Pensando Systems Inc.

Balakrishnan Raman

Murty Subba Rama Chandra Kotha

Harinadh Nagulapalli

Vishwas Danivas

Sanjay Shanbhogue

Krishna Doddapaneni

Sarat Kamisetty

Pirabhu Raman

Chaitanya Huilgol

Varada Raja Kumar Kari

Akshaya Nadahalli

Varagur Chandrasekaran

Prabu Thayalan

Vijay Sampath

Maruthi Ram Namburu

Allen Hubbe

John Cruz

Dontula Venkata Ratnananda Ganesh

Rathina Sabapathy Sabesan

Michael Brian Galles

Vipin Jain

Raja Rao Tadimeti

Madhava Rao Cheethirala

Sameer Kittur Subrahmanya

Neel C Patel

Tuyen Van Quoc

J Bradley Smith

Hareesh Ramachandran

Andrew Boyer

Arun Selvarajan

Chandrasekaran Swaminathan

Akshay Nadahalli

Nitish Bhat

A networking device can use sync request packets to synchronize connection objects with a peer. Real or virtual flow misses occurring in the data plane while processing a network packet indicate that synchronization is needed. A virtual flow miss occurs when a data plane connection object is obsolete. A sync request packet containing connection metadata and the network packets is sent to the peer. The peer uses the connection metadata to create or update its version of the connection object then returns a sync ack packet that contains the network packet. Receiving a sync ack packet can indicate that the locally stored version of the connection object and the peer's version of the connection object are consistent. The network packet may be processed normally when the locally stored version of the connection object and the peer's version of the connection object are consistent.

Systems and methods for flow configuration syncing between network infrastructure devices

SR-IOV (single root IO virtualization) capable PCIe devices can implement virtual functions (VFs) that are assigned to VMs running on a host machine, thereby speeding IO operation by writing directly to the VMs' memory while bypassing the hypervisor managing the VMs. As such, VFs thwart the dirty page tracking that hypervisors use to minimize VM downtime when the VM is migrated between hosts. The SR-IOV PCIe devices can help resolve this problem by maintaining dirty page tracking data for VMs running on the host machine. The SR-IOV PCIe devices bypassing the hypervisor while writing into a memory page of the VM can set the dirty page tracking data to indicate the memory pages that are dirty (i.e., written to by the VF), and can provide access to the dirty page tracking data. The hypervisor can thereby obtain and use the dirty page tracking data.

Methods and systems for using a peripheral device to assist virtual machine IO memory access tracking

A networking device can include a packet processing pipeline circuit and a processor. The packet processing pipeline circuit can be configured to implement a data plane and the processor can be configured to implement a control plane. The packet processing pipeline circuit and the processor can also be configured to send a plurality of heartbeat packets on multiple paths to a second networking device. The data plane can produce and send the heartbeat packets to the second networking device within a heartbeat period. The data plane may send a second plurality of heartbeat packets on multiple paths to the second networking device with a second heartbeat period. The heartbeat packets can have unique packet five tuples that include an IP address of the second networking device.

Systems and methods for a networking device to send heartbeat packets on multiple paths to a second networking device

HA peers can include networking devices that have data planes and control planes that configure the data plane to use status data in a memory for processing network packets of network flows. The HA peers synchronize the status data such that one peer can take over when another fails. When a HA peer is brought up, data plane syncing synchronizes data for new network flows but not existing network flows. A first bulk sync operation synchronizes data for existing flows but not for new flows. A second bulk sync operation can synchronize the data for flows that changed state during the first bulk sync operation. Data plane syncing can sync data for all flows after the first bulk sync operation.

Systems and methods for high availability node pairing without incurring connection drops

Synchronizing the databases maintained by network appliances can support high availability or high throughput topologies, but also consumes the devices' processing resources. To address that resource consumption, the network appliance's packet processing pipeline circuits can process synchronization packets to thereby synchronize the databases. A local data structure can be in a first local state. Processing a network packet can result in changing the local data structure to a second local state. A state sync packet can include state transition data that indicates a state difference between the first local state and the second local state. The state sync packet can be sent to a peer device that is configured to process the state transition data using the peer device's packet processing pipeline circuit. The peer device's packet processing pipeline can use the state transition data to update a peer device data structure that is in the peer device.

Methods and systems for distributed high speed state synchronization

Data centers often run long lived services such as web servers that are intended to run for hours, days, or even longer before being torn down and replaced with another instance of the long-lived service. Currently, many applications are being implemented with microservice architectures that run short lived services that start up, implement an operation, and are then torn down. An aspect of starting up a service is creating administrative data structures such as InfiniBand queue pairs. A packet processing pipeline having a DMA output stage can be configured to create the administrative data structures, thereby increasing the rate at which the administrative data structures are created. As a result, services running in data centers can be started up more rapidly and efficiently.

Methods and systems for using a packet processing pipeline to accelerate InfiniBand administrative operations

Network traffic flows can be processed by routers, switches, or service nodes. Service nodes may be ASICs that can provide the functionality of a switch or a router. Service nodes can be configured in a circular replication chain, thereby providing benefits such as high reliability. The service nodes can implement methods that include receiving a first packet that includes a source address in a source address field and that includes a destination address in a destination address field, routing the first packet to a selected service node that is in a circular replication chain that includes a plurality of service nodes that have local flow tables and are configured for chain replication of the local flow tables, producing a second packet by using a matching flow table entry of the first packet to process the first packet, and sending the second packet toward a destination indicated by the destination address.

Methods and systems for service state replication using original data packets

Methods and systems for implementing traffic mirroring for network telemetry are disclosed. An embodiment of a method for implementing traffic mirroring for network telemetry involves identifying network traffic at a network appliance that is to be subjected to traffic mirroring for network telemetry, and selecting from available options of transmitting enhanced mirrored network traffic from the network appliance to a collector, wherein the enhanced mirrored network traffic is generated at the network appliance by at least one of compressing and encrypting the network traffic, and transmitting mirrored network traffic from the network appliance to the collector without compressing or encrypting the network traffic.

Methods and systems for implementing traffic mirroring for network telemetry

InfiniBand transport protocol today supports RDMA operations such as read and write with each operation having an opcode defined in the InfiniBand standard. Currently, new RDMA operations require extending the transport protocol by defining a new opcode, its respective header and enhancing InfiniBand implementations to support this new behavior. A more robust way of extending RDMA without requiring an expanding set of opcodes is to register computer code by associating it with a code key similar to a memory key. An InfiniBand channel adapter receiving an RDMA request that includes a code key executes the associated computer code, perhaps compiling it first, in response to receiving the RDMA request. The RDMA response returned to the requester includes an execution result indicating an outcome of executing the executable computer code.

Methods and systems for extending RDMA for dynamic/programmable remote operation registration and invocation

The InfiniBand transport protocol supports the concept of a SRQ (shared receive queue) by which multiple QPs (queue pairs) can share the same receive queue resources. According to the InfiniBand Specification, when a SRQ is enabled, flow control needs to be disabled. The lack of flow control mechanism results in there being no fairness guarantees across multiple requesters. Fairness across requesters can be obtained by implementing a SRQ configured to receive request messages from requesters initiating transactions that consume WQEs (work queue elements) of the SRQ, monitoring consumption of the WQEs by the requesters, determining that a requester has a WQE consumption exceeding a policing threshold, and in response to determining that the WQE consumption of the requester exceeds the policing threshold, sending a response message to the requester that results in reducing the WQE consumption of the requester.

Methods and systems for fairness across RDMA requesters using a shared receive queue

Virtual functions (VFs) running on SR-IOV (single root IO virtualization) capable PCIe devices can migrate in association with VMs using the VFs. A SR-IOV capable PCIe device installed in a host computer can implement the VFs. A VM running on the host and associated with the VF can use the VF to obtain a service such as network communications or access to a NAS device. Migrating the VF in association with the VM can include halting the VM in a VM state on the host, halting the VF in a PCIe state and then obtaining a PCIe state data, restarting the VF in the PCIe state on a second PCIe device of a second host based on the PCIe state data, and restarting the VM in the VM state on the second host, wherein the VM is configured to use the VF on the second PCIe device.

Methods and systems for migrating virtual functions in association with virtual machines

A multitude of data transfer queues can have data transfer operations that are scheduled for a processing circuit to perform. Some of the data transfer queues may submit so many or such large data transfer operations that others receive little or no attention. The situation can be resolved in the data plane via a processing circuit that performs the data transfer operations in conjunction with priority evaluation operations that can assign the data transfer queues to different scheduler priority classes. A scheduler can schedule data transfer operations based on the scheduler priority classes of the data transfer queues.

Methods and systems for adaptive network quality of service for latency critical applications

Load balancing of NVMe targets based on real time metrics can be obtained for NAS appliances mirroring a namespace by assigning the NAS appliances to service sets that include an active load balancing set, a monitored inactive set, and an out of service set. Storage performance metrics of the NAS appliances can be tracked by monitoring IO operations for accessing a NAS mirroring the namespace in a non-volatile memory. Based on the storage metrics, NAS appliances can be moved from one of the service sets to another. Dummy IO operations can be used to track the storage performance metrics of monitored inactive NAS appliances such that a monitored inactive NAS may be moved to the active load balancing set when certain performance constraints are met.

Methods and systems for NVMe target load balancing based on real time metrics

Increased fairness for small vs large NVMe IO commands for accessing a non-volatile memory namespace provided by a network attached storage appliance can be realized by placing NVMe submissions received by a NVMe SQ on a first fabric queue set or a second fabric queue set based on a fairness policy. The first fabric queue set accesses the namespace via a first fabric connection. The second fabric queue set accesses the namespace via a second fabric connection. Accessing the namespace via the fabric connections results in NVMe completions that are merged from the fabric queue sets onto an NVMe completion queue. A process producing the NVMe submissions and receiving the resulting NVMe completions may be unaware of the multiple fabric queue sets.

Growing community of inventors

Fremont, CA, United States of America

Balakrishnan Raman