Foldable ingress buffer for network apparatuses

A method comprising:

• receiving data at a plurality of inputs of a network apparatus, each input receiving up to a maximum portion size of the data each input clock cycle of a plurality of input clock cycles;
• multiplexing two or more inputs of the plurality of inputs by a plurality of packed ingress interfaces;
• sending data unit portions of the data to writers coupled to the packed ingress interfaces, the data unit portions forming Transport Data Units (“TDUs”);
• the writers concurrently writing in each of a plurality of memory access cycles the data unit portions to a folded buffer, the folded buffer comprising a plurality of memory banks arranged in a grid having columns and rows, a sum of memory bank widths for each row of the folded buffer is sized to accommodate a maximum-size TDU divided by a number of folds, each writer writing each data unit portion of TDUs of a maximum TDU size to a different bank in the plurality of memory banks, at least two portions of the TDU written in each column of the grid, the number of portions in each maximum-size TDU that are written to any given column of the grid is the same as the number of folds;
• reading, by a reader, the TDUs from the folded buffer up to a number of portions equal to the number of folds each memory access cycle of the memory access cycles;
• sending the TDUs to one or more downstream components of a network apparatus.

One or more non-transitory computer-readable media storing instructions that, when executed by one or more computing devices, cause performance of:

• receiving data at a plurality of inputs of a network apparatus, each input receiving up to a maximum portion size of the data each input clock cycle of a plurality of input clock cycles;
• multiplexing two or more inputs of the plurality of inputs by a plurality of packed ingress interfaces;
• sending data unit portions of the data to writers coupled to the packed ingress interfaces, the data unit portions forming Transport Data Units (“TDUs”);
• the writers concurrently writing in each of a plurality of memory access cycles the data unit portions to a folded buffer, the folded buffer comprising a plurality of memory banks arranged in a grid having columns and rows, a sum of memory bank widths for each row of the folded buffer is sized to accommodate a maximum-size TDU divided by a number of folds, each writer writing each data unit portion of TDUs of a maximum TDU size to a different bank in the plurality of memory banks, at least two portions of the TDU written in each column of the grid, the number of portions in each maximum-size TDU that are written to any given column of the grid is the same as the number of folds;
• reading, by a reader, the TDUs from the folded buffer up to a number of portions equal to the number of folds each memory access cycle of the memory access cycles;
• sending the TDUs to one or more downstream components of a network apparatus.