Switch fabric end-to-end congestion avoidance mechanism
Aspects of a switch fabric end-to-end congestion avoidance mechanism are presented. Aspects of a system for end-to-end congestion avoidance in a switch fabric may include at least one circuit that enables reception of a congestion notification message that specifies a traffic flow identifier. The circuitry may enable increase or decrease of a current rate for transmission of data link layer (DLL) protocol data units (PDU) associated with the specified traffic flow identifier as a response to the reception of the congestion notification message.
1. A method for congestion avoidance, the method comprising:
receiving, using a network device, a congestion notification message having a request to modify a data transmission rate; and
upon receiving the congestion notification message:
resetting a congestion notification timer,
determining whether a data transmission rate modification wait time has elapsed,
modifying the data transmission rate in response to determining that the data transmission rate modification wait time has elapsed, and
maintaining the data transmission rate in response to determining that the data transmission rate modification wait time has not elapsed.
2. The method of claim 1 , wherein the data transmission rate is a rate for transmitting data link layer (DLL) protocol data units (PDUs).
3. The method of claim 1 , further comprising:
determining whether the request to modify the data transmission rate is a data transmission rate decrease request or a data transmission rate increase request.
4. The method of claim 1 , wherein the data transmission rate modification wait time specifies a minimum amount of time to wait before decreasing the data transmission rate after a last decrease in the data transmission rate, and wherein modifying the data transmission rate comprises decreasing the data transmission rate.
5. The method of claim 1 , further comprising:
determining whether the data transmission rate is less than a maximum data transmission rate if the request to modify the data transmission rate is a data transmission rate increase request; and
in response to determining that the data transmission rate is not less than the maximum data transmission rate, ignoring the congestion notification message.
6. The method of claim 1 , wherein the data transmission rate modification wait time specifies a minimum amount of time to wait before increasing the data transmission rate after a last increase in the data transmission rate, and wherein the method further comprises:
in response to determining that the data transmission rate is less than a maximum data transmission rate and that the data transmission rate modification wait time has elapsed, increasing the data transmission rate.
7. The method of claim 1 , wherein the congestion notification message further comprises a fabric area network (FAN) flow indicator.
8. The method of claim 7 , further comprising:
selecting a traffic management queue (TMQ) based on the FAN flow indicator; and
modifying the data transmission rate for data link layer (DLL) protocol data units (PDUs) associated with the TMQ.
9. The method of claim 1 , further comprising:
generating the congestion notification message by a peer network device based on one or more congestion indications that are triggered by one or more switching devices external to both the network device and the peer network device.
10. A network device, comprising:
a processor;
a memory; and
a network interface controller (NIC) configured to:
receive a congestion notification message having a request to modify a data transmission rate, and
upon receiving the congestion notification message:
reset a congestion notification timer,
determine whether a data transmission rate modification wait time has elapsed,
modify the data transmission rate in response to determining that the data transmission rate modification wait time has elapsed, and
maintain the data transmission rate in response to determining that the data transmission rate modification wait time has not elapsed.
11. The network device of claim 10 , wherein the data transmission rate is a rate for transmitting data link layer (DLL) protocol data units (PDUs).
12. The network device of claim 10 , wherein the NIC is further configured to:
determine whether the request to modify the data transmission rate is a data transmission rate decrease request or a data transmission rate increase request.
13. The network device of claim 10 , wherein the data transmission rate modification wait time specifies a minimum amount of time to wait before decreasing the data transmission rate after a last decrease in the data transmission rate, and wherein the NIC is further configured to:
decrease the data transmission rate in response to determining that the data transmission rate modification wait time has elapsed.
14. The network device of claim 10 , wherein the NIC is further configured to:
determine whether the data transmission rate is less than a maximum data transmission rate; and
in response to determining that the data transmission rate is not less than the maximum data transmission rate, ignore the congestion notification message.
15. The network device of claim 10 , wherein the data transmission rate modification wait time specifies a minimum amount of time to wait before increasing the data transmission rate after a last increase in the data transmission rate, and wherein the NIC is further configured to:
in response to determining that the data transmission rate is less than the maximum data transmission rate and that the data transmission rate modification wait time has elapsed, increase the data transmission rate.
16. The network device of claim 10 , wherein the congestion notification message further comprises a fabric area network (FAN) flow indicator.
17. The network device of claim 16 , wherein the NIC is further configured to:
select a traffic management queue (TMQ) based on the FAN flow indicator; and
modify the data transmission rate for data link layer (DLL) protocol data units (PDUs) associated with the TMQ.
18. The network device of claim 10 , wherein the congestion notification message is generated by a peer network device based on one or more congestion indications that are triggered by one or more switching devices external to both the network device and the peer network device.
19. The network device of claim 10 , wherein the NIC is further configured to:
receive a data link layer (DLL) protocol data unit (PDU);
determine whether a congestion indication is contained within the DLL PDU; and
transmit a second congestion notification message in response to determining that the congestion indication is contained within the DLL PDU.
20. A network device, comprising:
a processor;
a memory; and
a network interface controller (NIC) configured to:
receive a congestion notification message having a request to modify a data transmission rate, and
upon receiving the congestion notification message:
reset a congestion notification timer,
determine whether a data transmission rate modification wait time has elapsed,
maintain the data transmission rate in response to determining that the data transmission rate modification wait time has not elapsed upon receiving the congestion notification message,
increase the data transmission rate in response to determining that:
the data transmission rate modification wait time has elapsed,
the request to modify the data transmission rate is a request to increase the data transmission rate, and
the data transmission rate is less than a maximum data transmission rate, and
decrease the data transmission rate in response to determining that:
the data transmission rate modification wait time has elapsed, and
the request to modify the data transmission rate is a request to decrease the data transmission rate.