×

Fibre channel arbitrated loop bufferless switch circuitry to increase bandwidth without significant increase in cost

  • US 7,430,171 B2
  • Filed: 01/15/2003
  • Issued: 09/30/2008
  • Est. Priority Date: 11/19/1998
  • Status: Expired due to Fees
First Claim
Patent Images

1. A process for carrying out dual simplex communications in a fibre channel arbitrated loop switch coupling a plurality of fibre channel arbitrated loop nets, comprising:

  • receiving and latching an open primitive and at least the source node address thereof identifying the address of said source node and receiving any following buffer credit ready primitives and either storing said ready primitives or counting a number thereof at a source port from a source node on a local fibre channel arbitrated loop net coupled to said source port, said open primitive directed to a remote destination node on the fibre channel arbitrated loop net coupled to a remote destination port;

    establishing a front channel connection through a backplane between said source port and said destination port;

    converting said open primitive to half duplex and transmitting said half duplex open primitive to said destination node over said front channel;

    receiving any ready primitives from said destination node and transmitting them to said source node via a back channel;

    receiving a connection request message from a third port coupled to a third node and granting said connection request by a connection grant message to said third port coupled to said third node, said connection grant message naming a backplane channel to use for a back channel connection between said source node and said third node;

    establishing a backchannel connection between said source port and said third port through said backplane channel named in said connection grant message;

    transmitting an open primitive directed to said source node from said third node to said source port;

    comparing a destination node address in said open primitive from said third port to a node address of the open primitive originally received by said source port, and, if the addresses match, quashing the open primitive from said third port and either transmitting one stored ready primitive received from said source node to said third node or transmitting a message which causes said third port to generate at least one ready primitive and transmit it to said third node;

    transmitting one frame of data from said third node to said source node via said back channel connection; and

    continuing to intercept ready primitives transmitted by said source node and send them to said third node and to receive frames of data transmitted in response to receipt at said third node of said ready primitives and transmit said frames to said source node via said back channel connection and mix said frames of data on said back channel connection with ready primitives received from said destination node.

View all claims
  • 3 Assignments
Timeline View
Assignment View
    ×
    ×