4 ms·
No, but datalink stacks have acted like they are for decades, and it's finally catching up with them.
by bandrami 10y ago
No, but datalink stacks have acted like they are for decades, and it's finally catching up with them.
- tener 10y agoI guess someone calculated that given current rate of sequential assignment of MAC ranges, we won't get to 4 and 6 prefixes in a very long time. That reasoning was broken as it assumed sequential assignment is the only way to go.
- packetized 10y agoUntil fairly recently (in IEEE/OUI terms, ~2012), the distribution has been entirely sequential. FTA: "I believe IEEE changed their strategy to attempt to purposefully higher the chance of collisions with MAC squatters, to encourage people to register and pay the fee"
- bandrami 10y agoNonetheless it was a broken parse: they needed to expend the memory to figure out what the actual protocol is and not jump up from the datalink layer if it's not called for. I guess I'm just saying I really can't blame IEEE here.
- packetized 10y agoOh, I don't blame the IEEE wholly - but I think that they probably changed tack after realizing that a number of manufacturers were simply not paying the fees to actually be assigned a block of OUIs. I view it as kind of a cudgel to be used against vendors, and end-users (read: netops/sysops) are caught in the middle.