Fast SDN updates using tree-based architecture

Efficient TCAM design based on dual port SRAM on FPGA

“…1 0 0 0 1010 [10] 0 0 0 0 1011 [11] 0 0 0 0 1100 [12] 0 1 0 1 1101 [13] 0 0 0 0 1110 [14] 0 0 0 0 1111 [15] 0 0 0 0 Comments: The use of 1 RAM gives us advantages and disadvantages such as a) Advantage: Simple and easy to use. b) Disadvantages: extremely large memory usage for bigdata data types.…”

“…1 0 0 1010 [10] 0 0 0 1011 [11] 0 0 0 1100 [12] 0 1 0 1101 [13] 0 0 0 1110 [14] 0 0 0 1111 [15] According to Principle 2, the output result will be retrieved when we get the address = key. For the example above, we have Output = RAM[1100] = 0101 →We have the key results that match rule 1 and rule 3.…”

“…The pipelined-lookup method in [2,9] is efficient and could be used to speed up grid of BRAMs. We decided not to use decision tree and binary tree method in [10][11][12][13][14] as this increases complexity and require more software solution. Other type of solution involves splitting and cutting the original rule set into many subsets [15,16] which faces the same complexity as tree based method.…”