WO2012075688A1 - 一种可以取代ipv6的网络层协议 - Google Patents

一种可以取代ipv6的网络层协议 Download PDF

Info

Publication number
WO2012075688A1
WO2012075688A1 PCT/CN2011/002057 CN2011002057W WO2012075688A1 WO 2012075688 A1 WO2012075688 A1 WO 2012075688A1 CN 2011002057 W CN2011002057 W CN 2011002057W WO 2012075688 A1 WO2012075688 A1 WO 2012075688A1
Authority
WO
WIPO (PCT)
Prior art keywords
network
super
block
url
protocol
Prior art date
Application number
PCT/CN2011/002057
Other languages
English (en)
French (fr)
Inventor
谭中飞
曹化喆
Original Assignee
Tan Zhongfei
Tsao Roger Huazhe
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tan Zhongfei, Tsao Roger Huazhe filed Critical Tan Zhongfei
Priority to EP11846093.0A priority Critical patent/EP2651088A1/en
Priority to US13/992,740 priority patent/US20130275608A1/en
Publication of WO2012075688A1 publication Critical patent/WO2012075688A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L69/00Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
    • H04L69/16Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
    • H04L69/167Adaptation for transition between two IP versions, e.g. between IPv4 and IPv6
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L2101/00Indexing scheme associated with group H04L61/00
    • H04L2101/60Types of network addresses
    • H04L2101/618Details of network addresses
    • H04L2101/659Internet protocol version 6 [IPv6] addresses
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L2101/00Indexing scheme associated with group H04L61/00
    • H04L2101/60Types of network addresses
    • H04L2101/668Internet protocol [IP] address subnets
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/74Address processing for routing
    • H04L45/741Routing in networks with a plurality of addressing schemes, e.g. with both IPv4 and IPv6
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L61/00Network arrangements, protocols or services for addressing or naming
    • H04L61/45Network directories; Name-to-address mapping
    • H04L61/4505Network directories; Name-to-address mapping using standardised directories; using standardised directory access protocols
    • H04L61/4511Network directories; Name-to-address mapping using standardised directories; using standardised directory access protocols using domain name system [DNS]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L61/00Network arrangements, protocols or services for addressing or naming
    • H04L61/50Address allocation
    • H04L61/5007Internet protocol [IP] addresses

Definitions

  • the present invention relates to a method of expanding the space of an internet website.
  • I ETF Internet Engineering Task Force
  • RFC1385 envisages the expansion of the Internet in the form of "in situ split autonomy”: only the geographically adjacent network areas are connected by gateways, and the gateway between network areas utilizes the newly added "network number” (not indicating whether 32-bit) URL translation to achieve addressing between different network areas.
  • Another valuable proposal in RFG1385 is to store the "network number" in a newly established IPP4 option for transmission to avoid disrupting the current Internet protocol framework.
  • the present invention takes the reasonable part of RFC1385 and Chinese Patent Application 200710073741 as the prior art, inherits the method of dividing the network area, using the new option of I Pv4 to store newly added information, and stacking the website, seeking to not destroy the compatibility with the current Internet. Expand the URL space on the premise of sex and with as few changes as possible. Not changing the existing addressing mode means continuing to use the 32-bit user URL to continue to use existing I Pv4 hardware, continue to use existing routing systems, and minimize network software changes.
  • Confirmation Networking is the highest level of "tree roots", and it is still a central tree Internet; in the present invention, the highest level of the Internet will be composed of equal network blocks, neither tree structure nor center.
  • the invention is a network layer communication protocol, which can be called a super-linked protocol.
  • the English standard translation name can use Ultra Link Internet Protocol, abbreviated as ULIP.
  • the English translation of the other main keywords is recommended:
  • Block Network Block, referred to as Block
  • Subnet block Child Block
  • Network block group Block Group
  • Proxy Super Gateway Proxy Super Gateway
  • Routing Security Stamp Routing Security Stamp
  • the sender's website refers to the "Source Address” of I Pv4; the “receiving website” refers to the "Destination Address” of I Pv4.
  • the super-linked protocol is formally a legal extension of the current I Pv4 protocol and does not require changes to protocols outside the network layer.
  • the present invention uses a new method to connect the network blocks.
  • the present invention has nothing in common with RFC1385 and Chinese Patent Application 200710073741, except for the network block, the new I Pv4 option, and the URL stacking.
  • the gist of the present invention is to implement the following concepts and methods for efficient connection between network blocks:
  • Net block URL Select a 32-bit URL for each network block as the entrance to the network block, called the network block URL.
  • the above connection method of the super gateway is the basis for establishing the network block URL and establishing a super route.
  • Net block group Each pair of network blocks has a set of network blocks connected by a super gateway, which is called a network block group. Net block within the network block group Mutual network blocks. The same group of network blocks are equal to each other, so the new Internet will have no center.
  • Subnet block can be connected under the network block, and the upper network block is called the parent network block.
  • the two network blocks in the new Internet may have the same network block URL. However, at least one of the network blocks of the parent network block, the grandmother network block, and the like may have at least one difference, thereby uniquely determining the network block. s position.
  • a full URL contains several layers of URLs. If the user's complete URL inside a network block is 2 layers, the network block is called a double-layer URL network block, which is called L2 network block; and so on, there may be L3, L4, and L5 network blocks. There is no "single layer URL block": There is no parent block in the L2 block. An L3 network block must be a subnet block of an L2 network block, and an L4 network block must be a subnet block of an L3 network block, and so on.
  • the maximum capacity of the Internet (excluding the network block URL) is about 17 million times that of I Pv6 after expanding according to the hyperlink agreement.
  • the existing Internet will automatically become one of the L2 network blocks, which can be referred to as the "old network".
  • the old network is expected to be the only multinational shared L2 network block.
  • the L2 network block group will become the highest layer structure of the new Internet.
  • Sister network group Short for sister group.
  • the same layer and the same mother network block can form more than one network block group; these are the same layer and the same mother network block group (but must have A limited connection, called a sister network block. All network blocks in each sister group have the same number of URL layers and a common parent network block (L2 sister groups do not have a parent network block). Building a sister group can save the super gateway.
  • the L2 network blocks are successively established, or a single L2 network block group is formed, or several interconnected L2 sister groups are formed according to geographical and communication requirements, but the old network is expected to be a common member of all L2 sister groups.
  • Super-linking protocol option This is a new option for I Pv4 established for the super-linked protocol. It is used to store the net block URL, the host URL of the start or end point, and other information that the packet must pass during the transfer, such as the following table. Shown. The introduction of this option allows the hyperlink protocol to avoid disrupting the current network protocol system and allowing users in the newly created network block to access unupgraded websites in the old network. The idea of this option comes from RFC1385, which is designed in a specific format in accordance with the handling and security needs of super-routing. UL I P identification code optional 154.
  • the 4-bit "cross-network times” is used to record the number of times the packet crosses the boundary of the network block. The initial value is 0. Packets with the highest number of cross-networks but not yet reached the destination network block will be discarded. 4-bit "process pointer” pointing The next set of substitution URLs required for super routing, with an initial value of 3, and reaches the "maximum process” when the packet reaches the destination block. There are also 8 digits reserved for future innovation.
  • the basis for the super-linked protocol option to pass through the IPv4 network is the current Internet Protocol's handling of unfamiliar IP options: The host does not process (including not discard) the unfamiliar IP options received (see Internet Standard Document RFC1122, Section 3.2.1.8) ).
  • the hyperlink protocol option can also be ported for other purposes, such as establishing a giant private network.
  • Another major difficulty such as the seamless connection between IPv6 and IPv4, is that after IPv6 packets are "translated" into IPv4 packets, the IPv6 URLs "have nowhere to hide” in the packets.
  • the IPv6 URL is stored in the hyperlink protocol option, the IPv6 host can complete access to the IPv4 website with the assistance of the gateway at the IPv4 and IPv6 junction.
  • This means that network blocks with different URL lengths can also be integrated under one network through the hyperlink protocol.
  • the IPv6 network should join the Internet in the form of a network block, and multiple networks can also be established. IPv6 network block.
  • the network block URL of the IPv6 network block should be one of the IPv4 URLs, so that all the network block URLs can be freely converted one-to-one between the 32-bit and 128-bit formats.
  • All IPv6 network blocks must be L2 network blocks. This aspect is due to the huge capacity of the IPv6 network block, there is no need to establish a deeper IPv6 network block; on the other hand, in order to maintain the hyperlink protocol option, at most one IPv6 address is included in the IPv4 network block to avoid ambiguity.
  • the super gateway uniquely maps all IPv4 URLs in the packet header to IPv6 URLs. The hyperlink protocol option will naturally migrate to the IPv6 protocol just like other IPv4 options.
  • Super-routing The process by which the super-gateway collectively sends packets to the target network block.
  • a super gateway is like a mixture of a common host and a router.
  • the super gateway on the network block except for the connection of the parent network block and the subnet block, announces communication with each other in the routing table inside the network block.
  • the cross-sister group's super gateway should also announce all the network blocks in the other side network block group.
  • the "sender URL” and “receiving URL” are not always “starting URL” and “end URL”.
  • the starting host needs to obtain the full URL of the destination host from the domain name service system, compare it with its own complete URL, and extract the necessary network blocks along the way from the data packet. URL.
  • the originating host sets the first required web block URL to "receipt URL". Since the super gateway communicates with each other in the routing table, and the super gateway that receives the data packet, the ingress URL can be either the current recipient address or not; if so, the super gateway is responsible for replacing the recipient URL with the next one. URL (net block URL or destination host URL); if not, no replacement is made, and the data packet is "trial" again in another network block. In order to comply with the IPv4 convention as much as possible, the Super Gateway will replace both the "Receipt URL" and the "Sender URL". The URL required for the above operation is saved in the newly established hyperlink agreement option.
  • the above method will cause a transmission path through different network blocks between the start point and the end point, as shown in FIG. One of them is the shortest path, and the rest is called the alternate transmission path.
  • the L2 network block group starting from an L2 network block in a network autonomous system (ie, a network with unified management: a country), the L2 network block outside the area returns to the area.
  • the alternate path of another L2 network block perhaps it should be eliminated.
  • the data transmission between the subnet blocks does not allow the alternate path through the parent network block.
  • the super gateway can request the modification of the routing table like a normal router to adjust the distribution of data flows between transmission paths.
  • the data packet will be sent to the super gateway with the current receiving address, so that the packet travel will be the shortest; but if the shortest path should be temporarily due to congestion, failure, etc.
  • the super routing system will naturally use the alternate path to offload the data packets.
  • the super-router also has the function of helping to realize the interworking between the new and old networks.
  • the normal host executing the super-linked protocol needs to organize the contents of the super-linked protocol option, including reversing the order in which the URL is saved; the old un-upgraded I Pv4 host "does not know"
  • the hyperlink protocol option so in the response packet sent by the I Pv4 host, the hyperlink protocol option will be sent back to the super gateway on the old network boundary.
  • the Super Gateway will be able to identify such packets and "represent" the I Pv4 host to organize the information in the new options and then send them to the new network host that initiated the access. In this way, users in the new network block will be able to access the un-upgraded old network server with the assistance of the super gateway, thus solving the main part of the new and old network interconnection problems.
  • the individual upgraded old network hosts can enter the full URL of both parties to actively access the new network block, but the un-upgraded host cannot. This will cause the nodes in the old network to upgrade to accept the super-connection agreement. Due to the need for two-way communication, the email service between the new network and the old network will be a problem before the upgrade of the old network's email server.
  • the solution is: Establish a "transfer station" in the old network, and upgrade the super-link protocol of the host of the transfer station.
  • Proxy super gateway A special gateway that is connected in parallel with a super gateway between two network blocks.
  • the proxy super gateway uses one of the specific reserved URLs, and the URLs of the two ports can be the same.
  • the proxy super gateway announces a close proximity to the proxy's super gateway in the routing table and uses the two port URLs of the proxy's super gateway for super-routing operations.
  • the proxy super gateway has the same privileges as the proxyed super gateway.
  • the proxy super gateway also announces to all super gateways that the proxy's super gateway announces.
  • the role of the proxy super gateway is to share the data flow to improve the network speed and improve the reliability of the super route.
  • the super-gateway and proxy super-gateway can perform shutdown maintenance, upgrade, etc. at any time.
  • Routing security stamp is a 32-bit binary integer password, stored in the super-linked protocol option, stamped and cleared by the first and last super-gateway that received the packet. If the packet passes only during the entire trip
  • a super gateway does not need to cover the routing security stamp.
  • the routing security stamp is used to prevent forgery of packets sent from one super gateway to another, and should only be circulated in one network autonomous region.
  • the super gateway on the border should clear the routing security stamp in the outbound packet.
  • the invention does not include details of the routing security stamp.
  • the routing security stamp can also be borrowed by the common routing system inside the new network block to perform security innovation of the routing protocol.
  • Super Gateway Management Protocol is a new protocol established for managing super gateways, similar to the current protocol used to manage routers.
  • the management protocol is an auxiliary agreement established for the security mechanism of the super-linked protocol. Like the routing security stamp, the super-management management protocol should only be circulated in one network autonomous region; the global identification number of the protocol can be unified, but the format and password of the protocol need not be shared.
  • the English translation of the Super Gateway Management Protocol can be referred to as Super Gateway Management Protoco l, abbreviated as SGMP.
  • the invention does not include details of the Super Gateway Management Protocol.
  • Multi-network block sharing of important servers That is, important server hosts appear in different network blocks with multiple ports (network cards) at the same time, and even the same website address can be used in each network block. Such a connection will shorten some important servers, such as the access time delay of the domain name server. However, it must be noted that the server thus connected will actually have the status of a super gateway. In order to prevent the "information smuggling" channel, the server that exists across the network block should be cautious, and the network block to which the port is connected should not exceed the "route security stamp" and the shared scope of SGMP. For example, the server should not be allowed to have ports that connect both the old network and the new network block.
  • Figure 1 shows the relationship between the port URL of the super gateway and the block URL of the two network blocks connected to both sides.
  • the super gateway is connected to the network block A and the network block D.
  • the port address of the super gateway is d, which is the network block address of the network block D.
  • the port address of the super gateway is a, which is the network block address of the network block A.
  • Figure 2 shows a new Internet consisting of four network blocks. As can be seen from Figure 2, all the super gateways connected to each network block use the same URL for the outer port, which is the net block URL of the network block.
  • the network block URL of network block A is a; the network address of network block P is p, and so on.
  • Figure 3 shows the new Internet shown in Figure 2, from!
  • the user in the 3 ⁇ 4 block s transmits the data packet to the server in the network block P.
  • Three of the different transmission paths that can be used The path from the block A directly into the network block P is the shortest path, and the other two are the alternate paths.
  • Figure 4 shows a new Internet consisting of six network blocks.
  • the network blocks, B, E (the network block URLs are a, b, e) are L2 blocks, and C, F, and R are L3 network blocks.
  • G is the subnet block of network block B, and F and R are the subnet blocks of E. This figure is intended to describe two examples of programming hyperlink protocols.
  • the first example of FIG. 4 is to send a data packet from the website s in the network block C to the website d in the network block F.
  • the two complete URLs required to compile the hyperlink protocol options are b : c : s and e : f : d. Incorporated by reference to section 4 of the "Detailed Implementation" below (Rule 20.6)
  • the highest layer s in the URL heap as the IPv4 sender URL, and the third layer b as the IPv4 receiver URL.
  • Table 2 Part of the packet header sent from (b: c : s) to (e: f: d), related to the superroute
  • the second example of Figure 4 is to send a packet from the URL t in the network block E to the website d in the network block F.
  • the two full URLs required to compile a hyperlink agreement are e:t and e:f:d.
  • e:t has only two layers of URLs
  • the same two addresses e on both sides are erased, and the obtained URL is t:e:f:d.
  • e:d Take the highest layer t in the network and the f in the third layer as the sender and receiver URLs of the IPv4.
  • the remaining e:d is the address table in the hyperlink protocol option.
  • Table 3 shows that
  • Table 3 From the (e: t) to the (e: f: d) packet header, the part related to the super 3 ⁇ 4f
  • Figure 5 shows the evolution of the parameters listed in Table 2 in the first example of Figure 4 as they pass through the super-gateway g1, g2, g3 to the end point.
  • the arrow with the polyline in the figure is drawn from the "process pointer" to the indicated URL.
  • the maximum capacity of the new Internet (excluding the network block URL) can reach 17 million times the IPv6 capacity, but this is done with one-third of the old network URL.
  • the network block URL is premised and therefore unrealistic. But even with the restriction of establishing 100 L2 network blocks, the maximum total capacity of the new Internet will still exceed
  • the new Internet in the future will only need to build hundreds of L2 network blocks; the size of the L2 network block group will not be determined by the capacity requirements of the website, but by the needs of self-built network blocks in each network autonomous region. Efforts to maximize the capacity of each subnet block should not be deliberate, and it is not even necessary to establish an L4 block.
  • the L2 network block will form the highest level of the Internet.
  • the data that can be referred to is the current cable capacity between China and the United States and the speed of the routers on the market: the former is 5120 Gb/s, and the latter is CRS-1 produced by Cisco. 40Gb/s per line. The ratio between the two is 128. In the future, communication capacity requirements and hardware speed will be further improved, but the above ratio should not be changed by orders of magnitude. Therefore, in the old network, a C-level network address capacity (256 URLs) can be reserved for each network block URL to connect to the proxy super gateway. In order to cover hundreds of L2 network blocks that may be established in the future, 100,000 URLs should be reserved in the old network as the L2 network block URL and the proxy super-gateway URL that may be needed. This is not difficult at present.
  • proxy hyper-gateway URLs can be reserved for each network block URL within the new network block.
  • the reserved URL of the proxy super gateway should be spread across different subnets. Select the net block URL of the old network
  • a network block URL for the old network This can select one of the L2 network block URLs reserved in the old network.
  • the network block URL cannot be determined by the network block itself, but by other network blocks. Therefore, determining the network address of the old network will not require the consent of the old network. In the future, when the old network is upgraded, it will naturally recognize the common choice of new network blocks.
  • the network block URL of the old network should be given to the port of the super network between the old network and each new network block in the new network block.
  • the old network Since the old network has been shared around the world, it is not a problem to select the network block URL of the new L2 network block.
  • Each new network block can select one for itself in the L2 network block URL reserved in the old network.
  • the net block URL should be assigned to the port of the super gateway between the new network block and the old network in the old network.
  • the software of the user host needs to be upgraded to support the hyperlink protocol.
  • the responsibility of the user host includes the preparation of the hyperlink protocol option and the handling of the hyperlink protocol option in the received packet, which includes two parameters: the order of the reverse URL table, the reset process pointer, and the number of cross-network times.
  • the initial value of "cross-network number" is 0, and the initial value of "process pointer” is 3.
  • the URL table is compiled according to the following rules (see two examples in Figure 4):
  • the number of URLs in the URL table must be an even number. Add 2 to the number of URLs in the URL table to get the maximum progress. (If the end point is in the I Pv6 network block, the lowermost layer of the "URL heap" is the IPv6 URL; at this time, since the last URL is different in length, although the number of URLs is still even, the maximum process will be odd, by the address table. The number of URLs in the file is increased by 5. The URL table is clamped in the I Pv6 network block. Since all the URLs are the same length, the above special processing is not required.) The user host should be able to upgrade itself to communicate with the local network block.
  • the corresponding upgrade software should be prepared for users to download, including at least user operating system patches, application software (such as browser, web server, email, FTP, Tel net, etc.) and domain name service system software that support the super-linked protocol. To do this, you need to do the following tasks:
  • the super gateway needs to have the ability to handle hyperlinks and routing security stamps. If the gateway in the market is not programmable enough, you can first replace it with a Linux workstation and then look for optimized dedicated hardware.
  • the super gateway on the border is the gateway to the network autonomous region, which is equivalent to customs. Therefore, it should also be responsible for intercepting data packets that should not cross the border.
  • Routing security stamp operation If the new recipient URL is the domestic network block URL and the number of cross-networks is 0, the data packet is covered with a security stamp. When the number of cross-networks is greater than 0, if the new recipient URL is a net block URL, check whether the security stamp is valid; if the new recipient URL is not the net block URL, the security stamp is cleared. If the packet goes out of the network (including entering the old network), the security stamp is also cleared.
  • the routing security stamp is actually a password between the super gateways inside the network autonomous area, which is used to ensure the authenticity of the content of the hyperlink protocol option. Unless there is an agreement between the network autonomous regions, the security stamp should be cleared when crossing the border.
  • the Super P3 ⁇ 4 connection between two network blocks belonging to different network autonomous regions will actually be two super gateways of "back to back", independent of each other, each with its own secrets, and generally does not share a security stamp.
  • the routers inside the new network block can use the mature (originally ready to be phased out) I Pv4 products on the market, without having to purchase expensive I Pv6-compatible new products that have not yet been fully tested.
  • the new I Pv6 network block can't be like this. Routing system protocols and software can be copied from the I-day network (the super-link protocol can be improved and innovated after implementation).
  • the domain name service system must be changed, but will not affect the operation of the old network.
  • the domain name service system will need to make two changes: One is the format of the report URL, and the domain name server will need to report the full URL of the target node and its layers.
  • the existing domain name service system in the old network is a central tree structure, the most authoritative of which is "root"
  • each L2 network block although it is theoretically possible to collect and store the name-address resolution data of all other L2 network blocks by the root domain name server, actually maintaining such a database would be a heavy burden. Since the name-address resolution only makes sense when the target network block can be connected, it is not necessary for the root domain name server to store the name-address resolution data of all other L2 network blocks, but should adopt a "de l egate" query. the way. To do this, the root domain name server only needs to be able to find the target L2 network block for the users it manages. The root name servers in each L2 network block are equal to each other and will often need to query each other.
  • the only part that may not be implemented in the future will be the existing root domain name server in the old network. This situation will need to wait until the old network is upgraded. Before the old network is fully upgraded, the root domain name servers in the old network will continue to operate, but their authority will continue to be limited to the old network; they will naturally accept queries from the root domain name servers in the new network block, but Before "self-upgrading", it is "not used to" to query the name and address of the root domain name server in the new network block. Therefore, users on the old network will not be able to query the URL and website of the new network block through the domain name system. In the future, when the number of users in the old network that need to connect to the new network block is large enough, the root domain name server in the old network will be forced to upgrade, which will be the beginning of a comprehensive upgrade of the old network.
  • a network autonomous community has several L2 network blocks, consider sharing them with a set of root name servers (for reliability, the root server is generally more than one): You can let each L2 network block share these root name servers. This way, a central domain name service system can still be formed in the network within the network autonomous region.
  • the internal domain name of a network block should be designed to facilitate the determination of which network block the target belongs to in order to simplify the query and increase the speed.
  • top-level domain names of the new Internet will only be national or national organizations (such as the United Nations), and other institutions that must be registered in a country will be placed under the country's top-level domain. Knowing the national domain name, it is much easier to determine the network block.
  • top-level domain names such as com, edu, gov, org, net, etc.
  • top-level domain names such as com, edu, gov, org, net, etc.
  • an international organization with a member state is using the "org" top-level domain name, it is recommended to migrate it to "i nt" in the new network.
  • the UN website should become www. un. org in the new network. . i nt.
  • the website that usually omits or does not have a national domain name will not be able to obtain access from the new network block if it is not placed under the old country name of the old network. It is recommended that such sites be temporarily placed under the US national domain name "us", and that in the event of an objection, they will be allowed to choose their own country.
  • the network layer protocol, the network address hardware identification code (MAC) conversion protocol ARP and RARP, and the control message protocol I GMP can be copied from the old network.
  • the super gateway should intercept packets that perform management operations across the border: Network management operations should be internal to the network autonomous region, and its instructions must be issued from within the territory.
  • routing system within a network block is similar to a post office. Private ownership is common in European and American countries, but the post office is a state-owned enterprise; this is obviously to ensure the neutrality of postal services. From this point of view, it is not appropriate for a routing system within a network block to be built and maintained by multiple enterprises. In addition, to ensure information security, to prevent the routing system from being hijacked and blocked, the routing system should not be completely Transparent, but should have an internal defense system to prevent foreign invasion.
  • the post office cannot determine where the letter originated, the postmark records the location and time of the letter entering the post office system, which provides the recipient with a set of important basic information; in contrast, the current routing system does not have this The minimum security mechanism.
  • the routing system should form a continuous and seamless system within each network block and should be seamlessly connected to the super gateway.
  • the backbone of the routing system and its entrance should be operated and managed by a unified organization. Even so, it is only possible to guarantee the authenticity of the user's information source within the scope of the network autonomous region.
  • a packet Before a packet is sent to the backbone of the routing system, it should be checked like a letter with a postmark at the entrance to ensure that it belongs to the distribution network. If it contains the ULIP option, it should also check whether the recipient URL is a super gateway.
  • the post office may not intercept messages with forged (or blank) outgoing addresses, because the presence of the postmark will enable the recipient to make a judgment; in the Internet, because the cost of sending is extremely low, it may be considered to discard packets that contain false penalties. .
  • the routing system is improved to implement the above security functions, the purpose of the routing security stamp may be changed; for example, the security stamp may be routed and sent to the user, and the relevant data such as the information source and the routing journey may be transmitted. Due to the limited number of super gateways, the above changes should not be difficult to implement.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Description

一种可以取代 I Pv6的网络层协议 技术领域
本发明涉及的是一种扩大互联网网址空间的方法。
背景技术
互联网所带动的信息革命, 在短短二十余年间显著地改变了世界, 这是最初的网络设 计者们始料未及的。 当初被认为是不可能用完的 32位网址现在居然已经用去大半。 除了 互联网的发源地美国之外, 世界其他各国都将很快面临网址枯竭。
互联网工程任务组(I ETF) 1998年推荐的 I Pv6协议虽然有可能提供巨量网址, 却与现 行的 I Pv4协议完全不兼容。 虽然全球网络专家在过去十年全力推动 I Pv6, 却由于不能与 I Pv4网无缝连接, 至今无法商业实施, 导致互联网发展的一些前沿被 I P网址短缺问题卡 住。 由于互联网发源地美国存有大量储备网址(截至 201 1年 5月, 美国获得的网址数为总 分配额的 45%, 人均 5个网址), 估计靠 I Pv4可撑 10年以上, 在此期间 I Pv6将举步維艰。
在 I ETF的技术文件 "征求意见公告 1 385号" (RFC1385 , 1992年)中曾经提出一种试 验性方案来扩大互联网的网址空间,其中首次提出了以划分多个具有 32位(二进制;下同) 网址的网络区域的方式来扩大网址空间, 并采取"向上方加号"的方法在全网内区分不同 的网络区域。 RFC1385所设想的是以 "就地分裂自治" 的方式扩大互联网: 只有地域毗邻 的网络区域之间通过网关互有连接, 网络区域之间的网关利用新加上的 "网络号" (未指 明是否 32位)进行网址翻译, 从而实现不同网络区域之间的寻址。 RFG1385中另一个有价 值的提议是, 将"网络号"储存在一个新设立的 I Pv4选项中进行传送,以避免破坏现行互 联网的协议框架。
中国专利申请 200710073741 (2007年)中也提出建立具有 32位网址的 "元地址区", 又称 "元 I P地址区", 并采取多层 32位网址堆叠的方法来区分不同的 "元地址区" (这相 当于在 RFC1385的基础上, "向上方加入多层 32位网络号")。 可惜的是, 该方案的其余 部分不尽合理, 例如没有利用 I Pv4选项, 而是要求改变传输层协议的识別号, 这就破坏 了与现行互联网的兼容性。
本发明以 RFC1385和中国专利申请 200710073741 中的合理部分作为现有技术, 继承 其中划分网络区域、 利用 I Pv4新选项储存新增加的信息、 和网址堆叠的方法, 寻求在不 破坏与现行互联网的兼容性、并尽可能少作改动的前提下扩大网址空间。不改变现有的寻 址模式, 就意味着继续使用 32位用户网址, 以达到继续使用现有 I Pv4硬件、 继续使用现 有的路由系统、 和尽可能少改变网络软件的目的。 这就要求以现有互联网为 "模板 "新建 具有 32位网址的网络单位。 我们将此网络单位重新取名 "网块",一是因为 RFG1385中没 有为网络区域取名; 二是因为, 就网络中的地位而言, 本发明中的 "网块"与中国专利申 请 200710073741 中的 "元地址区" 有本质上的不同: 后者从属于现有互联网, 以现有互
1
确认本 联网为最高层的 "树根", 仍然是一个有中心的树型互联网; 而在本发明中, 互联网的最 高层将由彼此平等的网块构成, 既非树状结构、 也没有中心。
本发明是一种网络层通讯协议, 可称为超联协议, 其英文标准译名可使用 Ultra Link Internet Protocol, 缩写为 ULIP. 其余主要关键词的英文译法建议:
网块 = Network Block, 简称 Block;
网块网址 = Block Address;
子网块 = Child Block;
网块群 = Block Group;
超级网关 = Super Gateway;
代理超级网关 = Proxy Super Gateway;
完全网址 = Complete Network Address, ;
超级路由 = Super Routing;
路由安全戳 = Routing Security Stamp;
起点 = Start Point;
终点 = Endpoint;
另外, "发方网址"系指 I Pv4的 " Source Address "; "收方网址 "系指 I Pv4的 " Destination Address" .
超联协议在形式上是现行 I Pv4 协议的合法扩充, 不要求改变网络层以外的协议。 本 发明采用新的方法连接网块。 除了网块、 I Pv4 新选项、 和网址堆叠三项外, 本发明与 RFC1385和中国专利申请 200710073741 没有其它共同之处。
发明内容
本发明的要点在于实现网块之间有效连接的下列概念和方法:
(1 ) 网块网址: 为每个网块选择一个 32位网址作为进入该网块的入口, 称为网块网址。
数据包从其它网块进入某网块之前, 最后经过的一定是该网块的网块网址。 作为对 比, 在中国专利申请 200710073741 中, 从不同的方向进入 "元地址区"需要经过不 同的网址; 而 RFC1385采用比较昂贵的网址翻译, 并不将 "网络号"赋予任何网关。 由于网块网址是从其它网块进入本网块的入口, 网块网址原则上不能由网块本身决 定, 而是由其它网块决定的。 起点、 终点处于同一网块的数据包, 在传送途中将不 经过网块网址。
(2) 超级网关: 为连接在任何两个网块之间的特殊网关。 每个超级网关的两个端口必须 使用互不相同的网址, 它们分别是超级网关另一側的网块网址, 如图 1 所示。 作为 对比, 中国专利申请 200710073741 中的 "元网关 "其两个端口使用相同的网址, 因 而与超级网关有本质的区别。
超级网关的上述连接方式是确立网块网址和建立超级路由的基础。
(3) 网块群: 每一对网块都有超级网关相连的一组网块, 称为网块群。 网块群内的网块 互为同群网块。 同群网块彼此平等, 因此新互联网将没有中心。
(4) 子网块、 母网块、 完全网址: 网块下方可以连接子网块, 上方的网块则称为母网块。
新互联网中的两个网块可以有相同的网块网址,但这时它们的母网块、祖母网块、… 等诸层网块网址当中必然至少有一个不同, 由此唯一地确定网块的位置。
借用中国专利申请 200710073741 中的 "网址堆叠" 方法, 从用户网址开始, 依次向 上堆叠本网块网址, 以及母网块网址、 祖母网块网址等等(如果存在的话), 就形成 用户的完全网址。 完全网址在全网范围唯一。
一个完全网址含有若干层网址。 如果一个网块内部的用户完全网址是 2层的, 则该 网块称为双层网址网块, 简称 L2 网块; 依此类推可以有 L3、 L4、 L5 网块。 不存在 "单层网址网块": L2网块没有母网块。 一个 L3网块一定是某个 L2网块的子网块, 一个 L4网块则必定是某个 L3网块的子网块, 依此类推。
在不突破 I Pv4数据包格式的前提下, 按照超联协议扩大后, 互联网的最大容量(不 包括网块网址)约为 I Pv6的 1700万倍。
现有互联网将自动成为 L2 网块之一, 可简称 "旧网"。 预计旧网将是唯一的多国共 享 L2网块。 L2网块群将成为新互联网的最高层结构。
(5) 姊妹网块群: 简称姊妹群。 在某些网块之间通讯需求不多的情况下, 同层、 同母的 网块可以形成多于一个网块群; 这些同层、 同母的网块群之间仅有(但必须有)有限 的连接, 称为姊妹网块群。 各姊妹群中的所有网块具有相同的网址层数和共同的母 网块(L2姊妹群则没有母网块)。 建立姊妹群可以节省超级网关。
相继建立的 L2 网块, 或者形成单一的 L2 网块群, 或者按照地域、 通讯需求等关系 形成若干个互相连接的 L2姊妹群, 但预计旧网将是所有 L2姊妹群的共有成员。
(6) 超联协议选项: 这是为超联协议建立的 I Pv4新选项, 用来存储数据包在传送途中必 须经过的网块网址、 起点或终点的主机网址、 以及其它信息, 如下面表格所示。 此 选项的引入, 使超联协议得以避免破坏现行的网络协议体系, 并使新建网块中的用 户可以访问旧网中未升级的网站。此选项的创意来自 RFC1385, 本发明按照超级路由 的橾作和安全需要设计了具体格式。 UL I P识別码可选 154.
表 1 : 超联协议选项的一般格式 (最大进程 =选项长度÷4)
Figure imgf000005_0001
其中"选项长度"的 8位数,在超联协议选项中最低 2位永远为 0,其余 6位作为 "最 大进程"使用。 4位的 "跨网次数"用于记载数据包跨越网块边界的次数, 初值为 0 ; 跨网次数达到最大而仍未到达终点网块的数据包将被丟弃。 4位的 "进程指针"指向 超级路由所需要的下一组代换网址, 初值为 3, 并在数据包到达终点网块时达到 "最 大进程"。 另外有 8位数留作将来创新之用。
超联协议选项在 IPv4网中通行的根据, 在于现行互联网协议关于陌生 IP选项的处 理规定: 主机不处理(包括不丢弃)收到的陌生 IP 选项(参见互联网标准文件 RFC1122, 第 3.2.1.8节)。
基于上述理由, 还可以将超联协议选项移植用于其它目的, 例如建立巨型私网。 又 例如 IPv6与 IPv4无缝连接的一个主要困难, 是将 IPv6数据包 "翻译" 成 IPv4数 据包后, IPv6网址在数据包中 "无处藏身"。 如果将 IPv6网址储存在超联协议选项 之内, 则在 IPv4与 IPv6交界处的网关的协助下, IPv6主机就可以完成对 IPv4网站 的访问。这说明具有不同网址长度的网块也可以通过超联协议整合在一个网络之下, 换句话说, 在超联协议之下, IPv6网应该以网块的形式加入互联网, 并且也可以建 立多个 IPv6网块。但有两项限制是必须的: 第一, IPv6网块的网块网址应该是 IPv4 网址之一,从而使所有网块网址可以 32位和 128位格式之间一对一自由变换;第二, 所有的 IPv6网块必须是 L2网块。 这一方面是由于 IPv6网块容量巨大, 没有必要建 立更深层的 IPv6网块; 另一方面是为了保持超联协议选项在 IPv4网块中至多含有 一个 IPv6网址, 以避免出现歧义。 当数据包进入 IPv6网块时, 超級网关将包头中 的所有 IPv4网址唯一地映射成 IPv6网址。 超联协议选项将像其它 IPv4选项一样, 自然地移植到 IPv6协议中。
超级路由: 超级网关集体协作将数据包送往目标网块的过程。
由于兼容性, 在同一个网块内传递的数据包将完全可以通过现行协议体系实现; 只 有跨网块传递的数据包才需要超级路由。
从普通路由系统角度看来, 超级网关像是普通主机和路由器的混合体。 网块上的超 级网关, 除了连接母网块和子网块的以外, 都在网块内部的路由表中宣布彼此相通。 当存在姊妹群时, 跨姊妹群的超級网关还应当宣布通往另一側网块群中所有的网块。 为避免混淆, 我们称始发数据包的主机为该数据包的 "起点" 主机; 最终接收数据 包的主机为该数据包的 "终点"主机; 而 IPv4数据包头中原有的 "Source Address" 和" Destination Address" 则分别译为 "发方网址" 和 "收方网址"。 在跨网块传 送的数据包中, "发方网址"和 "收方网址" 并不总是 "起点网址"和 "终点网址"。 当数据包需要从起点主机送往另一个网块内的终点主机时, 起点主机需要向域名服 务系统索取终点主机的完全网址, 与自身的完全网址比较, 从中提取数据包沿途必 经的网块网址。
起点主机将第一个必经的网块网址设为 "收方网址"。 由于超级网关在路由表中彼此 相通, 收到数据包的超级网关, 其入口网址既可以是当前收方网址, 也可以不是; 如果是, 则超级网关负责将收方网址置换为下一个必经的网址(网块网址或终点主机 网址); 如果不是, 则不作置换, 让数据包在另一网块中再次 "试投"。 为了尽可能 符合 IPv4的惯例, 超级网关将同时置换 "收方网址"和 "发方网址"。 上述操作所 需网址保存在新设立的超联协议选项内。 当网块群含有两个以上网块时, 上述方法将在起点和终点之间造成经过不同网块的 传输路径, 如图 3。其中一条是最短路径, 其余称为备用传输路径。 网块群越大, 则 备用传输路径越多; 如此, 网块之间的传递过程将类似于普通路由, 但却不能由普 通路由系统执行, 因此称为超级路由。
出于安全考虑, 可以从超级路由中除去某些备用路径。 例如在 L2网块群中, 从一个 网络自治区(Autonomous System, 即具有统一管理的一片网络: 也可以是一个国家) 内的某 L2网块出发、 经过区外的 L2网块回到该区内另一个 L2网块的备用路径, 或 许就应当在剔除之列。 另外, 为防止造成结构性数据流拥挤, 子网块之间的数据传 输不容许有经过母网块的备用路径。
根据数据流量和自身状态, 超级网关可以像普通路由器一样请求修改路由表, 从而 调节数据流在各传输路径之间的分配。 当网块和超级网关都工作正常时, 数据包将 被送到拥有当前收方网址的超級网关, 这样数据包的行程将最短; 但如果本应该最 短的路径由于拥塞、 故障等原因而暂时不是最短时, 由于超级网关对路由表的调节, 超级路由系统将自然地利用备用路径对数据包进行分流。
此外, 超级路由还具有协助实现新、 旧网之间互通的功能。 在对收到的数据包进行 回应之前, 执行超联协议的正常主机需要整理超联协议选項中的内容, 包括反转其 中保存网址的顺序;旧 中未升级的 I Pv4主机"不认识 "超联协议选项,因此在 I Pv4 主机发出的回应数据包中, 超联协议选项将被照抄发回到旧网边界上的超级网关。 超级网关将可以识别这类数据包, 并 "代表" I Pv4主机整理新选项内的信息, 然后 送到发起访问的新网主机。 如此, 新网块中的用户将可以在超级网关协助下访问未 升级的旧网服务器, 从而解决新、 旧网互通难题的主要部分。
在旧网全面升级之前, 个别升级的旧网主机可以自行输入双方的完全网址主动访问 新网块, 但未升级的主机则不能, 这将促使旧网中的节点升级接受超联协议。 由于需要双向通讯, 新网块与旧网之间的电子邮件服务在旧网的电邮服务器升级之 前将是一个问题。 解决办法是: 在旧网内建立 "转递站", 并对转递站的主机进行超 联协议升级。
(8) 代理超级网关: 与某个超级网关平行连接于两个网块之间的特殊网关。 代理超级网 关使用特定的预留网址之一, 其两个端口的网址可以相同。 代理超级网关在路由表 中宣布近距离通往被代理的超级网关, 并使用被代理的超级网关的两个端口网址进 行超级路由操作。 除了不使用自己的网址之外, 代理超级网关与被代理的超级网关 享有享有相同的权限。 另外, 代理超级网关也宣布通往被代理的超级网关宣布通往 的所有超级网关。
代理超级网关的作用, 在于分担数据流以提高网速, 和提高超级路由的可靠性。 在 有代理超级网关存在的情况下, 超级网关和代理超级网关可以随时进行关机维护、 升级等橾作。
(9) 路由安全戳: 是一个 32位二进制整数密码, 保存在超联协议选项中, 由第一个和最 后一个收到数据包的超级网关分別加盖和清除。 如果数据包在整个行程当中只经过
援引加入 (细则 20.6) 一个超级网关, 则不需要盖路由安全戳。 路由安全戳用来防止伪造从一个超级网关 发往另一个超级网关的数据包, 应该仅流通于一个网络自治区境内。 边境上的超级 网关应当将出境数据包中的路由安全戳清除。 本发明不包括路由安全戳的细节。 路由安全戳还可以被新网块内部的普通路由系统所借用,进行路由协议的安全创新。
(10)超级网关管理协议: 是一个为管理超级网关而设立的新协议, 类似于现行的、 用来 管理路由器的协议。 超级 Is)关管理协议是为超联协议的安全机制设立的辅助协议。 像路由安全戳一样, 超级冃关管理协议应当仅流通于一个网络自治区内; 全球可以 统一该协议的识别号, 但不必共享协议的格式和密码。 超级网关管理协议的英文译 名可采用 Super Gateway Management Protoco l ,缩写为 SGMP。 本发明不包括超级网 关管理协议的细节。
(11 )重要服务器的多网块共享: 即重要的服务器主机以多个端口(网卡)同时出现于不同 网块内, 甚至可以在各网块中使用同一个网址。 这样的连接将缩短一些重要的服务 器, 譬如域名服务器的访问时间延迟。 但必须注意的是, 这样连接的服务器将实际 上具有超級网关的地位。 为了防止出现 "信息走私"通道, 设立跨网块存在的服务 器应当谨慎, 其端口所连接的网块应当不超出 "路由安全戳 "和 SGMP的共享范围。 例如, 不应当容许服务器同时具有连接旧网和新网块的端口。
(12)上述概念和方法并不需要锁定于 I Pv4的具体细节。 例如, 即使 I P冋址长度不是 32 位,或者数据包格式与 I Pv4有所不同,仍然可以通过本发明的概念和方法将网络 "复 制"成 R块, 上述其它概念也仍然可以运用。 通过移植超联协议, 可以将任何有中 心的网络扩大成为对应于 "多极世界"的无中心网络。 附图说明
图 1表示超级网关的端口网址和两側所连接的两个网块的问块网址之间的关系。其中的超 级网关连接网块 A和网块 D。 在网块 A中, 该超级网关的端口网址为 d, 此即网块 D的网 块网址; 在网块 D中, 该超级网关的端口 址为 a , 此即网块 A的网块网址。
图 2表示一个由 4个网块组成的新互联网。从图 2中可以看出,连接在每个网块上的所有 超级网关, 其外側端口使用相同的网址, 此即该网块的网块网址。例如网块 A的网块网址 为 a ; 网块 P的冋块网址为 p, 等等。
图 3表示在图 2所示的新互联网中,从! ¾块八中的用户 s传送数据包到网块 P中的服务器 d可以使用的不同传输路径当中的三个。其中从冋块 A直接进入网块 P的路径是最短路径, 另外两条是备用路径。
图 4表示由 6个网块组成的新互联网。 其中网块 、 B、 E (网块网址分别为 a,b, e) 为 L2 冋块, C, F, R为 L3网块。 其中 G是网块 B的子网块, F和 R是 E的子网块。 本图意在描 述两个编制超联协议选项的例子。
图 4的第一个例子是从网块 C中的网址 s向网块 F中网址 d发送数据包。编制超联协议选 項所需的两个完全网址分别为 b : c : s 和 e : f : d。 根据下文 "具体实施方式" 第 4小节的 援引加入 (细则 20.6) 规则, 由于 b≠e, 规则的第 (2) 步没有擦掉任何网址: 因此第 (3) 步获得的网址堆为 s:c:b:e:f:d。 取此网址堆中最高层的 s作为 IPv4发方网址, 第 3层的 b做为 IPv4收方 网址, 剩下的 c:e:f:d即为超联协议选項中的网址表。 因此最大进程为 4+2=6。
表 2: 从 (b: c : s)发往(e: f: d)的数据包头中, 与超级路由相关的部分
Figure imgf000009_0001
图 4的第二个例子是从网块 E中的网址 t向网块 F中网址 d发送数据包。编制超联协议选 项所需的两个完全网址分别为 e:t和 e:f:d。 根据规则, 由于 e:t只有两层网址, 两边 相同的两个 址 e被擦掉一个, 获得的网址堆为 t:e:f:d。取此网扯堆中最高层的 t和第 3层的 f 作为 IPv4的发方网址和收方网址, 所余 e:d即为超联协议选项中的网址表, 最 大进程为 2+2=4。 如表 3:
表 3: 从 (e: t)发往 (e: f: d)的数据包头中, 与超级 ¾f由相关的部分
Figure imgf000009_0002
图 5表示在图 4的第一个例子中, 表 2中所列参数在经过超级网关 g1、 g2、 g3传送到终 点过程中的演变。 图中带折线的箭头从 "进程指针"划到所指示的网址。 具体实施方式
1. 在旧网中预留网块网址和代理超级网关网址
实施之前, 首先应该考虑新互联网可能达到的规模, 作出相应的规划。 在不突破 IPv4 数据包格式的前提下, 实施超联协议后, 新互联网的最大容量(不含网块网址)可达 IPv6容量的 1700万倍, 但这以三分之一的旧网网址作为网块网址为前提, 因而是不 现实的。 但即使以建立 100个 L2 网块为限制条件, 新互联网的最大总容量仍将超过
援引加入 (细则 20.6) I Pv6的容量。 再退一步, 即使不建立 L5网块, 则在 1 个 L2网块之下的 L3、 L4网块 的最大总容量为 2. 9 X 10"。 这些网址如果均句地分配在地球表面(包括极地和海洋), 则每平方米可得 5万多亿个网址。 应该说人类用不完哪怕接近这个数量的网址。 因此 应该用不着建立太多的 L2网块。
另一方面, 我们认为人类对信息的需求不可能超出自身有时间处理和理解的容量, 因 此对网址的需求应该不会无限扩张。即使全球人口达到 200亿,平均每人 1000个网址, 也仅够装满 5000个网块; 对于像中国这样的国家, 5亿家庭各分配 256个网址约需 30 个网块; 16亿人口每人 1000个网址需要不到 400个网块。
因此, 预计未来的新互联网将仅仅需要建立数百个左右的 L2 网块; L2 网块群的规模 将不是决定于网址容量需求, 而是由于各网络自治区自建网块的需要。 应当不需要刻 意作出各层子网块容量最大化的努力, 甚至不一定需要建立 L4 网块。 L2 网块将形成 互联网的最高层。
关于代理超级网关的需求极限的估计, 可以参照的数据是当前中国与美国之间的光缆 容量和市场上在售路由器的速度: 前者为 5120Gb/s, 后者如思科生产的 CRS-1 , 为每 线 40Gb/s。 两者比值为 128。 将来通讯容量需求和硬件速度将会进一步提高, 但上面 的比值应当不会有数量级上的变化。 因此在旧网中, 可以为每个网块网址预留一个 C 级分网的网址容量(256 个网址)以连接代理超级网关。 为了涵盖未来可能建立的数百 个 L2网块,应当在旧网中预留 10万个左右的网址作为 L2网块网址和可能需要的代理 超级网关网址, 这在目前不难做到。
考虑到未来区域内通讯需求极有可能超过国际通讯需求, 可以在新网块内为每个网块 网址保留更多代理超级网关的网址。
为了实现数据包分流的目的, 代理超级网关的保留网址应该分散在不同的分网中。 选定旧网的网块网址
为旧网确定一个网块网址: 这可以在旧网中预留的 L2网块网址当中选择一个。前面说 过, 网块网址原则上不能由网块本身决定, 而是由其它网块决定的, 因此确定旧网的 网块网址将不需要征得旧网的同意。将来旧网升级时将自然承认新建网块的共同选择。 旧网的网块网址应该赋予旧网与每个新网块之间的超级网关在新网块中的端口。
选定新 L2网块的网块网址
由于旧网已经有全世界共享, 选择新 L2网块的网块网址也不是问题。每个新网块可以 在旧网中预留的 L2网块网址当中为自己选择一个。网块网址应该賦予该新网块与旧网 之间的超级网关在旧网中的端口。
用户主机
用户主机的软件需要升级以支持超联协议。 用户主机的责任, 包括编制超联协议选项 和处理收到数据包中的超联协议选项, 后者包括颠倒网址表的次序、 重置进程指针和 跨网次数两个参数。 超联协议选项中, "跨网次数"初值为 0, "进程指针"初值为 3。 网址表按下面的规则 编制(参见附图 4中的两个例子):
(1 ) 将起点和终点的完全网址并排放置, 最高层在上, 并使最高层网块网址对齐;
(2) 比较各层的两个网址: 从最高层开始向下进行。
• 若两者不同, 则停止比较;
• 若两者相同且两边现有网址数均为 2, 则将该层擦掉并停止比较;
• 若两者相同且其中一边现有 2个网址,则擦掉该层两个网址之一,并停止比较:
• 若两者相同且两边现有网址数均大于 2, 则将该层擦掉并继续比较下一层;
(3) 将起点完全网址一边剩下的部分倒转次序(使最低层的用户网址在上),并整个移 置于终点完全 R址的剩余部分之上, 形成一个 "网址堆";
(4) 若此"网址堆 "中只剩两个网址, 则属于网块内部通讯, 两个网址分別用作发方 网址和收方网址。这时可以选择省略超联协议选项;如果需要借用路由安全戳进 行创新来强化网块内的路由系统, 则可以编制空的超联协议选项,并将最大进程 和进程指针均设为 2;
(5) 如果是两个冋块之间的通讯, 则 "网址堆 "中将至少有 4个网址。 这时从"网址 堆"的最高层开始,将第一个网址移走作为初始发方网址, 第三个网址移走作为 初始收方网址, 其余网址存入超联协议选项中作为 "网址表"。
网址表中的网址个数必定是偶数。 将网址表中的网址个数加 2, 即得到最大进程。 (若 终点在 I Pv6网块中, 则 "网址堆"最下层的一个是 IPv6网址; 这时由于最后一个网 址长度不同, 虽然网址个数仍是偶数, 但最大进程将是奇数, 由网址表中的网址个数 加 5得到。在 I Pv6网块中縞制网址表,由于所有网址长度相同,不需要上述特殊处理。) 用户主机应当能够自行升级, 以实现与当地网块以外进行通讯。 为此应当准备好相应 的升级软件供用户下载, 至少包括支持超联协议的用户操作系统补丁、应用软件 (如浏 览器、 网站服务器、 电邮、 FTP、 Tel net 等) 、 和域名服务系统软件。 为此需要做下 列各项工作:
(1 ) 为超联协议定义一组新的应用编程界面(API)。
(2)选择合适的 L inux版本和配套的网站服务器(如 Apache)、域名服务器, 按照新定 义的 API进行超联协议升级。
(3) 按照新定义的 AP I编写 WINS0GK替代函数包, 作为微软视窗的补丁。 同时提请微 软升级支持超联协议。
(4) 选择适当的浏览器和电邮服务等软件, 升级支持超联协议。
超级网关
超级网关需要具有处理超联协议选项和路由安全戳等功能。 如果市场上的网关可编程 性不够用, 可以先用 L inux工作站代替, 然后再寻求优化的专用硬件。
援引加入 (细则 20.6) 超级网关的责任包括:
(1 ) 对超联协议选项中的信息进行安全检查(包括安全戳检查), 丟弃不合格或过期的 的数据包。 并在必要时通过报文控制协议通知数据包的发方。 边境上的超级网关 为网络自治区之门户, 相当于海关, 因此还应当负责拦截不应该越境的数据包。
(2) "补偿超级路由"搡作(仅适用于连接旧网的超级网关): 来自旧网的数据包, 若 其中进程指针等于最大进程、且旧网的网块网址出现在网址表的底层而不是顶层, 则可以认定是来自未升级的 I Pv4主机的回应数据包。超级网关需要颠倒网址表的 次序、 重置进程指针和跨网次数。
(3) 在收方网址与自己的入口网址相等时执行网址代换: 让发方网址与进程指针处的 网址对换,将进程指针加 1 ,再让收方网址与进程指针处的网址对换。参见图 5. (若 超级网关入口为 IPv6而出口为 I Pv4, 则在协议翻译之后只有发方网址是 IPv6网 址: 在发方网址对换前, 需要移动网址, 在网址表内为此 IPv6网址留出空间; 发 方网址对换是 "取 4存 6", 进程指针需要加 4。 收方网址对换后, 进程指针照常 加 1.若超级网关入口为 I Pv4而出口为 I Pv6, 则网址对换之前的网址表中, 应与 收方网址对换的最后一个网址为 I Pv6网址: 协议翻译则在网址对换前进行。 )
(4) 路由安全戳操作: 若新的收方网址是境内的网块网址, 且跨网次数为 0, 则为数 据包盖安全戳。 当跨网次数大于 0时, 若新的收方网址是网块网址, 则检查安全 戳是否有效; 若新的收方网址不是网块网址, 则清除安全戳。 若数据包通过本网 块出境(包括进入旧网), 也要清除安全戳。
(5) 处理完数据包之后, 在转发之前将跨网次数加 1。
(6) 接受和处理超级问关管理协议传来的数据、 指令; 执行规定的操作。
路由安全戳实际上是网络自治区内部超级网关之间的密码, 用于保证超联协议选项的 内容的真实性。 除非网络自治区之间有协议, 否则安全戳应该在越境时清除。 连接在 分属不同网络自治区的两个网块之间的超级 P¾关实际上将是 "背对背"的两个超级网 关, 互相独立, 各有各的秘密, 一般不共用安全戳。
路由器和路由系统
新网块内部的路由器可以采用市场上成熟的(原来准备淘汰的) I Pv4产品, 而不必购买 昂贵、 尚未经过充分实用检验的兼容 I Pv6的新产品。 当然, 新建 I Pv6网块不能如此。 路由系统协议和软件可以从 I日网照抄(超联协议实施之后可以改进和创新)。
域名服务系统
域名服务系统必须改变,但将不影响旧网的运行。域名服务系统将需要进行两项改变: 其一是报告网址的格式,域名服务器将需要报告目标节点的完全网址及其层数。
另一项改变是增加根域名服务器的主动查询功能, 具体实施并不复杂: 只需要根域名 服务器把自己暂时当成一个下层服务器向对方查询即可, 可借用现有的协议来实现。 旧网中现有的域名服务系统是一个有中心的树型结构, 其中具有最高权威的是 "根
10
援引加入 (细则 20.6) (root) "域名服务器。 由于新互联网没有中心, 因此将不可能建立和維持一个全球统 一的域名服务器"树"。但每个网块内部将可以照抄现有的名一址解析体系, 建立自己 的域名服务系统, 每个网块内的域名服务器将形成一个树型结构。 如果需要, 每个 L2 网块都可以拥有自己的根域名服务器。
在每个 L2网块中, 虽然在理论上可以由根域名服务器收集并储存所有其它 L2网块的 名一址解析数据, 但实际上维持这样的数据库将是沉重的负担。 由于名一址解析只在 目标网块可以接通时才有意义,因此没有必要让根域名服务器储存所有其它 L2网块的 名一址解析数据, 而应该采取 "委托(de l egate) " 查询方式。 为此, 根域名服务器只 需要有能力为自己所管辖的用户找到目标 L2网块即可。 各 L2网块中的根域名服务器 彼此平等, 将经常需要互相查询。
唯一暂时不一定能实施的部分将是旧网中现有的根域名服务器, 这种情况需要等到旧 网升级时方能改变。 在旧网全面升级之前, 旧网中的根域名服务器将可以继续运行, 只是其权威性将继续限于旧网之内; 它们将自然地接受来自新网块中的根域名服务器 的查询, 但在自身升级之前却 "不习惯" 向新网块中的根域名服务器查询名一址解析 数据。 因此, 旧网中的用户将不能通过域名系统查询新网块的网址和网站。 将来当旧 网中需要连接新网块的用户数量足够大时, 旧网中的根域名服务器将被迫升级, 这将 是旧网全面升级的开端。
如果一个网络自治区拥有若干个 L2网块, 可以考虑让它们共享一組根域名服务器(为 了可靠性, 根服务器一般多于一个): 可以让各 L2网块共享这些根域名服务器。 这样 仍然可以在网络自治区之内的网络中形成有中心的域名服务系统。
由于跨网块的名一址查询采用委托方式, 第二层域名之下的域名无须全球统一, 各网 块将自然拥有按照自己的需要创造内部域名的自由。一个网块的内部域名, 其结构在 设计上应该方便确定目标属于哪一个网块, 以简化查询、 提高速度。
经过协商, 多个网络自治区也可以共同设立一个机构统一管理网块。 在这种情况下也 可能形成一个大面积共享的、 有中心的域名服务系统。 因此在超联协议之下, 虽然互 联网是天然的无中心结构, 但在自愿加盟的基础上仍然可以形成中心。
域名体系
由于建立新网块如此容易, 很难设想将来会有两个国家愿意共享一个网块。 因此, 预 计新互联网的顶层域名将只能是国家、 或以国家为成员的国际组织(譬如联合国), 其 它必须在某国注册才能存在的机构将只能置于该国的顶层域名之下。 知道国家域名, 确定网块就容易多了。 而当前旧网中除上述之外的其它所谓顶层域名(如 com, edu, gov, org, net等)如果跨网块使用, 就需要查表以确定某一机构属于哪个国家, 这将 使域名服务系统的工作量成倍增加, 有害无益, 因此这些域名应该降格, 而不应该成 为新网块中的顶层域名。 在旧网中, 目前的域名可以继续使用, 只是其中需要降格的 顶层域名将不为新网块所承认, 因而只限于在旧网内使用。 由于网块内部的查询可以 省略顶层域名, 上述降格过程将是自然的, 不会影响旧网的运行。 对于旧网中已经存 在的国家域名, 比较简单的处理办法是让它们继续存在, 以代表各国在旧网内已经建 立的网络; 在新网块中另取新的国家域名。
若有以国家为成员的国际组织正在使用 "org"顶层域名, 建议在新网块中将其迁移到 " i nt" 之下, 例如联合国的网站在新网块中应该成为 www. un. org. i nt。 而旧网中其 余通常省略或没有国家域名的网站, 在新网块中如果不置于某个旧网原有的国家域名 之下, 将不能获得来自新网块的访问。 建议暂时将这类网站置于美国的现有国家域名 "us "之下, 将来在发生异议时让其自行选择国家归属。
除 I P协议之外的网络层协议, 网址一网卡硬件识别码(MAC)转换协议 ARP和 RARP, 和 控制报文协议 I GMP等等都可以从旧网照抄。超级网关应当拦截跨境执行管理操作的数 据包: 网络管理操作应当是网络自治区的内部事务, 其指令必须从境内发出。
网块内部的创新: 超联协议蘊含的有界结构将打破互联网的 "无边界"现状。 新建网 块中,一切从头开始, 这是合理规划网络结构、 强化网络安全、 自主创新的难得机会: 新建网块中的通讯,容易实现"内外有別 "。譬如传输层和更高层的互联网协议(如 TCP、 UDP、 HTML等), 与网块外部通讯时可沿用现有协议, 但在网块内部则不必受此限制。 为了增进新网块的安全和效率, 可以对现有协议进行改进, 或在改进不能解决问题时 创建新的替代协议。 改进或新设计的协议, 可先在一个网块内试验, 然后在网络自治 区的所有新网块内部通行。 只要新协议更安全、 更合理, 被普遍接受就只是时间问题。 另一个值得解决的问题是增强网块内部路由系统的安全。 从功能看, 网块内的路由系 统类似于邮局。 欧美诸国普遍实行私有制, 但邮局却都是国营企业; 这显然是为了保 证邮递服务的中立性。 从这一点看来, 一个网块内的路由系统由多个企业来搭建和共 同维持是不恰当的; 另外, 为保证信息安全, 为防止路由系统被劫持和堵塞, 路由系 统也不应该是完全透明的, 而是应该设有一个内部防御体系防止外来入侵。
虽然邮局不能确定信件源自何处, 但邮戳记载了信件进入邮局系统的地点和时间, 这 为收信人提供了一組重要的基本信息; 相比之下, 现行的路由系统还不具备这一最起 码的安全机制。 要保证数据包传递的安全性和网络信息的可信性, 路由系统应该在每 个网块内形成一个连续无缝的系统, 应该与超级网关无缝连接。 实际上在整个网络自 治区内, 路由系统的主干及其入口都应该由一个统一的机构进行运营和管理。 但即便 如此, 也最多只可能在本网络自治区范围内保证用户信息来源的真实性。
一个数据包被送进路由系统的主干之前, 应该像信件盖邮戳一样, 在入口处核对发方 网址, 确保其属于本分网; 如果其中含有 ULIP 选项, 则还应该检查收方网址是否超 级网关之一, 并给合法数据包加盖路由安全戳(对于跨网块数据包, 可直接使用 ULIP 选项中的安全戳空间; 对于网块内部通讯, 可使用最大进程 =2的 ULIP选项来承载路 由安全戳), 注明数据包进入路由系统的位置, 并表明此数据包已经通过检查。 邮局未 必拦截有伪造(或空白)发信地址的邮件, 因为邮戳的存在将使收信者能够做出判断; 在互联网中, 由于发送的代价极低, 因此可以考虑丢弃发现含有伪造成分的数据包。 如果路由系统改进后可以实现上述安全功能, 则路由安全戳的用途可以改变; 例如可 以使用明码路由安全戳并且发给用户, 传递信息源和路由旅程等相关数据。 由于超级 网关数量有限, 上述改变应该不难实行。

Claims

本发明是一种新的网络层协议, 以现存互联网为模板建立同样大小的网块, 不改变 I Pv4的框架以及其它各层网络协议,在现行的 I Pv4协议中加入一个新的选项以储存 网块寻址所需的信息,通过增加网块的方式扩大互联网网址空间。该新协议的特征为: 给每个网块赋予一个确定的入口网址, 称为网块网址; 网块通过特殊网关(称为超级 网关)相连, 跨网块传递的数据包通过超级网关之间的协作(称为超级路由)送往目标 网块。按照该协议新建的网块可以与现有互联网单向无缝连接, 新网块中的用户将可 以访问现存互联网中未升级的的网站, 从而避免 I Pv6难以启动的困难。 该新协议可 称为超联协议, 由下列概念和方法构成:
(1 ) 网块网址: 网块的入口网址。 数据包从其它网块进入某网块之前, 最后经过的一 定是该网块的网块网址; 起点、 终点处于同一网块中的数据包, 在传递途中则不 经过网块网址。
(2) 超级网关:连接两个网块的特殊网关。每个超级网关的两个端口拥有不同的网址, 分别是另一側的网块网址。
(3) 网块群: 每一对网块都有超级网关相连的一组网块, 称为网块群。 网块群内的网 块互为同群网块。
(4) 超联协议选项: 这是为超联协议建立的新的 I Pv4选项, 用来存储数据包在传送 途中必须经过的网址和其它辅助信息。
(5) 超级路由:超级网关集体协作将数据包送往目标网块的过程。每个网块上的大部 分超级网关都在网块内部的路由表中宣布彼此相通; 超级网关通过网址代换, 将 数据包辗转送到目标网块中,并在起点和终点之间造成经由不同网块的多条备用 传输路径。 这样的传递过程类似于普通路由, 但却不能由普通路由系统执行, 因 此称为超级路由。超级网关可以通过普通路由表调节数据流在各传输路径之间的 分配。 超级路由还负责协助实现新、 旧网互通。
(6) 代理超级网关: 与某个超级网关平行连接于两个网块之间的一个或多个特殊网 关。 代理超级网关的端口网址属于特定的预留网址范围, 其两个端口的网址可以 相同。 代理超级网关在路由表中宣布通往被代理的超级网关, 并借用被代理的超 级网关的两个端口网址进行超级路由操作。
(7) 路由安全戳: 一个 32位二进制整数密码, 是超联协议的安全机制的組成部分。
路由安全戳仅在一个网络自治区(Autonomous System)内部的超級网关之间流 通, 用来防止伪造从一个超级网关发往另一个超级网关的数据包。 路由安全戳保 存在超联协议选项中, 由第一个和最后一个收到数据包的超级网关分别加盖和清 除。 国境上的超級网关还负责清除出境数据包中的路由安全戳。 本发明不包括路 由安全戳的细节。
(8) 超级网关管理协议: 是一个为管理超级网关而设立的、 作为安全机制组成部分 的辅助协议。 此协议类似于现行的、 管理路由器所使用的的协议, 但应该仅在 一个网络自治区境内使用; 全球可以统一该协议的识别号, 但不必共享协议的 格式和密码。 本发明不包括超级网关管理协议的细节。
2. 将权利要求 1 中的概念和方法应用于本发明实施之前、 IPv4升级之后的网络层协议。
3. 将权利要求 1 中的概念和方法推广应用于连接具有不同网址长度的网块。 例如连接 IPv4网块和 IPv6网块, 解决 IPv4与 IPv6共存期内两者互通的困难。
PCT/CN2011/002057 2010-12-09 2011-12-09 一种可以取代ipv6的网络层协议 WO2012075688A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP11846093.0A EP2651088A1 (en) 2010-12-10 2011-12-09 Network layer protocol for replacing ipv6
US13/992,740 US20130275608A1 (en) 2010-12-09 2011-12-09 Network-Layer Protocol Substituting IPv6

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201010582838.8 2010-12-09
CN2010105828388A CN102075519A (zh) 2010-12-10 2010-12-10 一种可以取代IPv6的网络层协议

Publications (1)

Publication Number Publication Date
WO2012075688A1 true WO2012075688A1 (zh) 2012-06-14

Family

ID=44033863

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2011/002057 WO2012075688A1 (zh) 2010-12-09 2011-12-09 一种可以取代ipv6的网络层协议

Country Status (4)

Country Link
US (1) US20130275608A1 (zh)
EP (1) EP2651088A1 (zh)
CN (1) CN102075519A (zh)
WO (1) WO2012075688A1 (zh)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102075519A (zh) * 2010-12-10 2011-05-25 谭中飞 一种可以取代IPv6的网络层协议
DE102016221690A1 (de) 2016-11-04 2018-05-09 Audi Ag Verfahren zum Übertragen von Datenpaketen zwischen einem Ethernet und einem Bussystem in einem Kraftfahrzeug sowie Gatewayvorrichtung und Kraftfahrzeug
US10693715B1 (en) 2017-10-26 2020-06-23 Amazon Technologies, Inc. Dynamic network address space allocation for virtual networks
CN116155797A (zh) * 2020-05-13 2023-05-23 华为技术有限公司 一种协议报文的处理方法、网络设备及计算机存储介质
CN112153753B (zh) * 2020-09-24 2022-09-16 维沃移动通信有限公司 网络连接方法及装置

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2006073583A2 (en) * 2004-12-30 2006-07-13 General Instrument Corporation Transparent communication between ipv4 private address spaces using ipv6
US20060268863A1 (en) * 2004-10-29 2006-11-30 Hui-Kai Chang Transparent address translation methods
CN101277309A (zh) * 2007-03-29 2008-10-01 汪涛 一种ip地址系统及在其中建立用户间通信连接的方法
CN102075519A (zh) * 2010-12-10 2011-05-25 谭中飞 一种可以取代IPv6的网络层协议

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7644171B2 (en) * 2001-09-12 2010-01-05 Netmotion Wireless, Inc. Mobile networking system and method using IPv4 and IPv6
KR100429900B1 (ko) * 2002-05-07 2004-05-03 한국전자통신연구원 IPv4 주소체계 확장방법 및 확장된 IPv4주소체계를 이용하여 계층적으로 레이블을 스위칭하는방법 및 그 기록매체
EP1420559A1 (en) * 2002-11-13 2004-05-19 Thomson Licensing S.A. Method and device for supporting a 6to4 tunneling protocol across a network address translation mechanism
GB2413464A (en) * 2004-04-21 2005-10-26 Orange Sa An inter-working unit with a protocol conversion or protocol encapsulation function, for use with dual stack user equipment on a packet radio network
CN101247308B (zh) * 2007-02-13 2011-04-27 上海亿人通信终端有限公司 基于网络处理器实现IPv6穿越IPv4的隧道报文处理方法
CN101232454B (zh) * 2008-02-02 2012-11-21 中国科学院计算技术研究所 一种整合的异构无线网络及其通信方法
JP5018654B2 (ja) * 2008-06-06 2012-09-05 沖電気工業株式会社 通信システム
US8243740B1 (en) * 2008-11-21 2012-08-14 Sprint Communications Company L.P. Using domain name server response and internet protocol version 6 to conserve internet protocol version 4 addresses
US8719337B1 (en) * 2009-04-27 2014-05-06 Junaid Islam IPv6 to web architecture

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060268863A1 (en) * 2004-10-29 2006-11-30 Hui-Kai Chang Transparent address translation methods
WO2006073583A2 (en) * 2004-12-30 2006-07-13 General Instrument Corporation Transparent communication between ipv4 private address spaces using ipv6
CN101277309A (zh) * 2007-03-29 2008-10-01 汪涛 一种ip地址系统及在其中建立用户间通信连接的方法
CN102075519A (zh) * 2010-12-10 2011-05-25 谭中飞 一种可以取代IPv6的网络层协议

Also Published As

Publication number Publication date
CN102075519A (zh) 2011-05-25
US20130275608A1 (en) 2013-10-17
EP2651088A1 (en) 2013-10-16

Similar Documents

Publication Publication Date Title
CN104396192B (zh) 不对称网络地址封装
EP1164754B1 (en) Methods and arrangements in a telecommunications system
CN102577255B (zh) 云计算中企业的第2层无缝站点扩展
CN104350725B (zh) 通过软件定义网络进行信息中心网络的无缝集成和独立演进方法
EP2569902B1 (en) Interconnecting members of a virtual network
EP1368947B1 (en) Addressing method and system for using an anycast address
EP1057309B1 (en) System and method for using domain names to route data sent to a destination on a network
US8103797B2 (en) Parameterized recursive network architecture with topological addressing
USRE41024E1 (en) Communication using two addresses for an entity
US6996621B1 (en) Method for supporting secondary address delivery on remote access servers
CN100484083C (zh) 一种地址转换方法及实现该方法的混合地址转换路由器
CN1954574B (zh) 以太网上的统一架构
CN101095134A (zh) 自动发起和动态建立有防火墙的服务器和有防火墙的客户机之间的安全的因特网连接的系统和方法
WO2012075688A1 (zh) 一种可以取代ipv6的网络层协议
CN1849796A (zh) 用于确定对等路由器的能力和配置的ttl探询技术
CN101321128A (zh) 通信设备、通信网络系统和通信方法
CN100471163C (zh) 在IPv6中用主机间隧道支持IPv4应用程序的方法
US9154571B2 (en) Publish/subscribe networks
CN101707605A (zh) 基于IPv4/IPv6协议翻译的IPSec穿越互连方法
US20070189329A1 (en) System for combining networks of different addressing schemes
CN101175096B (zh) 基于源路由的可扩展ip网络的实现
Naderi et al. A review of IPv6 multihoming solutions
Zhou et al. P6P: A peer-to-peer approach to Internet infrastructure
CN102158477A (zh) 一种通信系统和信息交互方法
JP2018157513A (ja) 通信制御装置、通信制御システム、通信制御方法及び通信制御プログラム

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 11846093

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 13992740

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE

REEP Request for entry into the european phase

Ref document number: 2011846093

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2011846093

Country of ref document: EP