Next-generation protocols related to IPv6

Next-generation protocols related to IPv6 are in the TCP/IP protocol set. Protocols that are directly active and indirectly interoperable with IP addresses include various application layer protocols, connection protocols, and TCP and UDP. This chapter will discuss the impact of IPv6 on other layer protocols if these protocols are modified or need to be modified to adapt to protocol layers: www.2cto.com, it is important to understand that these protocols use IP addresses. Because so many systems rely on a large number of TCP/IP network technologies and application protocols, it is important that the upgrade of IP does not have to perform extensive upgrades to the upper or lower layer protocols of IP addresses. Therefore. In addition to IPv4, most existing TCP/IP applications, software and hardware can work with IPv6. Application Layer: WWW and e-mail are the most widely used applications today. WWW and e-mail users must point to services on the Internet to work. Traditionally, these customers can accept the host name or IP address of the node. When using a domain name, you can call the Domain Name System (DNS) to obtain the IP address corresponding to the host name, and then use it at the transport and Internet layers.
For a simple application, it is easy to make it work with IPv6: you can rewrite the software so that it can accept and correctly handle IPv4 and IPv6 addresses; or you can only access it by host name. The former method retains the application's ability to directly address nodes, but is relatively complicated. The latter method only removes features that are not used or even unnecessary by most users. However, considering the features of IPv6 Security, service quality, or other features, some applications want to use IPv6, which requires more extensive updates. Transport Layer: In most cases, IP addresses have nothing to do with the application layer protocol, but are closely related to the transport layer protocol. Both the UDP and TCP pseudo headers use the source IP address and destination IP address, and the TCP circuit is defined by the IP address and port number of the source and destination nodes. If you want to interoperate with IPv6, and can correctly calculate the pseudo header. For TCP, the implementation must also be able to manage the circuit that gives IPv6 addresses. After the release of the first IPV6 RFC, some concerns emerged. TCPng is required to supplement IPng. At present, TCP has a problem when dealing with mobile nodes: Determine the IP addresses of the source and target nodes required for the TCP circuit. If the IP addresses of one or both parties change in the TCP interaction device, the circuit identifier will be faulty. This occurs when a mobile node switches from a network address to another network address. For example, a node on a train or car uses a wireless network for access, or the node linked to the network changes the ISP situation for a better rate at night. Www.2cto.com is caused by the fact that TCP has at least no mechanism yet to allow IP addresses to be changed in the connection. If a node receives the source IP address in the TCP segment and the address negotiated when the TCP circuit is created is different, the node considers the TCP segment to belong to another circuit. This means that the Mobile IP address currently does not support the activation of the TCP circuit from one network address to another. The problem of TCPng is much more complicated than that of running TCP connections supporting network address translation. The problem is that supporting such an address will lead to a security vulnerability: attackers can easily impersonate a node that is converted from one network to another, just as an authorized node is converted from one network to another. To solve this problem, a major upgrade of TCP is required. You must introduce a mechanism to enable a node to prove itself to other nodes when its IP address is changed. Currently, if a Mobile IP address switches the network between TCp connections, it must negotiate the connection again after the switchover. In a sense, TCPng is necessary to support seamless interoperability between mobile devices and hosts. Link Layer: Compared with the above layer, link layer protocols such as Ethernet and ATM are rarely affected due to IPV6 upgrades. This is because these protocols only encapsulate the upper-layer datagram into the link layer frame. However, this does not mean that IPv6 has no effect on the link layer protocol. For example, ATM uses a point-to-point circuit to transmit data across networks. For services that require that Ipv6 package to deliver multiple nodes, ATM must pay special attention. Link Layer issues that may be affected by IPv6 include path MTU discovery and Address Resolution Protocol (ARP). These protocols need to be modified to support 128 of Ipv6 addresses. The IPv6 domain name system can be used to expand Internet applications. DNS is an important factor: it makes it easy to map names to IP addresses. DNS uses a hierarchical namespace. Some servers at each level of soybean oil help map names to addresses. The host name may be in the form of "host.organization.com", indicating that the host is in the domain organization. If the node in organization.com needs to find the host, j queries the local DNS server, which maintains the host Name and address information in organization.com. It simply searches for the host, the 32-bit IP address corresponding to the host is used to answer node requests. If a node other than organization.com needs the IP address of host.organization.com, It queries the local DNS server, which must be queried and maintained. the upper-level server of the com network domain information. Then, the upper-level server directs the request to the DNS server in the organization.com domain, and the server will eventually respond to the request, send the requested IP address to the local DNS server, and then the local DNS server sends the information to the node that sends the request. So far, everything went well. However, DNS was originally designed to process 32-bit Ipv4 addresses. RFC1886 (DNS extension Supporting Ipv6) describes the necessary modifications to make DNS support for IPv6. This RFC is very short. It briefly describes three modifications made to make DNS apply to Ipv6. * Create a new resource record type (called the AAAA record type) to map the Ipv6 address whose name is 128-bit. IPv4 resource records use the record type. * Create a new domain, that is, .IP6.int, which is used to add IPv6 host addresses to support address-based search. That is, the request node wants to know the domain name of an Ipv6 address. An IPv4 address also has a similar facility, namely, A. in-addr.arpa. * You must modify the existing DNS query so that it can not only locate the active IPV4 address, but also handle the coexistence of IPv4 and IPv6 addresses. Www.2cto.com Address Resolution Protocol and Neighbor Discovery: IPv6 does not execute Address Resolution Protocol (ARP) or Reverse Address Resolution Protocol (RARP ). ARP is not used in IPv6 for the following reasons: first, ARp depends on IPv6 and ICMPv6 packets using group packets, this means that there is no need to reconstruct ARP for each different type of network using ARP. Any node supporting IPv6 and multicast should also support Neighbor Discovery. It is important to support multicast, especially at the link layer. Like broadcast, multicast is easily implemented on networks that support multiple channels to access the same media at the same time, such as Ethernet. However, for so-called non-broadcast multiple access (NBMA) networks, such as ATM and frame relay, multicast is difficult to process ................. RFC1970 (IPV6 Neighbor Discovery) describes the neighbor discovery mechanism. It provides several different uses, including the following support: * route discovery, that is, it helps hosts identify local routers * prefix discovery. The node uses the stimulus value to determine the address prefix of the local link address and the address prefix that must be sent to the route forwarding. * Parameter discovery. This mechanism helps the node determine information such as the Local Link MTU.
* Automatic address configuration. Used for automatic configuration of IPV6 nodes. * Address resolution. Replace ARP and RARP to help the node identify the local node 9 (neighbor) from the destination IP address. * Next Hop is determined. It can be used to determine the next destination of the package, that is, to determine whether the destination of the package is on the current link. If the next hop in this link is the destination, otherwise, the package needs to be selected, and the next hop is the vro, and the neighbor finds that it can be used to determine the vro to be used. Www.2cto.com * neighbors cannot be monitored. Neighbor Discovery helps determine whether a neighbor (destination node or router) is reachable. * Duplicate address monitoring. It is also found that it can be used to help the node determine whether the address it wants to use is occupied on the local link * redirection. Sometimes the forwarding router selected by the node is not the best choice for the package to be forwarded. In this case, the forwarding route can be relocated to the node to send the packet to the best router. * A route notification requires the router to periodically advertise its availability and the link and Internet parameters used for configuration. * Router request. The host can request the local router to send the router notice immediately. * Neighbor announcement. When a node receives a request from a neighbor request or its link layer address changes, it sends a neighbor notification message. * Neighbor requests. The node sends a neighbor request message to request the neighbor's link layer address to verify that it obtains and maintains the accessibility of the neighbor link layer address in the cache, or verify that his own address is unique on the link. * Redirection. The router sends a redirection message to notify the host. It is not the best router for a specific destination. Routers send their router notification messages through multicast, so that nodes on the unified link can construct their own list of available Default routers: www.2cto.com Neighbor Discovery can also be used to achieve other objectives, including; * link layer address changes. For the same network, a node can have multiple interfaces. If the node knows that its link layer address has changed, you can send several multicast packets to notify other nodes of their address changes. * For inbound load balancing, note that nodes receiving a large number of business flows may have multiple network interfaces, which can be identified by neighbors using one IP address. The router load balancing can be achieved by letting the router omit the source link layer address when sending its router notification package. In this case, each time the node of the router is located to send a package to the next day, the router must perform Neighbor Discovery, And the router can select the link layer interface for receiving the package to correspond to this node. Www.2cto.com * any on-demand address. Any VOD address represents a set of unicast addresses. packages sent to any VOD address are delivered to any of these addresses. Generally, any VOD address is used to represent a node set that provides the same service. That is, a node that sends a packet to any VOD address does not care which node set responds. Because multiple members of any VOD address may respond to requests to the link layer address, the neighbor discovery mechanism requires that the node may receive multiple responses and handle them correctly. * Proxy notice: If a node cannot correctly respond to a neighbor request, the neighbor discovery mechanism allows a node to represent the node.