Firmware download with adaptive lost packet recovery
View Patent ↗Disclosed are methodologies for implementing a firmware download to endpoints in a mesh network. A firmware package is divided into a number of blocks, each block containing a number of packets and sent as a broadcast to endpoints in a wireless mesh network. A number of redundancy packets are sent to permit calculation of lost packets. A maximum number of redundancy packets to be sent is determined and a forward error correction code is developed depending on the maximum number of redundancy packets and the total number of firmware packets to be sent. A first block of redundancy packets, less than the maximum number, is sent followed by additional blocks of redundancy packets depending on reports from the receiving endpoints.
1. A method for downloading firmware, comprising:
dividing an image of firmware to be downloaded into a plurality of blocks, each block containing a plurality of packets;
broadcasting, by at least one transmitting apparatus, the plurality of blocks;
appending at least one redundancy packet to the broadcast of each of the plurality of blocks; and
computing missing packets within the blocks based on the at least one redundancy packet;
wherein the last block of the plurality of blocks is adjusted in length to fit an exact number of packets to be downloaded.
2. A method as in claim 1 , wherein the number of redundancy packets is a percentage of the total number of packets broadcast.
3. A method as in claim 2 , wherein the percentage is about 5%.
4. A method as in claim 1 , wherein the combined length of the plurality of packets and the at least one redundancy packet corresponds to the length of a forward error correcting code word.
5. A method as in claim 4 , wherein the forward error correcting code word is a Reed-Solomon code word.
6. A method as in claim 5 , wherein computing missing packets comprises Reed-Solomon decoding using an erasure procedure.
7. A method as in claim 1 , further comprising:
reporting computation of missing packet results; and
broadcasting at least one additional redundancy packet based on the reported results.
8. A method as in claim 7 , wherein each broadcast and computation step uses the same Reed-Solomon code.
9. A method as in claim 8 , wherein the Reed-Solomon code is based on a predetermined maximum number of packets to be broadcast.
10. A method as in claim 1 , further comprising sending an acknowledgment request.
11. A method for transferring software, comprising:
dividing the software into a plurality of blocks, each block containing a plurality of packets;
establishing a predetermined maximum number of redundancy packets to be downloaded with the plurality of packets;
setting a forward error correction code word to a length equal to that of the total of the plurality of software packets and the predetermined maximum number of redundancy packets;
transferring, by at least one transmitting apparatus a first block of a subset of the plurality of software packets;
transferring a first block of a subset of the predetermined maximum number of redundancy packets; and
computing missing software packets based on the first block of the subset of redundancy packets using the forward error correction code word.
12. A method as in claim 11 , wherein the forward error correction code word is a Reed-Solomon code word.
13. A method 11 as in claim 12 , further comprising:
transferring at least one second block of a subset of the predetermined maximum number of redundancy packets; and
computing missing software packets based on the first and at least one second blocks of the subsets of redundancy packets.
14. A method as in claim 11 , wherein the predetermined maximum number of redundancy packets is a percentage of the total number of software packets to be transferred.
15. A method as in claim 14 , wherein the percentage is about 5%.
16. In a wireless mesh network including a head end and at least one cell relay configured to provide communications between the head end and one or more endpoints, a method for updating endpoint firmware, comprising:
dividing an image of firmware to be downloaded into a plurality of blocks, each block containing a plurality of packets;
broadcasting, by at least one transmitting apparatus the plurality of blocks;
appending at least one redundancy packet to the broadcast of each of the plurality of blocks;
computing missing packets within the block based on the at least one redundancy packet;
reporting computation of missing packet results to the cell relay; and
broadcasting, by at least one transmitting apparatus, from the cell relay at least, one additional redundancy packet based on the reported results.
17. A method as in claim 16 , wherein the number of redundancy packets is a percentage of the total number of packets broadcast.
18. A method as in claim 17 , wherein redundancy is about 5%.
19. A method as in claim 16 , wherein the plurality of packets is broadcast from the cell relay to the one or more endpoints.
20. A method as in claim 16 , wherein the last block of the plurality of blocks is adjusted in length to fit an exact number of packets to be downloaded.
21. A method as in claim 16 , wherein the combined length of the plurality of packets and the at least one redundancy packet corresponds to the length of a forward error correcting code word.
22. A method as in claim 21 , wherein the forward error correcting code word is a Reed-Solomon code word.
23. A method as in claim 22 , wherein computing missing packets comprises Reed-Solomon decoding using an erasure procedure.
24. A method as in claim 16 , wherein each broadcast and computation step uses the same Reed-Solomon code.
25. A method as in claim 24 , wherein the Reed-Solomon code is based on a predetermined maximum number of packets to be broadcast.
26. A method as in claim 16 , further comprising sending an acknowledgment request from the cell relay.
27. A method as in claim 16 , wherein at least one of the endpoints comprises a meter.