Intermediate node, an end node, and method for avoiding latency in a packet-switched network

Patent metadata

Field	Value
Publication number	US-9979652-B2
Application number	US-201414902627-A
Country	US
Kind code	B2
Filing date	Jul 29, 2014
Priority date	Aug 2, 2013
Publication date	May 22, 2018
Grant date	May 22, 2018

Abstract

The present invention is related to an intermediate node, an end node, and method for avoiding latency in a packet-switched network.
According to the invention, the maximum bandwidth per flow is determined locally and transferred from the sending node to the receiving node. At each node, the information is updated if a lower local value is computed. By adjusting the transmission rate based on the maximum bandwidth per flow receiving at the receiving node, the link capacity can be optimally utilized and congestion can be minimized.

First claim

The invention claimed is:

1. A method for avoiding latency in a packet-switched network, the network comprising a sending node, a receiving node, and a plurality of intermediate network nodes, the plurality of intermediate network nodes being mutually coupled by a plurality of links thereby forming paths for routing of packet flows between sending nodes and receiving nodes, the sending nodes including the sending node, the receiving nodes including the receiving node, the method comprising:
at an intermediate node among the plurality of intermediate network nodes, receiving one or more packets to be transported by a first link, the one or more packets belonging to the same packet flow having encoded therein information regarding a first maximum bandwidth per flow, and extracting the information;
at the intermediate node, determining the bandwidth utilized by the packet flows for the first link;
at the intermediate node, calculating a second maximum bandwidth per flow using the utilized bandwidth, a predefined target bandwidth, and a previously calculated maximum bandwidth per flow;
comparing the second maximum bandwidth per flow with the first maximum bandwidth per flow;
if the second bandwidth per flow is lower than the first maximum bandwidth per flow, communicating the second maximum bandwidth per flow as a communicated maximum bandwidth per flow to the sending node corresponding to the packet flows.

2. The method according to claim 1 , wherein communicating the
calculated maximum bandwidth per flow comprises encoding the calculated maximum bandwidth in one or more packets to be sent to a next node in the path(s) of the packet flows.

3. The method according to claim 1 , further comprising the sending node communicating an initial maximum transmission rate per flow as the maximum bandwidth per flow to a next node in the path;
the method further comprising the receiving node communicating the maximum bandwidth per flow encoded in the packet(s) it has received to the sending node, the sending node adjusting its maximum transmission rate in accordance with the maximum bandwidth it has received.

4. The method according to claim 1 , wherein the communicating the second maximum bandwidth per flow includes encoding corresponding information in a packet header and/or packet payload of the one or more packets to be sent over the first link.

5. The method according to claim 1 , wherein the communicating the second maximum bandwidth per flow comprises using a dedicated control protocol.

6. The method according to claim 1 , wherein the communicating the second maximum bandwidth per flow comprises encoding corresponding information using a plurality of packets to be sent over the first link by setting at least one predetermined bit in a packet header, the plurality of packets including the one or more packets, the number of packets included in the plurality of packets correlates with the maximum bandwidth per flow.

7. The method according to claim 1 , further comprising the intermediate node replacing and/or updating the encoded information regarding the maximum bandwidth per flow prior to forwarding the one or more packets, if the second maximum bandwidth per flow is lower than the first maximum bandwidth per flow.

8. The method according to claim 1 , wherein the calculating a maximum bandwidth per flow (Rpf) using the utilized bandwidth (BW), a predefined target bandwidth (U) and a previously calculated maximum bandwidth per flow (Rpf_p) is performed in accordance with the following equation:
Rpf=U×Rpf _ p /BW.

9. The method according to claim 8 , further comprising estimating a number of flows (#f) per link in accordance with the following equation:
# f =BW/ Rpf _ p.

10. The method according to claim 8 , further comprising:
estimating the number of highly responsive flows or barely responsive flows, or the bandwidths utilized by the highly responsive flows or the barely responsive flows, wherein
the calculating of the maximum bandwidth per flow of the highly responsive flows is performed based on a difference between the target bandwidth and a previously determined utilized bandwidth.

11. The method according to claim 10 , further comprising determining a number of highly responsive flows (#fa) per link using a difference between the previously calculated maximum bandwidth per flow (Rpf_p) and a third maximum bandwidth per flow (Rpf_pp) calculated previously to the previously calculated maximum bandwidth per flow, and a difference between a previously determined utilized bandwidth (BW_p) and a currently determined utilized bandwidth (BW), in accordance with the following equation:
# fa =(BW−BW_ p )/( Rpf _ p−Rpf _ pp )

wherein the maximum bandwidth per flow is calculated in accordance with the following equation:
Rpf =( U−BL )/# fa

wherein BL is the combined bandwidth of the barely responsive flows, which is computed in accordance with the following equation:
BL =BW−# fa×Rpf _ p.

12. The method according to claim 1 , wherein
the intermediate node is communicatively coupled to a plurality of different next nodes corresponding to different paths using respective links, each link being capable of accommodating a plurality of flows, wherein the method for avoiding latency is separately performed per link.

13. A method for operating a receiving node for a packet-switched network, the network comprising a sending node, the receiving node, and a plurality of intermediate nodes, the intermediate nodes being mutually coupled by a plurality of links thereby forming paths for routing of packet flows between sending nodes including the sending node and receiving nodes including the receiving node, the method comprising:
receiving one or more packets having encoded therein information regarding a maximum bandwidth per flow;
extracting the information per flow after the one or more packets are separated per flow; and
communicating the maximum bandwidth per flow to at least one sending node that has sent the one or more packets.

14. A method for operating a sending node for a packet-switched network, the network comprising the sending node, a receiving node, and a plurality of intermediate nodes, the intermediate nodes being mutually coupled by a plurality of links thereby forming paths for routing of packet flows between sending nodes including the sending node and receiving nodes including the receiving node, the method comprising:
receiving one or more packets corresponding to a packet flow having encoded therein information regarding a maximum bandwidth per flow;
extracting the information; and
pacing packets to be sent to a next node, which packets belong to the packet flow, wherein the pace at which packets are sent is set in accordance with the maximum bandwidth per flow.

Assignees

• Provenance Asset Group Llc

Inventors

• De Schepper Koen
• Tsang Ing-Jyh

Classifications

• H04L47/263 Primary

  Rate modification at the source after receiving feedback · CPC title

• H04L47/12 Primary

  Avoiding congestion; Recovering from congestion · CPC title

• H04L47/31

  by tagging of packets, e.g. using discard eligibility [DE] bits · CPC title

• H04L47/115

  using a dedicated packet · CPC title

• H04L43/0882

  Utilisation of link capacity · CPC title