Load balancing among network links using an efficient forwarding scheme

• US 20190327173A1
• Filed: 04/22/2018
• Published: 10/24/2019
• Est. Priority Date: 04/22/2018
• Status: Abandoned Application

• Current Assignee: Mellanox Technologies TLV Ltd. (NVIDIA Corporation)
• Original Assignee: Mellanox Technologies TLV Ltd. (NVIDIA Corporation)
• Inventors: Gafni, Barak, Levy, Gil
• Application Number:
  US15/959,234
• Publication Number:
  US 20190327173A1
• Time in Patent Office: Days
• CPC Class Codes: H04L 45/245 Link aggregation, e.g. trunking
• CPC Class Codes: H04L 45/7453 using hashing
• CPC Class Codes: H04L 47/125 by balancing the load, e.g. traffic engineering
• CPC Class Codes: H04L 47/2441 relying on flow classification, e.g. using integrated services [IntServ]
• CPC Class Codes: H04L 47/628 based on packet size, e.g. shortest packet first

Abstract

A network element includes multiple output ports and circuitry. The multiple output ports are configured to transmit packets over multiple respective network links of a communication network. The circuitry is configured to receive from the communication network, via one or more input ports of the network element, packets that are destined for transmission via the multiple output ports, to monitor multiple data-counts, each data-count corresponding to a respective output port, and is indicative of a respective data volume of the packets forwarded for transmission via the respective output port, to select for a given packet, based on the data-counts, an output port among the multiple output ports, and to forward the given packet for transmission via the selected output port.

Claims

1. A network element, comprising:

multiple output ports, configured to transmit packets over multiple respective network links of a communication network; and

circuitry, configured to;

receive from the communication network, via one or more input ports of the network element, packets that are destined for transmission via the multiple output ports, and forward the received packets for transmission to the communication network via the output ports;

store forwarded packets that are awaiting transmission in multiple queues corresponding to the multiple output ports;

monitor multiple data-counts, each data-count corresponding to a respective output port, and is indicative of a respective data volume of the packets that were forwarded to a respective queue for transmission via the respective output port; and

based on the data-counts, select for a given packet an output port among the multiple output ports, and forward the given packet for transmission via the selected output port.

2. The network element according to claim 1, wherein the circuitry is configured to select the output port in accordance with a criterion that aims to distribute traffic evenly among the multiple output ports.

3. The network element according to claim 1, wherein the circuitry is configured to check a respective amount of data forwarded, in a recent interval, to each of the multiple output ports, and to select the output port to which the amount of data forwarded in the recent interval is minimal among the multiple output ports.

4. The network element according to claim 1, wherein the circuitry is configured to select the output port by determining an amount of data to be transmitted via the selected output port before switching to a different output port.

5. The network element according to claim 1, wherein the circuitry is configured to assign to the multiple output ports multiple respective weights, and to distribute traffic among the multiple output ports based on the assigned weights.

6. The network element according to claim 1, wherein first and second output ports are coupled to respective first and second network links that support respective first and second different line-rates, and wherein the circuitry is configured to select the first output port or the second output port based at least on the first and second line-rates.

7. The network element according to claim 1, wherein the circuitry is configured to select the output port in accordance with a predefined cyclic order among the multiple output ports.

8. The network element according to claim 1, wherein the packets destined to the multiple output ports belong to a given traffic type, and wherein the circuitry is configured to select the output port based at least on the given traffic type.

9. The network element according to claim 1, wherein the circuitry is configured to select the output port by refraining from forwarding to a given output port packets of a priority level for which the given output port is paused or slowed down by flow control signaling imposed by a next-hop network element.

10. The network element according to claim 1, wherein the circuitry is configured to assign a packet-flow to a given output port, and to re-assign the packet-flow to a different output port in response to detecting that a time that elapsed since receiving a recent packet of the packet-flow exceeds a predefined period.

11. The network element according to claim 1, wherein the packets destined to the multiple output ports have different respective delivery priorities, and wherein the circuitry is configured to select the output port based at least on the delivery priority of a packet destined to the multiple output ports.

12. The network element according to claim 1, wherein the multiple output ports belong to a first load-balancing group and to a second load-balancing group, wherein at least one output port has a respective data-count that is shared by both the first and second load-balancing groups, and wherein the circuitry is configured to select an output port in the first load-balancing group based on the shared data-count while taking into consideration a port selection decision carried out previously for the second load-balancing group.

13. A method, comprising:

in a network element, transmitting packets via multiple output ports of the network element over multiple respective links of a communication network;

receiving from the communication network, via one or more input ports of the network element, packets that are destined for transmission via the multiple output ports, and forwarding the received packets for transmission to the communication network via the output ports;

storing forwarded packets that are awaiting transmission in multiple queues corresponding to the multiple output ports;

monitoring multiple data-counts, each data-count corresponding to a respective output port, and is indicative of a respective data volume of the packets that were forwarded to a respective queue for transmission via the respective output port; and

based on the data-counts, selecting for a given packet an output port among the multiple output ports, and forwarding the given packet for transmission via the selected output port.

14. The method according to claim 13, wherein selecting the output port comprises selecting the output port in accordance with a criterion that aims to distribute traffic evenly among the multiple output ports.

15. The method according to claim 13, wherein selecting the output port comprises checking a respective amount of data forwarded, in a recent interval, to each of the multiple output ports, and selecting an output port to which the amount of data forwarded in the recent interval is minimal among the multiple output ports.

16. The method according to claim 13, wherein selecting the output port comprises determining an amount of data to be transmitted via the selected output port before switching to a different output port.

17. The method according to claim 13, and comprising assigning to the multiple output ports multiple respective weights, and distributing traffic among the multiple output ports based on the assigned weights.

18. The method according to claim 13, wherein first and second output ports are coupled to respective first and second network links that support respective first and second different line-rates, and wherein selecting the output port comprises selecting the first output port or the second output port based at least on the first and second line-rates.

19. The method according to claim 13, wherein selecting the output port comprises selecting the output port in accordance with a predefined cyclic order among the multiple output ports.

20. The method according to claim 13, wherein the packets destined to the multiple output ports belong to a given traffic type, and wherein selecting the output port comprises selecting the output port based at least on the given traffic type.

21. The method according to claim 13, wherein selecting the output port comprises refraining from forwarding to a given output port packets of a priority level for which the given output port is paused or slowed down by flow control signaling imposed by a next-hop network element.

22. The method according to claim 13, and comprising assigning a packet-flow to a given output port, and re-assigning the packet-flow to a different output port in response to detecting that a time that elapsed since receiving a recent packet of the packet-flow exceeds a predefined period.

23. The method according to claim 13, wherein the packets destined to the multiple output ports have different respective delivery priorities, and wherein selecting the output port comprises selecting the output port based at least on the delivery priority of a packet destined to the multiple output ports.

24. The method according to claim 13, wherein the multiple output ports belong to a first load-balancing group and to a second load-balancing group, wherein at least one output port has a respective data-count that is shared by both the first and second load-balancing groups, and wherein selecting the output port comprises selecting an output port in the first load-balancing group based on the shared data-count while taking into consideration a port selection decision carried out previously for the second load-balancing group.