Characterizing data flow in a network based on a footprint measurement and taking action to avoid packet loss including buffer scaling or traffic shaping by adapting the footprint measurement

1. A method for characterising a data flow to be transferred over a network path of a network, said network path comprising at least one network device susceptible of network congestion, said method comprising:

determining a footprint measure of said data flow, said footprint measure being indicative of a possible difference between the total amount of incoming data and the total amount of outgoing data in said at least one network device over a time interval having a duration of one or more time units, whereby said time unit is so chosen that individual data units of the data flow are distinguishable at a byte level by said at least one network device;

classifying said data flow according to said footprint measure; and

taking counteraction to avoid packet loss in accordance with the classified data flow, wherein taking counteraction comprises scaling said buffering means of said at least one network device.

2. The method for characterising as in claim 1 , further comprising determining at least one further characteristic of said data flow.

3. The method for characterising as in claim 2 , wherein said at least one further characteristic is a parameter of the group of parameters comprising {delay, jitter, timing sensitivity, priority level}.

4. The method for characterising as in claim 1 , further comprising comparing the footprint measure with the capacity of buffering means for storing at least a part of said data flow, said buffering means being comprised in said at least one network device.

5. The method for characterising as in claim 1 , further comprising determining a quality-of-service level based on the classification of said at least one data flow.

6. The method for characterising as in claim 5 , wherein the quality-of-service level is selected from a set of predefined quality-of-service settings.

7. The method for characterising as in claim 1 , wherein said network is an Ethernet based IP network.

8. The method for characterising as in claim 7 , wherein said Ethernet based IP network is based on lossless Ethernet components.

9. The method for characterising as in claim 1 , wherein determining the footprint measure is performed off-line.

10. A device for characterising a data flow to be transferred over a network path of a network, said network path comprising at least one network device susceptible of network congestion, said device comprising:

processing means for determining a footprint measure of said data flow to be transferred, said footprint measure being indicative of a possible difference between the total amount of incoming data and the total amount of outgoing data in said at least one network device over a time interval having a duration of one or more time units, whereby said time unit is so chosen that individual data units of the data flow are distinguishable at a byte level by said at least one network device;

means for classifying said data flow according to said footprint measure; and

means for taking counteraction to avoid packet loss in accordance with the classified data flow by scaling said buffering means of said at least one network device.

11. The device recited in claim 10 , wherein the means for classifying said data flow is arranged for receiving at least a part of said footprint measure and comprises a classification engine for classifying said data flow based on said footprint measure.

12. The device as in claim 10 , further comprising means for determining at least one further characteristic of said data flow.

13. The device as in claim 10 , arranged for being connected to a database wherein at least one further characteristic of said data flow is stored.

14. A non-transitory computer readable medium having a module embodied therein for classifying a data flow to be transmitted over a network path of a network for data transfer, the module in the computer product including:

computer readable program code to receive at least a part of a footprint measure of the data flow, the footprint measure being indicative of a possible difference between the total amount of incoming data and the total amount of outgoing data in a network device of the network path over a time interval having a duration of one or more time units, the network device being susceptible of network congestion, whereby the time unit is so chosen that individual data units of the data flow are distinguishable at byte level by the network device;

computer readable program code to a classify the data flow based on the footprint measure of the data flow; and

computer readable program code to take counter action to avoid packet loss in accordance with the classified data flow by scaling the buffering means of the at least one network device.

15. The non-transitory computer readable medium of claim 14 wherein the module further includes:

computer readable program code to determine a quality-of-service level based on the classification of the at least one data flow;

wherein the quality-of-service level is selected from a set of predetermined quality-of-service settings.

16. A method for characterising a data flow to be transferred over a network path of a network, said network path comprising at least one network device susceptible of network congestion, said method comprising:

determining a footprint measure of said data flow, said footprint measure being indicative of a possible difference between the total amount of incoming data and the total amount of outgoing data in said at least one network device over a time interval having a duration of one or more time units, whereby said time unit is so chosen-that individual data units of the data flow are distinguishable at a byte level by said at least one network device;

classifying said data flow according to said footprint measure; and

taking counteraction to avoid packet loss in accordance with the classified data flow, wherein taking counteraction comprises shaping the traffic over said network path at the level of individual data units, wherein shaping the traffic comprises adapting said footprint measure in said at least one network device.

17. The method for characterising as in claim 16 , further comprising determining at least one further characteristic of said data flow.

18. The method for characterising as in claim 17 , wherein said at least one further characteristic is a parameter of the group of parameters comprising {delay, jitter, timing sensitivity, priority level}.

19. The method for characterising as in claim 16 , further comprising comparing the footprint measure with the capacity of buffering means for storing at least a part of said data flow, said buffering means being comprised in said at least one network device.