Hi, Well, it's clearly explained in this paragraph isn't it ?
*In order to maintain the lowest-latency connection between* *client and server, the server’s virtual machine is migrated* *between virtual machine hosts located in different parts of* *the building, to always be as close to the mobile clients as* *possible. This move must be seamless: as the VM moves,* *applications on the VM and the mobile client should con-* *tinue to communicate, without requiring changes to their IP* *or MAC addresses. * One specific VM runs on an Hypervisor and there is often (always?) more than only one of it in a virtualized environment. So when you have to run a VM, you have to choose on which Hypervisor to do it. More generally VMs can have to move for load balancing purpose (to equilibrate charge between hypervisors) Hope it helps, ArnaudM ** 2013/3/24 Cyril Auburtin <cyril.aubur...@gmail.com> > In this old paper > http://conferences.sigcomm.org/sigcomm/2008/papers/p513-ericksonA.pdf, I > don't understand much why "VM moves" (2nd paragraph), aren't they all > somewhere physically on a server behind the router, I'd like to have more > explanation about where VM are run. > > Thanks > > _______________________________________________ > openflow-discuss mailing list > openflow-discuss@lists.stanford.edu > https://mailman.stanford.edu/mailman/listinfo/openflow-discuss > >
_______________________________________________ openflow-discuss mailing list openflow-discuss@lists.stanford.edu https://mailman.stanford.edu/mailman/listinfo/openflow-discuss
