Nicolas Williams wrote: > On Wed, Oct 17, 2007 at 11:24:26AM -0400, Bill Sommerfeld wrote: > >> On Wed, 2007-10-17 at 16:07 +0100, Paul Jakma wrote: >> >>> Firmware 'near' the NIC (I've never seen an explanation of the exact >>> mechanism) interposes itself between hardware and OS and 'hijacks' >>> traffic to that port. It never makes it to the OS. >>> >> one implication of this is that, when this is in use, the NIC cannot be >> part of a L2 aggregation (because the upstream switch will load-spread >> some of the management traffic flows to other ports in the aggregation). >> >> And using it with other L2/L3 redundancy technologies (such as IPMP and >> OSPF-MP) is going to be tricky (the shared management IP address must >> not be seen as reachable via other NICs). >> > > My impression was that the AMT chip gets its own IPv4 address. >
That is an option. But it is not the default configuration. In the default configuration, both the host and the embedded environment use the same MAC address, using DHCP, and thus share an IP address. If static IP addressing is used, then the embedded environment can be assigned its own MAC address, and its own IP address. This is only true if the NIC is the onboard wired NIC. For WiFi, I believe that only a shared MAC address configuration is possible. (And only a small number of configurations have the WiFi support using only Intel WiFi cards, of course.) And yes, I do realize that it should have been possible to assign two MAC addresses and let each do DHCP on its own behalf. However, for reasons I don't understand, this is not the design choice that Intel took. -- Garrett
