Routes to support seamless migration: location identification/Identity

The location identification/Identity separation Protocol (LOCATOR/ID separation Protocol) provides a packet routing method that enables seamless migration of virtual machines (VMS) between servers without changing the terminal software, Allow mobile devices to migrate seamlessly between Wi-Fi and wireless networks. The LISP IETF RFC draft is scheduled to be released this fall. Prior to this agreement was a draft RFC in early 2012, as well as several public domain Lisp implementations and a beta network established in 2010.

Why do I need a location identification/identity separation protocol?

Traditional IP forwarding mechanisms do not apply to nodes that migrate between networks. Because the IP address (IPv4 and IPV6) contains a network number and a NIC address, forwarding is more difficult. The network number can be used to forward packets to the appropriate network, and the NIC address can be used to select network nodes. Once the node is moved to another network, the network segment on the address is invalidated. If you want to update the terminal IP address when the node is moving, you need to modify the node software, but this method does not apply to VMS or mobile devices.

How does the location identification/identity separation protocol work?

In a traditional network, the terminal node does not move, and the node that needs to send the packet finds the terminal identifier (end point Identifier, EID) of the target node through DNS to find or check packets received from the target. Eid is the IP address of the target node, while the routing location identity (Routing Locator, Rloc) can precisely determine the destination address of the local network. Packets will be sent from the local network to the router that provides the Internet access. This router is responsible for determining how to reach the target network using a standard routing protocol.

Definition of position identification/identity separation protocol

The Lisp standard contains the following definitions:

The eid--terminal identifier is the IP address that is assigned to the terminal when it is initialized, and it is typically determined by the Dynamic Host Configuration Protocol (Dynamics host Configuration Protocol). The Eid of a device or virtual machine will not change during network migration.

The etr--Exit Channel router (egress tunnel Router) is the router that is responsible for transferring packets from the Internet to the local network, while the target terminal node is located on the local network.

The itr--gateway router is the router that is responsible for transferring packets from the local network to the Internet, while the source node is located on the local network. As packets move in two directions, routers will be the ITR of packets leaving the network, and the ETR of packets from the Internet to network nodes.

The rloc--routing location identity refers to the network identifier portion of an IP address. It can be used to transfer packets from ITR to ETR.

When a node moves, the traditional packet forwarding method is invalidated. The node that sends the packet learns the Eid of the target via DNS or a packet that was received in the traditional way before. However, in mobile networks, the rloc contained in the Eid is not valid. The Eid of a mobile node is allocated when the node is initialized. When a node is moved, the Rloc part can no longer be used to forward packets to a moving target.

Lisp solves this problem by adding an extra IP header to the packet that leads to the mobile node. The Portal channel router (ITR) learns the current location of the mobile node through the mapping feature defined by the position identification/Identity separation Protocol (LISP). When ITR receives a packet from a local network node, it queries the Lisp mapping feature. The mapping feature Returns a (response) Rloc that determines the path to the current location to the target node.

ITR then adds a traditional header in front of the left packet and then adds the Rloc learned from the mapping service. The packet is then forwarded to the Internet via a standard routing protocol. Lisp does not need to modify any routing protocol or Internet core router.

When the destination network is reached, the receiving port Exit Channel router (ETR) deletes the added header. It then forwards the packet from its local network to the target node. For the target node, the packet received is as if it had never moved. So Lisp does not require any modifications to the node software.

ITR will cache the mapping service query results, so you do not need to query each subsequent packet. If the target node moves when the connection is maintained, the mapping service provides an updated rloc.

Mobility Benefits of Location identification/identity separation protocol

Agreement:

Allows a site to switch to an Internet service provider without renumbering the terminal node

Enable multihomed nodes to spread the load across multiple links

At the same time support IPV4 and IPv6, simplify IPv4 to IPv6 transition

mapping function of location identification/identity separation protocol

After receiving the ITR request, the Lisp mapping feature responds to the rloc required for the remote node. There are several proposed designs, but the current work includes some network components as "mapping Servers" and "Mapping parsers". ETR sends a "map registration" message to the mapping server on a regular basis. The map registration message enumerates the Eid currently available through the ETR.

The mapping server is connected through a generic routing encapsulation (Generic Router encapsulation, GRE) channel. Each mapping server uses an Edge gateway patrol (Border Gateway Patrol, BGP) to distribute Eid information from the ETR to other mapping servers. True Eid to rloc information is stored on the ETR. However, there is not a single database that stores Eid to rloc mappings centrally.

When a ETR needs to forward a packet but does not know which rloc to use, it sends a "map request" to the mapping parser. The mapping parser uses the information learned through BPG to determine the corresponding ETR. The mapping parser then forwarded the mapping request to ETR, ETR the original Itr.itr will append additional header information to the packet and add the received Rloc to the destination address field by sending the mapping response.