaining egress sFlow monitoring of the first port when the queue associated with a first port is persistently congested until the queue associated with the first port is no longer persistently congested.

Abstract: A hash collision mitigation system includes a source device coupled to a destination device by a networking device. The networking device receives a first packet from the source device, and uses first packet information included in the first packet to perform a hash operation on the first packet and generate a hash value. The networking device then determines that the hash value has caused a hash collision with a first hash value forwarding entry in a hash forwarding table in the networking device and, in response, generates a first Policy Based Routing (PBR) entry for the first packet. The first PBR entry is configured to provide for the forwarding of at least one second packet received from the destination device and directed to the source device, and that first PBR is stored in a Ternary Content Addressable Memory (TCAM) in the networking device.

Abstract: A microburst visibility system includes a switch device coupled to each sender device(s) and receiver device(s). The switch device receives flows transmitted by the sender device(s) to the receiver device(s). For each flow that include data packets that are stored in a first queue, the switch device provides a first queue identifier for the first queue in a flow record associated with that flow and modifies, for each data packet that is transmitted as part of that flow, flow information in data packet counter(s) that are associated with the first queue identifier in the flow record associated with that flow. When the switch device detects a microburst in the first queue, it captures the flow information in each data packet counter that is associated with the first queue identifier in each of the flow records included in the switch device, and provides it for analysis.

Abstract: Embodiments of the present invention include systems and methods for adjusting RED configuration according to the available buffer space for a queue in a switch. In one or more embodiments, a method comprises the steps of: initializing minimum and maximum thresholds of RED associated with a queue; determining an available free space for the queue, wherein a data packet for the queue is discarded by a dynamic thresholding when a length of the queue reaches the available free space; determining an allowable free space (AFS) for the queue that is a multiplication of the available free space to an allowance factor (AF); and, when the length of the queue reaches the AFS, calculating a ratio of the minimum threshold to the maximum threshold and updating the maximum threshold to the AFS and updating the minimum threshold to a multiplication of the ratio to the AFS.

Abstract: Embodiments of the present invention include systems and methods for adjusting RED configuration according to the available buffer space for a queue in a switch. In one or more embodiments, a method comprises the steps of: initializing minimum and maximum thresholds of RED associated with a queue; determining an available free space for the queue, wherein a data packet for the queue is discarded by a dynamic thresholding when a length of the queue reaches the available free space; determining an allowable free space (AFS) for the queue that is a multiplication of the available free space to an allowance factor (AF); and, when the length of the queue reaches the AFS, calculating a ratio of the minimum threshold to the maximum threshold and updating the maximum threshold to the AFS and updating the minimum threshold to a multiplication of the ratio to the AFS.

Abstract: A microburst visibility system includes a switch device coupled to each sender device(s) and receiver device(s). The switch device receives flows transmitted by the sender device(s) to the receiver device(s). For each flow that include data packets that are stored in a first queue, the switch device provides a first queue identifier for the first queue in a flow record associated with that flow and modifies, for each data packet that is transmitted as part of that flow, flow information in data packet counter(s) that are associated with the first queue identifier in the flow record associated with that flow. When the switch device detects a microburst in the first queue, it captures the flow information in each data packet counter that is associated with the first queue identifier in each of the flow records included in the switch device, and provides it for analysis.

Abstract: Embodiments are directed to a process and system for avoiding temporary traffic loss in a network having first and second computers and provider edge devices, by retaining, in each provider edge device, path information including a MAC address for the first host computer in a control plane as a dynamic address, and maintaining, in each provider edge device, its own routing table listing network addresses of the other provider edge devices as multi-homed peers. A failure recovery process in each provider edge devices advertises all multi-homed peer network addresses as addresses that each respective provider edge device has dynamically learnt in the control plane such that when a provider edge device receives a route with a locally connected host, it re-advertises the host address on its own.

Abstract: Embodiments are directed to a process and system for recovering from data link failure in a network, by retaining, in a first provider edge device, path information for a first host coupled to a second host over the network in a control plane, upon indication of withdrawal of the first host from a second provider edge device, starting a timer controlling a period of re-advertisement and re-routing to the first host, removing, in the first provider edge device, the first host from a data plane, and generating, in the first provider edge device, a new route to advertise the first host address to cause other provider edge devices to receive the new route and adjust their respective control plane routes prior to expiration of the timer.

Abstract: Embodiments are directed to a process and system for optimizing traffic paths for orphaned hosts in a VXLAN system, by configuring virtual link trunking (VLT) peers to advertise MAC addresses learned from all multi-homed hosts in the system using Anycast VXLAN tunnel endpoint-Internet Protocol address (VTEP-IP); configuring the virtual link trunking (VLT) peers to advertise MAC addresses learned from all single-homed hosts in the system using a secondary VTEP-IP; directing unicast traffic destined to the single-homed hosts to directly connected VLT peers using the Secondary VTEP-IP; and directing Broadcast, unknown unicast, and multi-cast (BUM) traffic destined to the single-homed hosts to directly connected VLT peers using the Inclusive Multicast Ethernet Tag route.

Abstract: Embodiments are directed to a process and system for optimizing traffic paths for orphaned hosts in a VXLAN system, by configuring virtual link trunking (VLT) peers to advertise MAC addresses learned from all multi-homed hosts in the system using Anycast VXLAN tunnel endpoint-Internet Protocol address (VTEP-IP); configuring the virtual link trunking (VLT) peers to advertise MAC addresses learned from all single-homed hosts in the system using a secondary VTEP-IP; directing unicast traffic destined to the single-homed hosts to directly connected VLT peers using the Secondary VTEP-IP; and directing Broadcast, unknown unicast, and multi-cast (BUM) traffic destined to the single-homed hosts to directly connected VLT peers using the Inclusive Multicast Ethernet Tag route.

Abstract: Embodiments are directed to a process and system for recovering from data link failure in a network, by retaining, in a first provider edge device, path information for a first host coupled to a second host over the network in a control plane, upon indication of withdrawal of the first host from a second provider edge device, starting a timer controlling a period of re-advertisement and re-routing to the first host, removing, in the first provider edge device, the first host from a data plane, and generating, in the first provider edge device, a new route to advertise the first host address to cause other provider edge devices to receive the new route and adjust their respective control plane routes prior to expiration of the timer.

Abstract: The present invention relates generally to an information handling system. Aspects of the present invention include an adaptive credit-based flow control in an information handling system. In embodiments, a request/grant credit system can be eliminated since the receiver can dynamically allocate credits based on demand at the sender. In embodiments, the sender can provide information related to its queue size to the receiver. The receiver can estimate queue size and demand based on the estimated queue size. In embodiments, the receiver allocates credits based on sender demand.

Abstract: An N-port, shared-memory switch allocates a shared headroom buffer pool (Ps) for a priority group (PG). Ps is smaller than a worst case headroom buffer pool (Pw), where Pw equals the sum of worst case headrooms corresponding to each port-priority tuple (PPT) associated with the PG. Each worst case headroom comprises headroom required to buffer worst case, post-pause, traffic received on that PPT. Subject to a PPT maximum, each PPT may consume Ps as needed. Because rarely will all PPTs simultaneously experience worst case traffic, Ps may be significantly smaller than Pw, e.g., Ps<(Pw/A) where M>=2. Ps may be size-adjusted based on utilization of Ps, without halting traffic to or from the switch. If Ps utilization exceeds an upper utilization threshold, Ps may be increased, subject to a maximum threshold (Pmax). Conversely, if utilization falls below a lower utilization threshold, Ps may be decreased.

Abstract: Systems and techniques for processing and/or forwarding packets are described. An ingress switch can use a QoS mapping mechanism to map a first set of Quality of Service (QoS) bits in a packet received from a customer to a second set of QoS bits for use in a Transparent Interconnection of Lots of Links (TRILL) packet which encapsulates the packet. The first set of QoS bits can be different from the second set of QoS bits. The TRILL packet can be processed and/or forwarded in the network based on the second set of QoS bits. At the egress switch, the TRILL packet can be decapsulated and the original packet with the original QoS bits (or QoS bits that are different from the original QoS bits) can be forwarded to the customer's network. In this manner, some embodiments of the present invention can preserve the QoS bits across a TRILL network.

Abstract: A system for managing networked devices comprising a plurality of racks of computing devices, each rack computing device further comprising a plurality of constraints and configured to operate one or more hosted systems, each rack comprising a bit vector stored in a data memory defining the minimum available set of constraints for each of the plurality of rack computing devices. A plurality of enterprises, each comprising one or more enterprise computing device configured to operate a virtual network system that is configured to interactively operate with the one or more hosted systems of one of the rack computing devices. A network function virtualization system configured to receive a service request and to determine that a new hosted system is required to respond to the service request, and to select a rack for instantiation of the new hosted system as a function of the plurality of bit vectors.