Cisco Technology, Inc.

Harsha Bharadwaj

Prabesh Babu Nanjundaiah

Sunil John Varghese

Ramesh Sivakolundu

Ankan Ghosh

Edward D Mazurek

Santhosh Hanumana Gowda

Nagendran Venugopalan

Ayan Banerjee

Subrata Banerjee

Kishan Kumar Kotari Uppunda

Joy J Chatterjee

Venu Gopal Tummala

Rajesh L G

Badrinarayanan Ramaswamy

J Michel Metz

Sumanth Sindiri

Pramod Menon

Paresh Gupta

A method of congestion mitigation may include determining whether a host is sending a read command or a write command to an NVMe controller, and in response to a determination that the host is sending the read command, transmitting the read command via a first transmission control protocol (TCP) connection between the host and the NVMe controller. The method may further include in response to a determination that the host is sending the write command, transmitting the write command via a second TCP connection between the host and the NVMe controller.

Storage aware congestion management

This disclosure describes techniques for performing Internet Protocol (IP) switch-based Storage Area Networks (SANS) analytics management. An example method is performed by an IP switch. The method includes identifying information associated with Non-Volatile Memory express (NVMe)/Transport Control Protocol (TCP) Protocol Data Units (PDUs) examined from an Application-Specific Integrated Circuit (ASIC) data path. The information can be stored based on the NVMe/TCP PDUs being included in separate corresponding TCP data segments, according to a handshake protocol. Alternatively or additionally, the information can be stored based on a metadata header being received as a first PDU in a TCP segment. The NVMe/TCP PDUs can be used to extract the information to populate connection and IO tables inside an ASIC. The storage I/O metrics may then be available from the tables.

Internet protocol (IP) switch-based analytics management

Techniques for moving buffers between ports, or virtual lanes of a port, of a networking device of a credited network while maintaining the ports in an active state without dropping any frames. The techniques may include determining that a number of buffers are to be reallocated from a first port of a networking device to a second port of the networking device. The techniques may also include causing a peer port connected to the first port to decrement, by the number, a transmit credit counter associated with the peer port. Based at least in part on determining that the peer port decremented the transmit credit counter, the first port may release the number of the buffers from a buffer pool associated with the first port, and the number of the buffers may be reallocated to the second port.

Non-disruptive trading of buffers between ports or port virtual lanes of a credited network

A method comprises: at a management entity configured to monitor a transceiver system having a transceiver to receive a signal that conveys data frames transmitted by a peer transceiver over a connection: collecting time series values for operating parameters of the transceiver associated with received data frames, including (i) a receive power, and (ii) counts of different media access control (MAC) layer frame errors that respectively indicate degradation levels for system performance ranked from highest to lowest; using the time series values, performing correlations of the receive power against the counts of the different MAC layer frame errors according to a correlation hierarchy that indicates which of the correlations correspond to which of the degradation levels, to produce correlation results; and responsive to the correlation results, determining a degraded component of the transceiver system.

Correlating transceiver parameters for insight into transceiver health

A method includes measuring input/output traffic for respective hosts that are connected to a Fibre Channel N_Port Virtualizer (FC-NPV) switch, which is in communication with a first N_Port ID Virtualization (NPIV) core switch via a first port channel and with a second NPIV core switch via a second port channel; determining that traffic carried on the first port channel between the FC-NPV switch and the first NPIV Core switch exceeds a predetermined threshold compared to traffic carried on the second port channel; and re-assigning traffic from a given host carried on the first port channel to the second port channel between the FC-NPV switch and the second NPIV core switch.

Host input/output based load balancing on fibre channel N_port virtualizer switch uplinks

An initiator emulator is implemented on a control plane of a switch fabric connected to target ports of a storage array having storage configured with logical partitions. After an initiator port of a server logs into the switch fabric and is blocked from discovering the target ports, the initiator emulator, acting as proxy for the initiator port, discovers information that indicates logical partition masking enforced at the target ports for the initiator port. The initiator emulator determines allowed (initiator (I), target (T)) (I, T) port combinations that should be allowed access via the switch fabric based on the information from the discovery. The initiator emulator configures the switch fabric with one or more zones based on the allowed (I, T) port combinations. The initiator emulator then sends to the initiator port an indication of a zone change to the switch fabric.

Complete autozoning in fibre channel SANs

A method is performed at a switch fabric that communicates with a storage array target port. The method includes sending frames to the target port responsive to receiving buffer-to-buffer (B2B) credits that indicate a receive buffer at the target port is available for the frames. The method further includes, in response to detecting a credit stall at the target port, operating in a virtual lane mode. The operating in the virtual lane mode includes: determining whether a frame destined for the target port is a command frame or a data frame; based on the determining, marking the frame to indicate that the frame is destined for a particular virtual lane among virtual lanes of the receive buffer; and receiving from the target port a per-virtual lane B2B credit that indicates the particular virtual lane is available and, in responsive, sending the frame to that virtual lane on the target port.

Use of virtual lanes to solve credit stall on target ports in FC SAN

One embodiment is a method including configuring a first network element of a fibre channel ('FC') network as a generator element, wherein the generator employs a link diagnostic protocol to cause a second network element comprising a peer of the first network element as a reflector element, wherein the first and second elements are connected via a link; entering a first diagnostic phase, wherein in the first diagnostic phase, diagnostic capabilities of the first and second elements are determined; and subsequent to completion of the first diagnostic phase, entering a second diagnostic phase in which a deep loopback test is performed, wherein the deep loopback test comprises a frame level loopback test for exposing an issue in a path between the first and second network elements beyond a Media Access Control ('MAC') layer.

Techniques for using frame deep loopback capabilities for extended link diagnostics in fibre channel storage area networks

In one embodiment, a device obtains one or more fabric port (F-port) counters and one or more extender port (E-port) counters in a storage area network (SAN). The device inputs the obtained F-port and E-port counters to a machine learning-based prediction model. The device uses the prediction model to predict a slow drain condition in the SAN, based on the counters input to the model. The device initiates a corrective measure in the SAN, based on the predicted slow drain condition.

Using machine learning to detect slow drain conditions in a storage area network

One embodiment is a method and includes periodically polling a plurality of interface counters associated with each of an edge port and an Inter-Switch Link ('ISL') port of a first fibre channel ('FC') switch, wherein a target device is connected to the edge port of the first FC switch, and a plurality of interface counters associated with ISL port of a second FC switch, wherein the ISL port of each of the first and second FC switches are connected to one another via an ISL; determining based on the polling of the various counters whether several conditions have been met for a predetermined number of times and if so, characterizing the edge port as a level 1 slow drain port and taking remedial action based on the characterization.

Just-in-time identification of slow drain devices in a fibre channel network

A method is performed at a Fibre Channel (FC) switch of an FC switch fabric through which servers connected to the FC switch fabric access logical partitions of a storage array connected to the FC switch fabric. The FC switch receives from the storage array information indicative of port zoning rules and logical partition masking that collectively define which server ports are permitted access to which storage array ports and to which logical partitions of the storage array. The FC switch generates from the information authorization rules for enforcing the port zoning rules and the logical partition masking, and programs the authorization rules into memory. The FC switch receives FC frames from the server ports. The FC frames convey respective input-output (IO) operations destined for the logical partitions. The FC switch authorizes each IO operation based on a lookup of the programmed authorization.

Fine granularity authorization control in FC-NVMe storage area network

A Fibre Channel (FC) or FC-over-Ethernet (FCoE) switch has ports to forward Input-Output (IO) requests, and service data transfers, between end devices in a storage area network. The switch receives at a port a time ordered sequence of IO requests for data transfers to be serviced by the port. Each IO request including a data length of the data transfer. The switch detects a microburst on the port for each IO request. To do this, the switch parses the IO request to retrieve the data length, determines a transfer time required to transfer the data length over the port, upon receiving a next IO request, determine whether a time interval between the IO request and the next IO request is less than the transfer time, and if the time interval is less than the transfer time, declaring a microburst on the port, otherwise not declaring a microburst.

Detecting and handling solicited IO traffic microbursts in a fibre channel storage area network

A host bus adapter of a target device is associated with a Fibre Channel driver and is connected to a Fibre Channel switch fabric. The host bus adapter receives from the switch fabric an inbound frame having a header including a source identifier and a virtual machine (VM) tag, stores a mapping between the source identifier and the VM tag, and passes the inbound frame to the Fibre Channel driver. The host bus adapter receives from the Fibre Channel driver an outbound frame having a header including a destination identifier, and determines, based on the mapping, whether there is a match at least between the source identifier of the inbound frame and the destination identifier of the outbound frame. If there is a match, the host bus adapter tags the header of the outbound frame with the VM tag, and transmits the tagged outbound frame to the switch fabric.

Priority tagging based solutions in fc sans independent of target priority tagging capability

A Fiber Channel (FC) or FC-over-Ethernet (FCoE) switch has ports to forward Input-Output (IO) requests, and service data transfers, between end devices in a storage area network. The switch receives at a port a time ordered sequence of IO requests for data transfers to be serviced by the port. Each IO request including a data length of the data transfer. The switch detects a microburst on the port for each IO request. To do this, the switch parses the IO request to retrieve the data length, determines a transfer time required to transfer the data length over the port, upon receiving a next IO request, determine whether a time interval between the IO request and the next IO request is less than the transfer time, and if the time interval is less than the transfer time, declaring a microburst on the port, otherwise not declaring a microburst.

Growing community of inventors

Bengaluru, India

Harsha Bharadwaj