Exchange of processing metric information between nodes
Method and nodes are provided for propagating between the nodes information about processing capacity of peer nodes. Two types of signals are sent from propagating nodes towards their peer nodes. A propagating node sends first signal type at a constant rate and a second signal type at a variable rate. A ratio of the rates calculated at a peer node represents the processing capacity available a propagating node. Sequence numbers may be added to signals of the first and second type in order to assist peer nodes in detecting capacity issues.
1. A method of obtaining information about a processing capacity of a node, the method comprising the steps of:
receiving from the node a first signal at a first rate;
receiving from the node a second signal at a second rate;
calculating a ratio between the second rate and the first rate; and
determining the processing capacity of the node based on a comparison of the ratio with a preset value;
wherein:
each successive instance of the first signal comprises a sequence number calculated based on a sequence number of a previous instance of the first signal;
an instance of the second signal received immediately after a latest instance of the first signal comprises a sequence number based on a sequence number of the latest instance of the first signal; and
each successive instance of the second signal received before a next instance of the first signal comprises a sequence number based on a sequence number of a previous instance of the second signal.
2. The method of claim 1 , wherein:
the first rate is constant; and
the second rate is variable.
3. The method of claim 1 , wherein:
the preset value is equal to a ratio between the second and first rates when the node has a normal processing capacity.
4. The method of claim 1 , further comprising the steps of:
updating a routing table based on the step of determining the processing capacity of the node; and
refraining from sending messages towards the node when the processing capacity of the node is low.
5. The method of claim 1 , further comprising the steps of:
monitoring changes in the ratio; and
detecting a change of processing capacity of the node when the ratio changes.
6. The method of claim 5 , wherein:
a decrease of the second rate indicates a reduction of the processing capacity.
7. The method of claim 1 , wherein:
the sequence number in each successive instance of the first signal is equal to the sequence number of the previous instance of the first signal plus the preset value;
the sequence number in the instance of the second signal received immediately after the latest instance of the first signal is equal to the sequence number of the latest instance of the first signal plus one; and
the sequence number in each successive instance of the second signal received before the next instance of the first signal is equal to the sequence number of the previous instance of the second signal plus one.
8. The method of claim 1 , further comprising the steps of:
calculating an expected sequence number of a next instance of the second signal based on the sequence number of the latest instance of the first signal or based on the sequence number of the previous instance of the second signal; and
detecting a capacity problem by receiving the next instance of the second signal comprising a sequence number that is not equal to the expected sequence number.
9. The method of claim 1 , further comprising the steps of:
calculating an expected sequence number of a next instance of the second signal based on the sequence number of the latest instance of the first signal or based on the sequence number of the previous instance of the second signal; and
wherein calculating the ratio between the second rate and the first rate comprises ignoring a given instance of the second signal comprising a sequence number that does not match the expected sequence number.
10. A method of propagating information about a processing capacity of a node, the method comprising the steps of:
determining an offered capacity value, the offered capacity value being equal to or less than a preset value indicative of a normal processing capacity of the node;
sending from the node a first signal periodically at a base rate; and
sending from the node a second signal periodically at a second rate indicative of the offered capacity value;
wherein the second rate is equal to the base rate multiplied by the offered capacity value.
11. The method of claim 10 , wherein:
the node comprises a plurality of interfaces;
each periodic first signal is sent on all interfaces; and
each periodic second signal is sent on all interfaces.
12. The method of claim 10 , wherein:
the offered capacity value is determined periodically.
13. The method of claim 10 , wherein:
the offered capacity value is set according to an actual processing capacity of the node.
14. The method of claim 10 , wherein:
the offered capacity value is set according to a preset percentage of the normal processing capacity of the node.
15. A node for propagating information about its processing capacity, comprising:
a processor;
a load indicator adapted to indicate a load level of the processor;
one or more interfaces adapted to send signals towards peer nodes; and
a control logic adapted to:
read the load level of the processor;
request the one or more interfaces to send a first signal at a first rate;
calculate a second rate based on a comparison of the load level and of a maximum amount of processing capacity of the node; and
request the one or more interfaces to send a second signal at the second rate;
wherein calculating a second rate based on a comparison of the load level and of the maximum amount of processing comprises setting the second rate to a maximum value when the load level is low.
16. The node of claim 15 , further comprising:
a clock adapted to initiate a sequence of the control logic at the first rate.
17. The node of claim 15 , wherein:
each successive instance of the first signal comprises a sequence number calculated based on a sequence number of a previous instance of the first signal;
an instance of the second signal sent immediately after a latest instance of the first signal comprises a sequence number based on a sequence number of the latest instance of the first signal; and
each successive instance of the second signal sent before a next instance of the first signal comprises a sequence number based on a sequence number of a previous instance of the second signal.
18. A node for obtaining information about a processing capacity of a peer node, comprising:
one or more interfaces adapted to receive signals from the peer node; and
a control logic adapted to:
receive from the one or more interfaces a first signal multiple times;
determine a first rate of the first signal;
receive from the one or more interfaces a second signal multiple times;
determine a second rate of the second signal;
calculate a ratio between the second and first rates; and
store in a peer table a processing capacity of the peer node based on the ratio;
and further comprising a clock; and
wherein the control logic uses timing information from the clock to determine the first and second rates, and stores the first and second rates in the peer table.
19. The node of claim 18 , wherein:
the first and second signals comprise sequence numbers;
the control logic is further adapted to compute an expected sequence number of a next second signal based on a sequence number of a latest first signal or based on a sequence number of a previous second signal; and
the control logic is further adapted to store the expected sequence number in the peer table.
20. The node of claim 19 , wherein:
the control logic is further adapted to detect a capacity problem by receiving a next second signal comprising a sequence number that is not equal to the expected sequence number.
21. The node of claim 19 , wherein:
calculating a ratio between the second and first rates comprises ignoring a given second signal comprising a sequence number that does not match the expected sequence number.