Talking about the routing mechanism in IPv4 and IPv6
Li Ruigang Fan Lihua
Summary
In standard IP protocols, the optimization of routing paths is maintained and updated by routing protocols. However, this mechanism is not applicable to the mobile node. The forwarding of communication-to-end nodes to mobile node data packets must always pass through their hometown agents. The routing optimization mechanism allows data to be routed directly in both directions. At the same time, routing also provides support for smoothing. In addition, due to the new features of IPv6, the IP protocol has better support for mobility.
Key words
Hometown agents, hometown address, transfer address, binding cache
Abstract
In standard IP routing routes are constantly optimized as routing tables are propagated throughout the Internet. In the case of mobile hosts that wish to use their home addresses this mechanism does not work. Thus while communicating sub-optimal routing often takes place. The problem results from the routing of packets to the mobile node via its home agent. Route optimization enables the datagrams to be routed directly in both directions. Route optimization also provides support for smooth handoffs. Additionally, with the new characteristics of IPv6, better mobility support can be Resulted from Internet Protocol.
Keyword
Home Agent, Home Address, Care-of Address, Binding Cache, Triangle Routing
1 Introduction
Mobile IP is a series of standards for extended IP network mobility developed by IETF (Internet Engineering Task Force, which is a fixed IP address that connects mobile terminals connected to any network and can continue to receive IP packets. RFC2002 IP Mobility Support is the main criterion of mobile IP, which describes the architecture and working mechanism of mobile IP.
2. The principle of mobile IP protocol
The working principle of the mobile IP protocol is approximately as shown below:
1) When the mobile terminal is in the net, it is still in communication with the traditional TCP / IP mode, and the mobile IP protocol is not required.
2) When the mobile terminal roams to a field network, the fixed IP address is still communicated with a fixed IP address. In order to be able to receive the IP packet to its IP packet, the mobile terminal needs to register the current location address to the home agent, this location address is the transfer address. Mobile IP can get an address and dynamically configure an address of the forwarding address-external agent in two ways.
3) After the home agent accepts registration from the transfer address, it will build a tunnel to the transfer address, and the IP packet of the intercepted transmission to the mobile terminal is sent through the tunnel to the external network transfer address.
4) Remove the tunnel package at the turnover address, restore the original IP packet, and finally send it to the mobile terminal so that the mobile terminal can receive these IP packets sent to it.
5) The mobile terminal sends an IP packet to the communication peer from the router or outer agent of the external network.
6) When the mobile terminal comes to another external network, you can continue to communicate only to update the registered input address to the home agent.
7) When the mobile terminal returns home network, the mobile terminal logs out of the translated address to the home agent, and the mobile terminal will communicate using traditional TCP / IP mode. 3. Move the routing problem in IPv4
The existing mobile IP solution has a famous triangle routing problem, as shown in Figure 2. The path passed by the communication peers sent to the mobile host is: Communication peer, hometown agent, field agent, mobile host. The path passed by the mobile host sent to the communication peer is: mobile host, field agent (default router), communication peer.
Hometown agent
Field agent
Mobile host
Communication peer
Figure 2, triangle route
The mobile host sent to the communication peer-to-end communication is used by the best route provided by the routing protocol, and the communication peer transmits to the mobile host cannot use the best route provided by the routing protocol. The root cause of triangular routing is that the communication peer does not know the current location of the mobile host, ie, I don't know the transfer address of the mobile host. If the communication peer knows the transfer address of the mobile host, he sends a message to the mobile host without first sending a hometown agent for the mobile host, and can be sent directly to the current transfers in the tunnel.
4. Routing optimization used in mobile IPv4
For existing mobile IP protocols, all packets that are sent to mobile nodes must first be sent to their hometown agents, and then pass each packet to the current location of the mobile node via the Tunnel technology. Although this method solves the problem of the existing IP protocol does not support terminal mobility, it is not ideal for performance from the perspective of the efficiency. The reason is that each time the data packet must go to the packet to reach the destination with the mobile node and the peer node. This indirect routing mechanism not only prolongs transmission delays, increasing the packet loss rate, but also brought unnecessary burden to the network and routers in invisible. Therefore, it must be optimized for existing mobile IP routing mechanisms. The following describes an optimization method currently being universally identified, and this optimization method is also recommended by the IETF Mobile IP Working Group.
Overall, this optimization method is mainly to save a cache pair within a peer node that communicates with the mobile node, and is used to record the hometown address and the transfer address of the mobile node. With the movement of the mobile node, this cache is constantly updated. With this cache mechanism, it is possible to implement a tunnel directly between the peer node and the mobile node to communicate, thereby avoiding the various drawbacks brought about by the home agent.
The key to this optimization technology is reasonable maintenance and update of address cache pairs within the communication-to-end node. When the communication peer is in communication with the mobile node, the data is first sent to the hometown agent. At this time, different from the traditional mobile IP protocol is that the hometown agent not only makes the data not only tunnel, but also forwards to the field agent, but also saves themselves. The address cache is sent to the communication peer so that the address binding cache of the mobile node is also stored in the communication pair node. In the communication process after this, the two-way direct selection of the mobile node and its communication peer node can be formed, thereby avoiding various defects caused by "triangular routing". During the mobile node continuously moves, each new network, the address cache in the hometown agent is updated, and notify all the communication to keep communication, so that the internal cache is synchronized. Moreover, like the address cache in the hometown agent, the address cache in the communication peer node also has its own living time. When the living time is exhausted, this cache must abandon and re-request new binding cache. In order to achieve smooth switching, the routing optimization mechanism saves the address of the previous field agent of the mobile node, so when the mobile node is registered in a new field agent, all the data forwarded all the data forwarded to this new new The field agent, thereby avoiding problems with data loss during the continuous movement of the mobile node.
This routing optimization mechanism seems to be quite intelligent, but it is difficult to achieve in the actual production process. Because this requires a large-scale improvement to the communication peer operating system, it is difficult to resolve the security issues between the mobile nodes and their communication peers. The expected situation will improve after IPv6 is widely used. Hometown agent
Field agent
Mobile host
Communication peer
Figure 3, Routing Optimization
5. Move IPv6 support for mobility
IPv6 supports the design of mobile IP (mobile IPv6), draws the successful experience of mobile IPv4 design, while also has many new features. Mobile IPv6 has many common features with mobile IPv4, but mobile IPv6 is now fully integrated into IP and has a lot improvement than mobile IPv4. Here, the main difference between mobile IPv6 and mobile IPv4 is summarized:
"Routing Optimization" in Mobile IPv4 is a built-in protocol part in mobile IPv6, not as an extension option, and mobile IPv4 nodes may not be supported. This integrated routing optimization feature allows any communication nodes and mobile nodes to communicate directly, and does not require the host network through the mobile node and forward the "triangle routing" problem that occurs based on mobile IPv4 protocols. The "Registration" function of mobile IPv4 and routing optimization functions are performed as a single protocol in mobile IPv6 rather than two separate (and different) protocols.
Let the mobile node and mobile IP effectively coexist with the router that performs "inlet filtering", mobile IPv6 and IPv6 itself integrate this feature. When sending a packet, the source address of the IP header uses the transfer address of the mobile node, so the packet can be normally filtered through the inlet. The host address of the mobile node is a host address target option with a packet such that the transfer address is transparent to the protocol layer above the IP layer.
The "external agent" used in mobile IPv4 is no longer needed in mobile IPv6, and the mobile node utilizes the features of IPv6, such as neighbor discovery and address automatically formulate, and no local router is required or operated anywhere.
When the mobile node leaves the host network, most of the groups sent to it use IPv6 routing heads rather than IP packages, while mobile IPv4 must use IP packages. The use of routing heads reduces the number of additional header bytes, and reduces the load of the mobile IP packet. However, in order to avoid the transfer group, the group is intercepted by the host agent, and the packet forwarded by the tunnel is still using the package. When the mobile node leaves the host network, its host agent cuts the packet to the host network,
Use IPv6 neighbor to discover rather than the ARP used by mobile IPv4. The use of neighbors enhances the aggressiveness of the protocol and simplifies the application of mobile IP because the link layer to be considered in ARP is required.
Move the dynamic host agent address discovery mechanism in IPv6, use IPv6, return a single reply to the mobile node, not the mechanism used by mobile IPv4, use IPv4 direct broadcast, and all host agents on the host link return a reply Give the mobile node. Mobile IPv6 mechanism is more efficient, more reliable, because only one packet is returned to the mobile node. Mobile nodes are unlikely to lose these replies because the protocol does not need to "block" reply.
6. Conclusion
Traditional IPv4 supports mobility far from meeting the needs of growing mobile nodes. For IPv4, it is expanded to achieve routing, and actively explore IPv6 support for mobility, and has always been the goal studied. The IPv6 protocol takes into account the support of mobility at the beginning of design, so we have reason to think that if you want to truly implement your mobility, IPv6 is an inevitable choice.
references
[1] C. Perkins., IP Mobility Support for IPv4, RFC 2002, October 1996
[2] Johnson D., Perkins C., Mobility Support in IPv6, Work in Progress, 22 October 2002
[3] C. Perkins., Johnson D., Route Optimization In Mobile IP, Work In Progress [4] C. Perkins, IP Encapsulation Within IP, RFC2003,
[5] J. Solomon Applicability Statement for IP Mobility Support, RFC 2005, October 1996
[6] IAMES D.SOLOMM, Mobile IP, Machinery Industry Press
[7] SLVANO GAL, IPv6 Network Interconnection and Cisco Router, Machinery Industry Press
[8] Perkins.c, Mobile IP: Design Principles and Practices, Addison-Wesley, C1998.