Crossroads Systems, Inc.

Other

International Business Machines Corporation

Robert A Reynolds

Keith M Arroyo

John B Haechten

Kenneth D Smeltzer

Geoffrey B Hoese

David W Sheffield

Stephen K Wilson

Robert J Scott

Brian R Smith

Neil T Wanamaker

Michael B Comer

John T Mabon

A method and system are disclosed for processing multiple commands to a streaming target device through a SCSI router in a Fibre Channel network provisioned with one or more Fiber Channel hosts. The system and method are implemented in the SCSI router and include creating a FIFO queue, receiving a first command from one of the Fibre Channel hosts in the Fibre Channel network, and forwarding the first command to the streaming target device. When a subsequent command is issued to the streaming target device, the SCSI router can receive the subsequent command and determine whether the streaming target device has completed execution of the previous command. If the streaming target device has completed execution of the previous command, the subsequent command can be forwarded to the streaming target device. If the streaming target device has not completed executing the previous command, the subsequent command can instead be placed in the FIFO queue. Upon execution of the previous command by the streaming target device, the method of this invention can forward the subsequent command from the FIFO queue to the streaming target device. The method of this invention can repeat for any additional subsequent commands.

Method and system for multi-initiator support to streaming devices in a fibre channel network

A method and system are disclosed for encapsulating SCSI protocol for data transmission between two or more nodes across a packet-based network. The method of the present invention includes the steps of, at each node in the network, identifying all other available nodes on the network, and the remote devices attached to those nodes; representing one or more of the attached remote devices such that they are made available to the node's local hosts; encapsulating the I/O phases between one or more local hosts and one or more of the remote devices; and repeating the encapsulating step for subsequent I/Os between one or more hosts and one or more devices. The step of encapsulating I/O phases between a local host and a remote device can further comprise encapsulating task management functions, error recovery functions and normal I/O processing functions. Each node can be a Fibre Channel-to-SCSI router.

Encapsulation protocol for linking storage area networks over a packet-based network

The present invention provides a method and system for mapping addressing of SCSI devices between two SANs connected by a SAN extender over a packet-based network with use of a Fibre channel protocol over large distances. The present invention seamlessly interconnects graphically distinct SANs such that they operate as if they were local to one another by providing a means to generically and dynamically map SCSI device addresses between two SANs. The present invention provides a method and system for accessing a device from a host, wherein the host and device are in separate SANs interconnected by a transport layer, and wherein the interface between said transport layer and each of said SANs is a node. This method comprises, at each node, the steps of: mapping the device address into an intermediary device identifier, and mapping the intermediary device identifier into an address accessible by the host. Each node can be a Fibre channel-to-SCSI router, and the intermediary device identifier can comprise a node identifier and a generic device identifier. The transport layer can be a packet-based network.

System for mapping addresses of SCSI devices between plurality of SANs that can dynamically map SCSI device addresses across a SAN extender

A method and system for maintaining node states across network address changes in a Fibre Channel network includes steps and instructions for maintaining in each of a plurality of nodes, including target and initiator nodes, a triplet table including a triplet of data having the network address, node, name, and port name for each of other ones of said plurality of nodes with which the particular node communicates. The instructions permit sending from each of the plurality of nodes to the other ones of the nodes with which each of the plurality of nodes communicates the triplet of data following the resumption of communication occurring after a break in communication. The triplet of data identifies the event that the nodes are communicating with the same nodes both before and after said break in communication, the invention further includes instructions for continuing communications among said nodes without performing further recovery steps.

Method and system for managing I/O transmissions in a fibre channel network after a break in communication

A method and system for detecting and correcting error conditions on sequential devices in a private loop direct access fiber channel network respectively include the steps and instructions for determining the state of an exchange using an initiator and initiating an appropriate sequence level recovery using the initiator. The invention further includes determining whether a target response is overdue using a timer in conjunction with internal driver state information for indicating that packet information may have been lost. The invention further requests exchange and sequence state information from the target for determining the need for corrective action and takes the needed corrective action, such as resending sequence information, requesting that the target resend sequence information, or providing early indication to the ULP that an error has occurred.

Error detection and recovery for sequential access devices in a fibre

A method and system for maintaining a unique reserve command relationship between an initiator and a target device in a Fibre Channel network across network address changes after a break in communication. The present invention maintains triplet tables containing data triplets, comprised of the network address, the port name, and the node name, for each initiator and each target device. Following a break in network communication that results in the network address of an initiator and/or a target device changing, the method of the present invention updates the recorded network addresses for the initiators and the target devices, maintains any previously-existing unique reserve command relationships and continues with I/O transmission. Although the network address of an initiator may change, the node name and port name of the initiator will remain the same. By comparing the initiator port name and node name contained in a reserve table maintained in the target device to the node name and port name corresponding to the now updated network address for an initiator with which it was in a unique reserve command relationship prior to the break in communication, a target device can then simply update the network address for the initiator and no disruption in I/O traffic will result.

Method and system for maintaining reserve command relationships in a

A subsystem and method permits multiple protocols, to run simultaneously over a single adapter with minimal overhead. First interface structures are defined having similar component properties of a buffer address, buffer length, and cross-memory descriptor. The first structures are chainable, whereby data may be gathered from or scattered to multiple discontiguous data buffers for a single operation. A common structure completed by protocol heads has a pointer to the chain of first interface structures. For each control element, multiple buffer descriptors are provided to describe the various user buffers. The chain of multiple buffer structures contains pointers to the first structures employed to map buffers and obtain DMA addresses for an adapter. A layered architecture is provided for while allowing flexibility of utilizing current interfaces. Multiple protocols running over a single adapter driver are effected wherein the same protocol head code may run over a different adapter supporting the interface. Buffer, DMA management, and adapter handling are centralized in the adapter driver, allowing for more efficient allocation of system resources. The hardware interface is isolated in the adapter driver whereby changes for an integrated adapter may be isolated to one section of code.

Multiple protocol device interface subsystem and method

A technique for providing in-place cleaning of a sealing structure for a rotating shaft in a materials processing apparatus wherein, during a cleaning operation, cleaning liquid is supplied in the form of a plurality of jets thereof from a source thereof to a rotating member having a sealing element, such as an O-ring, positioned about the rotating shaft. The cleaning fluid produced a scrubbing action on critical surfaces of the sealing structure to clean them. A plurality of jets of a drying fluid, such as pressurized air, is then supplied from a source thereof to the critical surfaces to remove excess cleaning liquid and to thoroughly dry such surfaces. The cleaning and drying operation can be activated to take place automatically and in-place using a pre-programmed processor without having to disassemble the sealing structure.

Growing community of inventors

Pflugerville, TX, United States of America

Robert A Reynolds