System and method to provide path records derived from ARP responses and peer-to-peer negotiation based on homogeneous fabric attribute in a high performance computing environment
Systems and methods for InfiniBand fabric optimizations to minimize SA access and startup failover times. Based on the combination of path record parameters defined with values from the Homogeneous Fabric/Subnet flag (or configuration parameter), as well as the GID and SLID information defined by IPoIB ARP requests and responses, there is no additional need for SA requests to obtain path records or path related information in general.
1. A system for providing a path record in a high performance computing environment, comprising:
one or more microprocessors;
a first subnet, the first subnet comprising
a plurality of switches, the plurality of switches comprising at least a leaf switch, wherein each of the plurality of switches comprise at least one switch port of a plurality of switch ports,
a plurality of host channel adapters, wherein each of the host channel adapters comprise at least one host channel adapter port of a plurality of host channel adapters ports, and wherein the plurality of host channel adapters are interconnected via the plurality of switches,
a plurality of hosts, and
a subnet manager, the subnet manager running on one of the plurality of switches and the plurality of host channel adapters;
wherein the subnet manager determines that a set of the plurality of host channel adapter ports and a set of the plurality of switch ports support a same set of capabilities;
wherein upon the subnet manager determining that a set of the plurality of host channel adapter ports and a set of the plurality of switch ports support the same set of capabilities, the subnet manager configures a subnet management agent (SMA) flag, the SMA flag indicating that a condition can be set for each of the plurality of host channel adapter ports and the plurality of switch ports;
wherein a packet is received from a remote port, the received packet comprising a source LID (local identifier) and a source GID (global identifier) at a port on a switch in the first subnet; and
wherein the source LID and the source GID are combined with the SMA flag to determine a complete path record between a port at which the packet is received and the remote port.
2. The system of claim 1 , wherein the same set of capabilities comprises at least one of maximum transmission unit, and supported data rate.
3. The system of claim 2 , wherein the same set of capabilities comprises each of maximum transmission unit, and supported data rate.
4. The system of claim 2 , wherein the determined set of the plurality of host channel adapter ports that support the same set of capabilities comprises each of the plurality of host channel adapter ports.
5. The system of claim 1 , wherein the remote port is located in a second subnet, the second subnet being connected to the first subnet via at least one router port; and
wherein the source LID is generated by the at least one router port.
6. The system of claim 5 , wherein the subnet manager sets a second SMA flag indicating that any source LID generated by the router port can be used to reach an associated port in a remote subnet.
7. The system of claim 6 , wherein the generated source LID is combined with the source GID, the SMA flag and the second SMA flag to determine the complete path record between the port at which the packet is received and the remote port.
8. A method for providing a path record in a high performance computing environment, comprising:
providing, at one or more microprocessors, a first subnet, the first subnet comprising
a plurality of switches, the plurality of switches comprising at least a leaf switch, wherein each of the plurality of switches comprise at least one switch port of a plurality of switch ports,
a plurality of host channel adapters, wherein each of the host channel adapters comprise at least one host channel adapter port of a plurality of host channel adapters ports, and wherein the plurality of host channel adapters are interconnected via the one or more plurality of switches,
a plurality of hosts, and
a subnet manager, the subnet manager running on one of the plurality of switches and the plurality of host channel adapters;
determining, by the subnet manager, that a set of the plurality of host channel adapter ports and a set of the plurality of switch ports support a same set of capabilities;
upon the subnet manager determining that a set of the plurality of host channel adapter ports and a set of the plurality of switch ports support the same set of capabilities, configuring, by the subnet manager can, a subnet management agent (SMA) flag, the SMA flag indicating that a condition can be set for each of the plurality of host channel adapter ports and the plurality of switch ports;
receiving a packet can be from a remote port, the received packet comprising a source LID and a source GID at a port on a switch in the first subnet; and
combining the source LID and the source GID with the SMA flag to determine a complete path record between a port at which the packet is received and the remote port.
9. The method of claim 8 , wherein the same set of capabilities comprises at least one of maximum transmission unit, and supported data rate.
10. The method of claim 9 , wherein the same set of capabilities comprises each of maximum transmission unit, and supported data rate.
11. The method of claim 9 , wherein the determined set of the plurality of host channel adapter ports that support the same set of capabilities comprises each of the plurality of host channel adapter ports.
12. The method of claim 8 , wherein the remote port is located in a second subnet, the second subnet being connected to the first subnet via at least one router port; and
wherein the source LID is generated by the at least one router port.
13. The method of claim 12 , wherein the subnet manager sets a second SMA flag indicating that any source LID generated by the router port can be used to reach an associated port in a remote subnet.
14. The method of claim 13 , wherein the generated source LID is combined with the source GID, the SMA flag and the second SMA flag to determine the complete path record between the port at which the packet is received and the remote port.
15. A non-transitory computer readable storage medium, including instructions stored thereon for providing a path record in a high performance computing environment, which when read and executed by one or more computers cause the one or more computers to perform the steps comprising:
providing, at one or more microprocessors, a first subnet, the first subnet comprising
a plurality of switches, the plurality of switches comprising at least a leaf switch, wherein each of the plurality of switches comprise at least one switch port of a plurality of switch ports,
a plurality of host channel adapters, wherein each of the host channel adapters comprise at least one host channel adapter port of a plurality of host channel adapters ports, and wherein the plurality of host channel adapters are interconnected via the one or more plurality of switches,
a plurality of hosts, and
a subnet manager, the subnet manager running on one of the plurality of switches and the plurality of host channel adapters;
determining, by the subnet manager, that a set of the plurality of host channel adapter ports and a set of the plurality of switch ports support a same set of capabilities;
upon the subnet manager determining that a set of the plurality of host channel adapter ports and a set of the plurality of switch ports support the same set of capabilities, configuring, by the subnet manager can, a subnet management agent (SMA) flag, the SMA flag indicating that a condition can be set for each of the plurality of host channel adapter ports and the plurality of switch ports;
receiving a packet can be from a remote port, the received packet comprising a source LID and a source GID at a port on a switch in the first subnet; and
combining the source LID and the source GID with the SMA flag to determine a complete path record between a port at which the packet is received and the remote port.
16. The non-transitory computer readable storage medium of claim 15 , wherein the same set of capabilities comprises at least one of maximum transmission unit, and supported data rate.
17. The non-transitory computer readable storage medium of claim 16 , wherein the same set of capabilities comprises each of maximum transmission unit, and supported data rate.
18. The non-transitory computer readable storage medium of claim 16 , wherein the determined set of the plurality of host channel adapter ports that support the same set of capabilities comprises each of the plurality of host channel adapter ports.
19. The non-transitory computer readable storage medium of claim 18 , wherein the remote port is located in a second subnet, the second subnet being connected to the first subnet via at least one router port; and
wherein the source LID is generated by the at least one router port.
20. The non-transitory computer readable storage medium of claim 19 , wherein the subnet manager sets a second SMA flag indicating that any source LID generated by the router port can be used to reach an associated port in a remote subnet; and
wherein the generated source LID is combined with the source GID, the SMA flag and the second SMA flag to determine the complete path record between the port at which the packet is received and the remote port.