On Mon, Oct 6, 2014 at 5:49 PM, Sri Gundavelli (sgundave) <[email protected]<mailto:[email protected]>> wrote:
On 10/6/14 2:16 PM, "Tom Herbert" <[email protected]<mailto:[email protected]>> wrote: > Clarifying question. Trying to understand the comment here; Not promoting a specific approach/technology. Does migration of "IP addresses and connection state" does not equate to Layer-3 mobility ? > I believe they are mostly equivalent. Yes. That was point. I almost thought there is a new mobility requirement that stands out here and was trying to understand that requirement. If you take the case of a VM Migration (or any other light weigh virtualization models such as LXC), the mobility requirements related to migration across Layer-3 boundaries are the exactly same as for what typical layer-3 mobility protocols are designed for. So, looking at the properties: A VM has an IPv4 address and/or a set of IPv6 prefixes; It can change its point of attachment moving across layer-3 network boundaries VM migration comes with Application continuity requirement; Handover latency should be very small (Ex: 50 msec) is desired VM's can move as a cluster or as a individual node. VM's can be multi-homed with application based path selection VM's can negotiate QoS policies and provide Guaranteed-bit rates for set of flows .. And so on and so forth =====> Now, if I replace "VM" with "mobile node", that's layer-3 mobility and we have few solution options there… > Yes, although the identifier is probably not routable. The locator (routable) > is either the outer IP address in an encapsulation, or maybe something like > upper 64 bits in IPv6 address (where lower 64 are set to identifier). LISP > and ILNP are good references here. Right. That's from the LISP jargon; Sure, its one of the protocol options. Regards Sri
_______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3
