JP6523987B2 - Network connection system and network connection method - Google Patents

Description

  The present invention relates to a technology of a network connection system and a network connection method.

  As an Internet connection method using an IPv6 (Internet Protocol Version 6) address, Non-Patent Document 1 describes a PPPoE (Point-to-Point Protocol over Ethernet) method using a PPP (Point-to-Point Protocol) tunnel, and IP And an IPoE (Internet Protocol over Ethernet) method for transferring the data over Ethernet (registered trademark).

  As described in Non-Patent Document 2 etc., the PPPoE method has no restriction on the number of ISPs (Internet Service Providers) and no restriction on connection sites such as the installation location of POI (Point Of Interface), and ISP companies There is a high degree of freedom for On the other hand, the operator who performs the relay needs to manage the session. For the user, the use of a tunnel reduces the MTU (Maximum Transmission Unit) size, which causes transmission overhead.

  As described in Non-Patent Document 3 and the like, the IPoE scheme has a limitation on the number of ISP companies and a limitation on connection points (place where POI is installed), and the ISP company has a low degree of freedom. On the other hand, the operator that performs the relay does not need to manage the session, and there is no transfer overhead for the user.

About NTT East, "FLET'S Hikari Next, FLET Hikari Light" Internet Connection Using IPv6 Address "", [online], May 20, 2011, [February 8, 2016 search], Internet <URL: https://flets.com/next/IPv6/> NTT East, NTT West, "NGN IPv6 ISP connection <tunnel method> service specification", [online], June 2010, [search on February 8, 2016], Internet <URL: https: // www. ntt-west.co.jp/open/ngn/pdf/IPv6_tunnel_service.pdf> NTT East, NTT West, "NGN IPv6 ISP Connection <Native System> Service Specification", [online], June 2010, [Search on February 8, 2016], Internet <URL: https: // www. ntt-west.co.jp/open/ngn/pdf/ipv6_native_service.pdf>

  In the PPPoE method, an extra header is added to tunnel (encapsulate) the packet, thereby reducing the MTU of user data, which incurs transfer overhead for the user.

  On the other hand, by using the IPoE method, it is possible to eliminate session management and reduce the transfer overhead for the user. However, since the transmission source (SRC) / destination (DST) address is not encapsulated, the POI is selected by the destination address, and the POI to be relayed can not be selected.

Note that there is a method in which information for identifying a communication destination (Identifier) and an actual transfer destination (Locator) are considered separately, and, for example, an ILA method has been proposed as a method that does not encapsulate an IP header.
For example, 3.7 of "Identifier-locator addressing for network virtualization"<URL:https://www.ntt-west.co.jp/open/ngn/pdf/ipv6_native_service.pdf> (ILA draft) proposed as an Internet draft Section 1 describes an area for storing a 32-bit virtual address in a 128-bit IPv6 address frame.
Also, “Identifier Locator Addressing with IPv6” <URL: https://www.ietf.org/proceedings/94/slides/slides-94-v6ops-3.pdf> contains an explanation of the ILA draft. .

  However, since the IPv6 address is 128 bits long, it can not be stored in the 32 bit virtual address proposed by the ILA method. If only the lower 32 bits of the IPv6 global address are extracted and stored in the virtual address, then only the lower 32 bits will match, but multiple addresses in which the upper 96 bits do not match will collide, and the virtual address The transmission source (SRC) / destination (DST) address can not be uniquely identified from the storage destination.

  Therefore, when providing the Internet connection using an IPv6 address, the present invention has as its main object to reliably transfer a packet without applying transfer overhead.

In order to solve the above-mentioned subject, the network connection system of the present invention,
A network connection system for flowing packets via a relay network between a plurality of network terminating devices that connect the relay network and an external network,
An address of the second network termination device in the relay network of the second network termination device, which is directed to the destination address of the received packet, and a destination address of the received packet. Generating a new destination address including the shortened destination address, adding the generated new destination address to the received packet, and transferring the packet to the relay network;
A second network terminating apparatus that has received the packet from inside the relay network reads the shortened destination address from the new destination address of the received packet, and the shortening corresponding to the shortened destination address The destination address before conversion is restored, and the packet to which the restored destination address is added is transferred to the outside of the relay network.

  As a result, the destination address sent to the relay network is shortened without using encapsulation, so that when the Internet connection using the IPv6 address is provided, the packet is reliably transferred without applying transfer overhead. Can.

  In the present invention, in the step of shortening the destination address in the first network termination apparatus, the address before shortening and the address after shortening are associated with the same name, and the DNS (Domain Name System) is used. When registered, it is characterized by including the corresponding address after shortening in the new destination address.

  As a result, it is possible to easily obtain an address that does not overlap with another shortened address as the shortened address.

  In the present invention, in the step of shortening the destination address, the first network termination apparatus shifts each data by a shift amount according to the position of each data constituting the address before shortening, and then shifts the data. It is characterized in that the address after shortening is generated by adding each subsequent data.

  As a result, even when an address that does not overlap with another shortened address is not pre-registered, the shortened address can be automatically generated.

In the present invention, in the step of the first network terminating apparatus adding the new destination address to the packet, a virtual network identifier (VNID) embedded in an address of a source apparatus of the received packet is received. It is characterized in that the second network termination apparatus to be the transfer destination indicated by the new destination address is determined.

  Thereby, even when the transit network is connected to a plurality of ISP companies, the communication of the user of each ISP company can be reliably transferred.

  According to the present invention, when providing an Internet connection using an IPv6 address, packets can be reliably transferred without the transfer overhead.

It is a block diagram of the network connection system of IPv4 concerning this embodiment. FIG. 2A shows a table that associates the input IF with the VLAN and the VNID. FIG. 2 (b) shows a table for determining an address from VNID. FIG. 2C shows the proposed format for improvement when storing in the address of the packet. It is a block diagram of the network connection system of IPv6 concerning this embodiment. 2 shows a name resolution table according to the present embodiment. 7 illustrates a calculation method of address shortening according to the present embodiment.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

First, the notation of the IP address described in the present embodiment will be described.
An IPv4 address is expressed by four decimal numbers separated by dots "." Like "100.64.0.2", and can represent an address space of 8 bits x 4 = 32 bits. Also, a decimal number following a slash "/" like "100.64 / 10" indicates the length (n bits) of the network address shown before the slash.

The IPv6 address is represented by eight hexadecimal numbers separated by colon ":" like "2000: 0: a: a: 4000: 1: 6404", and 16 bits x 8 = 128 bits. It can represent the address space.
Hereinafter, in the present specification, “the m-th field” is described to indicate whether the data is separated by the first colon. For example, the first field of “2000: 0: a: a: 4000: 1: 6404: 1” = “2000”, and the fifth field = “4000”.
Furthermore, according to the IPv6 standard, the following abbreviation rules are also used.
(Rule 1) "0" at the beginning of each field can be omitted.
(Rule 2) When the fields of "0000" are continuous, they can be omitted once with "::".

FIG. 1 is a block diagram of an IPv4 network connection system.
In the network connection system, access from each in-home device (A14, B24, C34) is made from the subscriber accommodation device 13 via each network termination device (A12, B22) to the ISP enterprise network (ISP-A11, ISP) It is configured to flow from B 21) to the Internet and reach the destination device 10.
That is, the subscriber accommodation device 13, the network termination device A12, and the network termination device B22 are network termination devices (NVE: Network Virtualization Edge) that connect the relay network and the external network.

Each device of the network connection system is configured as a computer having a central processing unit (CPU), a memory, a storage unit (storage unit) such as a hard disk, and a network interface.
The computer operates a control unit (control means) configured of respective processing units described later by the CPU executing a program (also referred to as an application or an abbreviation of the application) read into the memory.

The in-home apparatus A14 is an apparatus placed at the user home of the ISP-A11, and the in-home apparatuses B24 and C34 are apparatuses placed at the user home of the ISP-B21.
First, the address “100.64.0.1” is assigned to the in-home device A14. As described above, an address belonging to the “100.64 / 10” network, which is an IPv4 address for CGNAT that has recently appeared due to exhaustion of the IPv4 address, is assigned to each home device, but another address may be assigned.
On the other hand, the subscriber accommodating device 13 and each network terminating device are managed by a relay carrier different from the above-mentioned ISP carrier. The relay operator assigns locator setting L11 “Lo: fd00: a: b: 8001 :: 1” to the network terminating device A 12 and locator setting L 12 “Lo: fd 00: a: b: 8002 :: 1 is allocated, and the locator setting L13 "Lo: fd00: a: b: 1: 1" is allocated to the subscriber accommodation device 13.

The ISP company or relay company may actually assign an IP address defined by the ISP-A 11 and the ISP-B 21 to each home device.
Further, although not described in FIG. 1, the relay operator manages a relay device for relaying packets in addition to the subscriber accommodation device 13 and each network termination device. This relay device exists between the subscriber accommodating device 13-the network termination device A12 and the network termination device B22.

  It is assumed that the in-home apparatus A14 transmits the packet P11 toward the destination apparatus 10. The transmission source (SRC) of the packet P11 is the address "100.64.0.1" of the home apparatus A14, and the destination (DST) is the address "1.2.3.4" of the destination apparatus 10. Further, “1001” is assigned by the relay operator as a VLAN value for logically separating the L2 domain on a subscriber basis in the packet P11. The subscriber accommodating apparatus 13 receives the packet P11 via “VLAN = 1001”.

  The subscriber accommodation device 13 has a table 91 for determining an input IF, a VLAN, and a VNID (virtual network identifier) shown in FIG. 2A, and a VNID to a standard identifier representation (SIR) prefix shown in FIG. 2B. And a table 94 for determining a Locator prefix from the VNID shown in FIG. 2B. Note that VNID is a term described in Section 3.6 and Section 3.7.1 of the ILA draft, and is an ID for identifying which ISP's communication in this example.

The subscriber accommodating apparatus 13 refers to the above-described tables 91, 93, 94, and transfers the packet P11 received in the following order as a packet P12.
(Procedure 1) From the input IF of the packet P11 and the VLAN, the table 91 is referenced to determine the VNID of the packet P12. For example, “VNID = 1” is obtained from “VLAN: 1001” of the packet P11 and the reception IF = Eth0.

(Procedure 2) Based on the VNID of the packet P12, the transmission source (SRC) of the packet P12 is determined based on the SIR prefix obtained by referring to the table 93. For example, SIR prefix “2000: 0: a: a” is obtained from “VNID = 1” in (Procedure 1). Then, according to the format described in Section 3.6 of the ILA draft (hereinafter, “ILA format”), the obtained “2000: 0: a: a” is stored in the first 64 bits, and then Type (3 bit) = 2 and “ By storing VNID = 1 ”(29 bits) and the lower 32 bits (= address“ 100.64.0.1 ”of the home apparatus A14),“ SRC: 2000: 0: a: a: 4000: 1: 6404: 1 ”. To determine
As described later with reference to FIG. 3, in the IPv6 network, the improvement proposal format 95 of FIG. 2C is used instead of the Type 2 ILA format.

(Procedure 3) Based on the VNID of the packet P12, the destination (DST) of the packet P12 is determined based on the Locator obtained by referring to the table 94. For example, Locator “fd00: a: b: 8001” can be obtained from “VNID = 1”. Then, according to the ILA format, the obtained "fd00: a: b: 8001" is stored at the first 64 bits, and then Type (3 bits), "VNID = 1" (29 bits), lower 32 bits (= destination device 10 By storing the address “1.2.3.4”, respectively, “DST: fd00: a: b: 8001: 4000: 1: 102: 304” is determined.
(Procedure 4) A packet P12 generated based on the determination contents of (Procedure 2) and (Procedure 3) is transmitted to the network terminating device A12. Since the network termination device A12 carries out the route advertisement R11 in advance of the network “fd00: a: b: 8001/64” which itself is in charge of transfer, the packet P12 reaches the network termination device A12.

When the network terminating device A12 receives the packet P12 from the subscriber accommodating device 13, it generates a packet P13 from the packet P12.
The transmission source (SRC) of the packet P13 is the seventh and eighth fields (that is, the lower 32 bits of the packet) of the transmission source (SRC) of the packet P12.
The destination (DST) of the packet P13 is the seventh and eighth fields of the destination (DST) of the packet P12.

  The network termination device A12 holds a table 92 shown in FIG. 2 (a). The table 92 has the same format as the table 91. The network termination device A12 reads the VNID (stored between the 67th bit and the 96th bit from the beginning) in the ILA format of the packet P12, refers to the table 92 from the VNID, and outputs the IF of the packet P13 and the VLAN. decide. Thus, the destination device 10 can receive the packet P13 designated by the IPv4 address via the ISP-A 11 even if the IPv6 relay network is crossed.

The packet transfer procedure has been described above in the order of the in-home apparatus A14 → (packet P11) → subscriber accommodation apparatus 13 → (packet P12) → network termination apparatus A12 → (packet P13) → ISP-A11 → destination apparatus 10.
Similarly, packets are transferred in the order of the in-home device B24 → (packet P21) → subscriber accommodating device 13 → (packet P22) → network terminal device B22 → (packet P23) → ISP-B21 → destination device 10. For example, in the case of the packet P22 which is "DST: fd00: a: b: 8002: 4000: 2: 102: 304", the network terminal device which has "fd 00: a: b: 8002/64" as the route advertisement R12. It reaches B22.
The same applies to the packet P31 from the in-home apparatus C34.

Further, the downstream packet to which the in-home apparatus A14 or the like is addressed from the network termination apparatus A12 or the like reaches the subscriber accommodation apparatus 13 which has been performing the route advertisement R13 of "fd00: a: b: 1/64".
That is, in the upstream packet, the subscriber accommodation device 13 refers to the tables 92 to 94, identifies VNID = 1 from the packet P11, and generates the packet P12 based on the identified information.
On the other hand, for downstream packets, the network termination devices A12 and B22 refer to the information equivalent to the table 93 held by itself, generate the transmission source IPv6 address, and then refer to the information equivalent to the table 94 held by itself. The locator prefix of the subscriber accommodation device 13 is acquired, and a destination IPv6 address used in the relay network is generated and transferred. The generated IPv6 packet is forwarded according to the route advertisement R13 of the subscriber accommodation device 13.

  Each of the tables 91 to 94 may be an external server in addition to being held in the above-described apparatus. At that time, although it is conceivable to acquire it from the outside through an authentication protocol such as Radius, or another protocol for directory service, in this example, it is assumed that data is written inside the device.

FIG. 3 is a block diagram of a network connection system of IPv6. The network connection system of FIG. 3 is configured to include a DNS 63 in addition to the respective devices of FIG. An IPv6 address is assigned to each in-home apparatus A74, B84 and each destination apparatus 61, 62.
In-home devices A74 and B84 are provided with a SIR address from a relay operator according to the policy of the ISP, using Router Advertisement (RA), Dynamic Host Configuration Protocol (DHCP) v6, DHCPv6-PD (prefix delegation), etc. .
In addition, for each device in the relay carrier network between the network termination devices A 72 and B 82 to the subscriber accommodation device 73, each address along the “network Prefix = fd00: a: b / 48” from the relay carrier (If it is the network terminal device A72, the locator setting L51 "Lo: fd00: a: b: 8001 :: 1", if the subscriber accommodation device 73 the locator setting L52 "Lo: fd00: a: b: 1 :: 1") is assigned Be

  In the subscriber accommodation device 73, the upper 96 bits of the destination (DST) address of the packet P51 received from the in-home device A74 are prefixed with the tenant (including VNID) prefix “In-network Prefix = fd00: a: b / 48” If not, the packet P52 is created according to the following procedure.

(Procedure 11) The IPv6 of the packet P51 is reversely referred to by referring to the name resolution table 96 of FIG. 4 managed by the DNS 63 (authority server). That is, a row (here, the second row) of the name resolution table 96 matching the IP address “DST: 2001: 2: 3: 4: 1: 2: 3: 4” of the packet P51 is obtained, and the name of the row is obtained. Get (Hostname "www.example-ntt.com").
(Procedure 12) A line corresponding to the name obtained in (Procedure 11) is searched, and an IP address whose “DNS record” of the searched line is “A record” is acquired. For example, among the first and second lines corresponding to the host name “www.example-ntt.com”, the “DNS record = A record” is the first line. The IP address of the first line = "5.6.7.8" is acquired.

(Procedure 13) Create the destination (DST) of packet P52 in which the IPv4 address (abbreviated address) obtained in (Procedure 12) above is embedded in the lower 32 bits of the improvement proposal format 95 shown in FIG. .
Here, the “improvement” of the improvement proposal format 95 is a notation for distinguishing from the existing ILA format of Type = 2, and uses a new type number of Type = 5.
The method of creating each field other than the lower 32 bits of the improvement proposal format 95 is the same as the processing of creating the packet P11 to the packet P12 in FIG. 1, but in FIG. On the other hand, in FIG. 3, the VNID itself is embedded in the address of the in-home apparatus A 74 ("1" of the sixth field).

The created packet P52 is transmitted from the subscriber accommodation device 73, and reaches the network terminating device A 72 which has been performing the route advertisement R51 of "fd00: a: b: 8001/64". Then, in the same manner as in FIG. 1, the packets are in the order of in-home apparatus A 74 → (packet P 51) → subscriber accommodation apparatus 73 → (packet P 52) → network termination apparatus A 72 → (packet P 53) → ISP-A 71 → destination apparatus 61. To be transferred.
Here, the network terminating device A 72 is the AAAA record of the host name from the host name “www.example-ntt.com” of the first line that matches the lower 32 bits “0506: 0708” of the destination (DST) of the packet P52. “2001: 2: 3: 4: 1: 2: 3: 4” is searched from the second line, and the search result is set as the destination (DST) of the packet P53.

If there is no “DNS record = A record” line in (Procedure 12), that is, if there is no IPv4 address on the host of IPv6 only, instead of (Procedure 12) as shown below (Procedure 12a ) Should be executed.
(Procedure 12a) As shown in FIG. 5, based on “DST: 2001: 2: 3: 4: 1: 2: 3: 4” of the packet P51, a 22-bit compressed address (abbreviated address), Create 28-bit data connected with the upper 6 bits for collision avoidance. The 28-bit data is registered in the name resolution table 96 in the space of 240.0.0.0/4 (class E) and is updated by D (Dynamic) DNS.

  In FIG. 5, with respect to the mth field of "DST: 2001: 2: 3: 4: 1: 2: 3: 4," shift to the left by (8-m) bits (if "<< 7, shift to the left by 7 bits) In the example of calculation, m = 1 to 8 fields are created, and the sum of them is used as a 22-bit compressed address. Also, when compressed addresses collide, if there is no collision as the upper 6 bits, “0” if there is no collision, “1” if it collides once, “2” if it collides twice, and so on. By combining with compressed addresses, collisions can be avoided. Such addition processing after shift of each field has a low computational load and can use a 22-bit compressed address space widely.

  The name (host name) at the time of registration in DDNS (the authoritative server of DNS 63) is a string obtained by converting the: (colon) notation of the IPv6 address as-(hyphen). The DNS 63 sets the A record as a temporary address generated by the method shown in FIG. 5 (the third line in FIG. 4), and the AAAA record as the IPv6 address of the destination (DST). (4th line in FIG. 4) Register each.

In the embodiment described above, when the subscriber (in-home device A 74) transmits the packet P51 to the device (destination device 61) having the global IPv6 address as the destination address, the subscriber accommodation device 73 is the network termination device. The address “fd00: a: b: 8001” of the network terminating device A 72 of the transfer destination and the address “DST: 2001” transmitted by the subscriber in order to transfer the packet without tunneling the packet to the ISP-A 71 across A72. : 2: 3: 4: 1: 2: 3: 4 "in combination with the shortened address" 5.6.7.8 ". Also, the network terminating device A 72 restores the received address to the original IPv6 address.
Thus, when relaying the ISP carrier network, it is possible to relay a packet to a connection destination desired by a specific ISP carrier without performing tunneling (capsulation of the header).

The network termination devices A12 and A72 refer to the Type column of the received packet, and in the case of an ILA format of Type = 2, the address contained in the lower 32 bits is an IPv4 address that has not been shortened, so shortening Restoration is unnecessary.
On the other hand, in the case of the improvement proposal format 95 of Type = 5, since the address included in the lower 32 bits is an address that has been shortened, the IPv6 address before the shortening is acquired by restoring the shortening.
That is, the source / destination address of the packet indicating each in-home device or each destination device is the same relay network that transmits both the packet of the IPv4 address shown in FIG. 1 and the packet of the IPv6 address shown in FIG. You may

  Here, the address shortening process may use any one-way conversion function, but as an example, mapping information between an IPv4 address and an IPv6 address registered as the same name in the DNS 63 may be used. Alternatively, when the mapping information is not registered in the DNS 63, an IPv4 address may be newly generated from the IPv6 address, and mapping information between the generated IPv4 address and the IPv6 address of the generation source may be used.

A plurality of ISPs (ISP-A11 and ISP-B21) are accommodated in the transfer network, and each ISP transfers the subscriber's packet through the route set by itself. Therefore, in order to specify which ISP packet belongs to, each packet in the transfer network is given a VNID.
This VNID can be identified from the combination of the in-home VLAN and the packet reception IF of the subscriber accommodation device 13 in FIG. 1 (table 91 of FIG. 2A). Alternatively, in FIG. 3, the VNID is distributed as a part (sixth field) of the IP address of each home device from a router advertisement (RA) or a DHCP server.

10, 61, 62 Destination device 11, 71 ISP-A
12, 72 Network termination A (second network termination)
13, 73 Subscriber accommodation device (first network termination device)
14, 24, 34, 74, 84 in-home apparatus 21, 81 ISP-B
22, 82 Network terminator B
63 DNS

Claims (5)

A network connection system for flowing packets via a relay network between a plurality of network terminating devices that connect the relay network and an external network,
The first network termination device that has received the packet from outside the relay network receives the address of the second network termination device in the relay network that is directed to the destination address of the received packet, and the destination address of the received packet. Generating a new destination address including the shortened destination address, adding the generated new destination address to the received packet, and transferring the packet to the relay network;
The second network termination apparatus having received the packet from inside the relay network reads the shortened destination address from the new destination address of the received packet, and the shortening corresponding to the shortened destination address is performed. A network connection system characterized in that the destination address before conversion is restored, and the packet with the restored destination address is transferred to the outside of the relay network. In the first network terminating device, in the process of shortening the destination address, the address before shortening and the address after shortening are associated with the same name and registered in DNS (Domain Name System) 2. The network connection system according to claim 1, wherein the new address after address is included in the new destination address when the address is shortened. The first network terminating device shifts each data by a shift amount according to the position of each data forming the address before shortening in the step of shortening the destination address, and then shifts each data after the shift. The network connection system according to claim 1, wherein an address after shortening is generated by adding. In the step of adding the new destination address to the packet, the first network termination device is based on a virtual network identifier (VNID) embedded in the address of the source device of the received packet. The network connection system according to any one of claims 1 to 3, wherein the second network termination device to be a transfer destination indicated by the new destination address is determined. A network connection method of a network connection system in which packets are flowed through the relay network between a plurality of network termination devices that connect the relay network and an external network,
The first network termination device that has received the packet from outside the relay network receives the address of the second network termination device in the relay network that is directed to the destination address of the received packet, and the destination address of the received packet. Generating a new destination address including the shortened destination address, adding the generated new destination address to the received packet, and transferring the packet to the relay network;
The second network termination apparatus having received the packet from inside the relay network reads the shortened destination address from the new destination address of the received packet, and the shortening corresponding to the shortened destination address is performed. A network connection method characterized in that a destination address before conversion is restored, and the packet with the restored destination address is transferred to the outside of the relay network.

Images