Method of optimizing a network
View Patent ↗An optimized network ( 300 ), includes providing a switch card topology ( 350 ) having a plurality of switching elements ( 305 ), where the plurality of switching elements are arranged to form a switch configuration ( 303 ). N number of payload interfaces ( 307 ) are coupled to the switch configuration, where each of the N number of payload interfaces is coupled to interface with one of a plurality of payload slots ( 308 ). A set of N payload module configurations ( 402, 502 ) is characterized by a sequential addition ( 320 ) of a payload module ( 304 ) into each of the plurality of payload slots, where the sequential addition of the payload module couples the payload module to the network. N number of payload interfaces are coupled to the switch configuration such that a latency function ( 616 ) is minimized for the switch configuration and the set of N payload module configurations.
1. A method of optimizing a network, comprising:
providing a switch card topology having a plurality of switching elements, wherein the plurality of switching elements are arranged to form a switch configuration;
providing N number of payload interfaces coupled to the switch configuration, wherein each of the N number of payload interfaces is coupled to interface with one of a plurality of payload slots;
providing a set of N payload module configurations, wherein the set of N payload module configurations is characterized by a sequential addition of a payload module into each of the plurality of payload slots, wherein the sequential addition of the payload module couples the payload module to the network; and
coupling the N number of payload interfaces to the switch configuration such that a latency function is minimized for the switch configuration and the set of N payload module configurations.
2. The method of claim 1 , wherein the latency function is a root-mean squared latency function.
3. The method of claim 1 , wherein the latency function is an arithmetic mean latency function.
4. The method of claim 1 , wherein the latency function is a normalized root-mean squared latency function.
5. The method of claim 1 , wherein the latency function is a normalized arithmetic mean latency function.
6. The method of claim 1 , wherein the sequential addition comprises a first end to a second end sequential addition.
7. The method of claim 1 , wherein the sequential addition comprises a second end to a first end sequential addition.
8. The method of claim 1 , wherein the plurality of switching elements comprises M number of switching elements coupled to the N number of payload interfaces, wherein each of the M number of switching elements has a plurality of ports, and wherein the sequential addition comprises populating all of the plurality of ports on one of the M number of switching elements, one of the M number of switching elements at a time.
9. A method of optimizing a network, comprising:
providing a switch card having a plurality of switching elements, wherein the plurality of switching elements are arranged to form a switch configuration;
providing N number of payload interfaces coupled to the switch configuration, wherein each of the N number of payload interfaces is coupled to interface with one of a plurality of payload slots;
providing a set of N payload module configurations, wherein the set of N payload module configurations is characterized by a sequential addition of a payload module into each of the plurality of payload slots, wherein the sequential addition of the payload module couples the payload module to the network; and
remapping the N number of payload interfaces to the switch configuration such that a latency function is minimized, wherein the latency function is a function of the switch configuration and the set of N payload module configurations.
10. The method of claim 9 , wherein the latency function is a root-mean squared latency function.
11. The method of claim 9 , wherein the latency function is an arithmetic mean latency function.
12. The method of claim 9 , wherein the latency function is a normalized root-mean squared latency function.
13. The method of claim 9 , wherein the latency function is a normalized arithmetic mean latency function.
14. The method of claim 9 , wherein the sequential addition comprises a first end to a second end sequential addition.
15. The method of claim 9 , wherein the sequential addition comprises a second end to a first end sequential addition.
16. The method of claim 9 , wherein the plurality of switching elements comprises M number of switching elements coupled to the N number of payload interfaces, wherein each of the M number of switching elements has a plurality of ports, and wherein the sequential addition comprises populating all of the plurality of ports on one of the M number of switching elements, one of the M number of switching elements at a time.
17. A method of optimizing a network, comprising:
providing a switch card topology having a plurality of switching elements, wherein the plurality of switching elements are arranged to form a switch configuration;
providing N number of payload interfaces coupled to the switch configuration, wherein each of the N number of payload interfaces is coupled to interface with one of a plurality of payload slots;
providing a set of N payload module configurations, wherein the set of N payload module configurations is characterized by a sequential addition of a payload module into each of the plurality of payload slots, wherein the sequential addition of the payload module couples the payload module to the network;
calculating a plurality of remapping functions for the set of N payload module configurations; and
selecting one of the plurality of remapping functions to minimize a latency function, wherein the latency function is a function of the switch configuration and the set of N payload module configurations.
18. The method of claim 17 , coupling the N number of payload interfaces to the switch configuration according to the one of the plurality of remapping functions selected to minimize the latency function.
19. The method of claim 17 , wherein the latency function is a root-mean squared latency function.
20. The method of claim 17 , wherein the latency function is an arithmetic mean latency function.
21. The method of claim 17 , wherein the latency function is a normalized root-mean squared latency function.
22. The method of claim 17 , wherein the latency function is a normalized arithmetic mean latency function.
23. The method of claim 17 , wherein the sequential addition comprises a first end to a second end sequential addition.
24. The method of claim 17 , wherein the sequential addition comprises a second end to a first end sequential addition.
25. The method of claim 17 , wherein the plurality of switching elements comprises M number of switching elements coupled to the N number of payload interfaces, wherein each of the M number of switching elements has a plurality of ports, and wherein the sequential addition comprises populating all of the plurality of ports on one of the M number of switching elements, one of the M number of switching elements at a time.
26. A method of optimizing a network, comprising:
providing a switch card having a plurality of switching elements, wherein the plurality of switching elements are arranged to form a switch configuration;
providing N number of payload interfaces coupled to the switch configuration, wherein each of the N number of payload interfaces is coupled to interface with one of a plurality of payload slots;
providing a set of N payload module configurations, wherein the set of N payload module configurations is characterized by a sequential addition of a payload module into each of the plurality of payload slots, wherein the sequential addition of the payload module couples the payload module to the network;
calculating a plurality of remapping functions for the set of N payload module configurations; and
minimizing an all-to-all transfer time in the network by selecting one of the plurality of remapping functions to minimize a latency function, wherein the latency function is a function of the switch configuration and the set of N payload module configurations.
27. The method of claim 26 , coupling the N number of payload interfaces to the switch configuration according to the one of the plurality of remapping functions selected to minimize the latency function.
28. A method of optimizing a switch card, comprising:
providing a plurality of switching elements, wherein the plurality of switching elements are arranged to form a switch configuration on the switch card;
providing N number of payload interfaces coupled to the switch configuration, wherein each of the N number of payload interfaces is coupled to interface with one of a plurality of payload slots;
providing a set of N payload module configurations, wherein the set of N payload module configurations is characterized by a sequential addition of a payload module into each of the plurality of payload slots, wherein the sequential addition of the payload module couples the payload module to the network; and
coupling the N number of payload interfaces to the switch configuration such that a latency function is minimized for the switch configuration and the set of N payload module configurations.
29. The method of claim 28 , wherein the latency function is a root-mean squared latency function.
30. The method of claim 28 , wherein the latency function is an arithmetic mean latency function.
31. The method of claim 28 , wherein the latency function is a normalized root-mean squared latency function.
32. The method of claim 28 , wherein the latency function is a normalized arithmetic mean latency function.
33. The method of claim 28 , wherein the sequential addition comprises a first end to a second end sequential addition.
34. The method of claim 28 , wherein the sequential addition comprises a second end to a first end sequential addition.
35. The method of claim 28 , wherein the plurality of switching elements comprises M number of switching elements coupled to the N number of payload interfaces, wherein each of the M number of switching elements has a plurality of ports, and wherein the sequential addition comprises populating all of the plurality of ports on one of the M number of switching elements, one of the M number of switching elements at a time.
36. A method of optimizing a network, comprising:
providing a switch card having a plurality of switching elements, wherein the plurality of switching elements are arranged to form a switch configuration;
providing N number of payload interfaces coupled to the switch configuration, wherein each of the N number of payload interfaces is coupled to interface with one of a plurality of payload slots;
providing a set of N payload module configurations, wherein the set of N payload module configurations is characterized by a sequential addition of a payload module into each of the plurality of payload slots, wherein the sequential addition of the payload module couples the payload module to the network; and
reordering a sequence of adding the payload module into each of the payload slots such that a latency function is minimized, wherein the latency function is a function of the switch configuration and the set of N payload module configurations.