Low latency flow control in data centers

A system for managing traffic between servers, the system comprises:

• first tier switches that are coupled to the servers;
• second tier switches that are coupled to the first tier switches and to third tier switches;
• and
• controllers;
• wherein each first tier switch comprises first queues;
• wherein each second tier switch comprises second queues;
• wherein the controllers are configured to control a traffic between the first tier switches and the second tier switches attributed to the traffic between the servers, (a) on, at least, a queue granularity;
• (b) while controlling some first queues to provide buffer extension to some second queues, and (c) while controlling some second queues to provide buffer extension to some first queues;
• and
• wherein at least one of the controllers is configured to control traffic between a certain first queue and certain second queues that provide buffer extension to the first queue by (a) monitoring a status of the certain queue, (2) detecting a potential congestion when a fullness level of the first queue exceeds a predefined threshold but is lower that a maximal fullness level, (c) sending a potential congestion alert, and (d) initiating a congestion prevention step related to the certain second queues.

A method for managing traffic between servers, the method comprises:

• controlling, by controllers of a system, traffic between servers, wherein the system comprises (i) first tier switches that are coupled to the servers;
• (ii) second tier switches that are coupled to the first tier switches and to third tier switches;
• wherein each first tier switch comprises first queues;
• wherein each second tier switch comprises second queues;
• wherein the controlling of the traffic comprises controlling a traffic between the first tier switches and the second tier switches that is attributed to the traffic between the servers, (a) on, at least, a queue granularity;
• (b) while controlling some first queues to provide buffer extension to some second queues, and (c) while controlling some second queues to provide buffer extension to some first queues;
• wherein the first tier switches and the second tier switches are positioned in multiple pods;
• and
• wherein the method further comprises comprising (a) preventing traffic between servers coupled to a same first tier switch from reaching any second tier switch, (b) preventing traffic between servers coupled to different first tier switches of a same pod from reaching any third tier switches, and (c) passing traffic between servers coupled to different pods through one or more third tier switches.

A method for managing traffic between servers, the method comprises:

• controlling, by controllers of a system, traffic between servers, wherein the system comprises (i) first tier switches that are coupled to the servers;
• (ii) second tier switches that are coupled to the first tier switches and to third tier switches;
• wherein each first tier switch comprises first queues;
• wherein each second tier switch comprises second queues;
• wherein the controlling of the traffic comprises controlling a traffic between the first tier switches and the second tier switches that is attributed to the traffic between the servers, (a) on, at least, a queue granularity;
• (b) while controlling some first queues to provide buffer extension to some second queues, and (c) while controlling some second queues to provide buffer extension to some first queues;
• and
• controlling, by at least one of the controllers, traffic between a certain first queue and certain second queues that provide buffer extension to the first queue by (a) monitoring a status of the certain queue, (2) detecting a potential congestion when a fullness level of the first queue exceeds a predefined threshold but is lower that a maximal fullness level, (c) sending a potential congestion alert, and (d) initiating a congestion prevention step related to the certain second queues.

A non-transitory computer readable medium that stores instructions for managing traffic between servers, the instructions, once executed by a system cause the system to execute the steps of:

• controlling, by controllers of the system, traffic between servers, wherein the system comprises (i) first tier switches that are coupled to the servers;
• (ii) second tier switches that are coupled to the first tier switches and to third tier switches;
• wherein each first tier switch comprises first queues;
• wherein each second tier switch comprises second queues;
• and
• wherein the controlling of the traffic comprises controlling a traffic between the first tier switches and the second tier switches that is attributed to the traffic between the servers, (a) on, at least, a queue granularity;
• (b) while controlling some first queues to provide buffer extension to some second queues, and (c) while controlling some second queues to provide buffer extension to some first queues;
• and
• controlling, by at least one of the controllers, traffic between a certain first queue and certain second queues that provide buffer extension to the first queue by (a) monitoring a status of the certain queue, (2) detecting a potential congestion when a fullness level of the first queue exceeds a predefined threshold but is lower that a maximal fullness level, (c) sending a potential congestion alert, and (d) initiating a congestion prevention step related to the certain second queues.