IP Library Granted Patent US 7,739,405
Granted Patent B2
US 7,739,405 · App. 12/274,158 · Granted Jun 15, 2010

Multicast group management in infiniband

View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 7,739,405
App. No.
12/274,158
Granted
Jun 15, 2010
Kind
B2
Abstract

A mechanism for managing multicast groups with send-without-receive (SWR) joiners without the use of traps on creation and deletion of groups is provided. Group information is maintained continuously while the SWR member exists. When an SWR join is attempted and the group does not already exist, the group information (MLID) is marked as used and the first switch to which the SWR packets are sent is routed to discard all packets sent to the group. When receiving members join the group, the routing is updated so that the SWR member begins sending to the receiving members. When the last receiving member leaves the group, the first switch is again routed to discard the packets.

Claims (28)

1. An apparatus for managing multicast groups in an InfiniBand system area network, the apparatus comprising:

InfiniBand end nodes in the InfiniBand system area network being final destinations for packets, said InfiniBand end nodes comprising a computer recordable-type medium;

said InfiniBand end nodes do not include switches or routers, and packets are not routed through said InfiniBand end nodes;

a Subnet Administration in a first InfiniBand end node that receives a join request from a second InfiniBand end node for joining a multicast group, wherein the second InfiniBand end node is connected to a first switch and wherein the join request is a send-without-receive request that specifies that the second InfiniBand end node is to be a send-without-receive member of the multicast group that can send packets but will not receive packets, and wherein the first InfiniBand end node is included within the InfiniBand system area network, and further wherein switches in the InfiniBand system area network route packets from the second InfiniBand end node but do not route any packets to the second InfiniBand end node;

the Subnet Administation determines whether the multicast group exists;

the Subnet Administration in the first InfiniBand end node creates the multicast group, wherein the multicast group is not created by a switch, wherein when the multicast group is created, the second InfiniBand end node is the only member of the multicast group and the multicast group includes no members that receive packets; and

the Subnet Administration routes the first switch to discard all packets for the multicast group if the multicast group does not exist.

2. The apparatus of claim 1 , wherein the Subnet Administration for creating the multicast group includes the Subnet Administration for assigning an InfiniBand multicast local identifier (MLID) to the multicast group, and further wherein the multicast group is not created by a switch.

3. The apparatus of claim 1 , wherein the Subnet Administration for routing includes the Subnet Administration for inserting an entry for the multicast group in a multicast routing data structure in the first switch.

4. The apparatus of claim 1 , further comprising:

the Subnet Administration, responsive to a join request from a receiving end node, for causing an update of at least one multicast routing table for at least one switch in the system area network to route packets for the multicast group to the receiving end node.

5. An apparatus for managing multicast groups in an InfiniBand system area network, the apparatus comprising:

InfiniBand end nodes in the InfiniBand system area network being final destinations for packets, said InfiniBand end nodes comprising a computer recordable-type medium;

said InfiniBand end nodes do not include switches or routers, and packets are not routed through said InfiniBand end nodes;

a Subnet Administration in a first InfiniBand end node that receives a leave request from a second InfiniBand end node for leaving a multicast group, wherein the multicast group has a first member at a third end node connected to a first switch, and wherein the multicast group is identified using an InfiniBand multicast local identifier (MLID), and wherein the first InfiniBand end node is included within the InfiniBand system area network, and further wherein the first member is a send-without-receive member that can send packets but will not receive packets, and still further wherein switches in the InfiniBand system area network route packets from the first member but do not route any packets to the first member;

the Subnet Administation determines whether the first member is the only remaining member in the multicast group; and

the Subnet Administration in the first InfiniBand end node routes the first switch to discard all packets from the first member if the first member is the only remaining member in the multicast group.

6. The apparatus of claim 5 , wherein the multicast routing data structure is indexed by the InfiniBand multicast local identifier (MLID).

7. The apparatus of claim 5 , further comprising:

the Subnet Administration, responsive to a join request from a receiving end node, for causing an update of at least one multicast routing table for at least one switch in the system area network to route packets for the multicast group to the receiving end node

8. A computer program product, stored in a computer recordable-type medium, for managing multicast groups in an InfiniBand system area network, the computer program product comprising:

instructions for receiving, by a Subnet Administration in a first InfiniBand end node, a join request from a second InfiniBand end node for joining a multicast group, wherein the second InfiniBand end node is connected to a first switch and wherein the join request is a send-without-receive request that specifies that the second InfiniBand end node is to be a send-without-receive member of the multicast group that can send packets but will not receive packets, and wherein the first InfiniBand end node is included within the InfiniBand system area network, and further wherein switches in the InfiniBand system area network route packets from the second InfiniBand end node but do not route any packets to the second InfiniBand end node, InfiniBand end nodes in the InfiniBand system area network being final destinations for packets, said InfiniBand end nodes do not include switches or routers, and packets are not routed through said InfiniBand end nodes;

instructions for determining whether the multicast group exists; and

instructions for creating, by the Subnet Administration in the first InfiniBand end node, the multicast group and routing the first switch to discard all packets for the multicast group if the multicast group does not exist, wherein when the multicast group is created, the second InfiniBand end node is the only member of the multicast group and the multicast group includes no member that will receive packets.

9. A computer program product, stored in a computer recordable-type medium, for managing multicast groups in an InfiniBand system area network, the computer program product comprising:

instructions for receiving, by a Subnet Administration in a first InfiniBand end node, a leave request from a second InfiniBand end node for leaving a multicast group, wherein the multicast group has a first member at a third end node connected to a first switch, and wherein the multicast group is identified using an InfiniBand multicast local identifier (MLID), and wherein the first InfiniBand end node is included within the InfiniBand system area network, and further wherein the first member is a send-without-receive member that can send packets but will not receive packets, and still further wherein switches in the InfiniBand system area network route packets from the first member but do not route any packets to the first member, InfiniBand end nodes in the InfiniBand system area network being final destinations for packets, said InfiniBand end nodes do not include switches or routers, and packets are not routed through said InfiniBand end nodes;

instructions for determining whether the first member is the only remaining member in the multicast group; and

instructions for routing, by the Subnet Administration in the first InfiniBand end node, the first switch to discard all packets from the first member if the first member is the only remaining member in the multicast group.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Nov 10, 2014
From: INTERNATIONAL BUSINESS MACHINES CORPORATION
To: LENOVO INTERNATIONAL LIMITED
Reel/Frame 034194/0280 →