JP2002232461A - Network system - Google Patents

Abstract

(57) [Summary] [Purpose] It is an object to perform network communication considering quality for each IP packet. An ISP device for accommodating a subscriber access network
0 and a relay device 40 for relaying the IP packet of the ISP device 20 on the Internet backbone, the ISP device 20 adds a high quality / low quality type quality identifier to the IP packet from the subscriber. On the network including the relay device 40,
Differentiation control (pass / discard, etc.) according to the quality identifier of the P packet is performed. In addition, a server B adjacent to the relay device B is provided. The relay device B stores the quality identifier added to the upstream IP packet to the server B, and stores the quality identifier in the downstream IP packet corresponding to the upstream IP packet from the server B. On the other hand, the same quality identifier is added.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a network system, and more particularly, to a network system including an ISP device for accommodating an access network of a subscriber and a relay device for relaying an IP packet of the ISP device on an Internet backbone.

[0002]

2. Description of the Related Art FIG. 20 is a diagram for explaining a conventional technology, and shows a typical network configuration of the Internet. The Internet employs a configuration in which a plurality of access networks A to C are mutually connected via an Internet backbone. 10 is a user terminal device, ALC is an access line accommodating device, and 20 'is an ISP (Internet Service Provider).
der) device, 40 'is I on the Internet backbone
A relay device for relaying P packets, and 80 is a server.

Each terminal device 10 is assigned an IP address. For data transfer on the Internet, an IP (Internet Protocol) packet in which information is converted into a variable-length packet and a header thereof is added with a communication destination, a transmission source address, and the like is added. The so-called IP communication for transfer is performed. ISP device 20 '
There is a connection between the telephone and the subscriber terminal 10 through a dial-up connection through a telephone line, or a connection always through a dedicated line or the like, and a certain bandwidth is secured for each subscriber.

[0004] TCP (Transmission Control Protocol)
l) is a connection-oriented reliable transport layer protocol that guarantees end-to-end data arrival and is reliable in this sense, but on the Internet backbone, all subscriber data is processed equally (best (Effort transfer), when the load on the relay device 40 'increases, the packet may be discarded indiscriminately. On the other hand, UDP
(User Datagram Protocol) is an unreliable connectionless transport layer protocol that is used for highly real-time audio and broadcast communications.
UDP does not participate in end-to-end data arrival confirmation and is unreliable in this sense. Conventionally, when it is necessary to confirm data arrival between end points, this is performed exclusively by a user application program.

[0005]

As described above, the conventional T
The CP performs communication that treats all packets equally (best-effort communication). For this reason, if the load on the relay device increases, the packets may be discarded indiscriminately.
On the other hand, since UDP does not allow packet discarding, TC
It is not used in the same network as P, but is operated in a separate network.

In the conventional TCP, when a user tries to download data from a certain server, even if a packet from the user reaches the server, the response (download) data is not always delivered to the user.
If the response data was discarded, the user had to retransmit the access packet, which was inefficient.

The present invention has been made in view of the above-mentioned problems of the prior art, and has as its object to provide a network system capable of performing network communication considering the quality of each IP packet.

[0008]

The above-mentioned problem is solved, for example, by referring to FIG.
Is solved. That is, the network system of the present invention (1) is an IS that accommodates a subscriber's access network.
In a network system including a P (Internet Service Provider) device 20 and a relay device 40 that relays an IP packet of the ISP device 20 on an Internet backbone, the ISP device 20 provides high quality / low quality for IP packets from a subscriber. Append a type quality identifier,
On the network including the relay device 40, differentiation control (pass / discard, etc.) according to the quality identifier of the IP packet is performed.

An identifier of priority information is added to an IP packet transmitted between subscriber terminals in the middle (ISP device 20), and the priority identifier is detected by each relay device 40 that passes through the packet. The corresponding differentiating control is performed. Then, when packet congestion occurs in the relay device 40, low quality type packets are preferentially discarded, and high quality type packets are not discarded as much as possible.
Therefore, network communication can be efficiently performed in consideration of the quality of each IP packet according to the purpose of the IP packet.

Preferably, in the present invention (2), in the present invention (1), a TCP / IP packet and a UDP /
IP packets are mixed on the same network,
The P device adds a high quality type quality identifier to the subscriber's UDP / IP packet.

[0011] For example, real-time UDP related to voice communication and the like
/ IP packets are mixed under the common differentiating control according to the present invention, and a high quality type quality identifier is added to a UDP / IP packet that cannot be discarded, so that the packet is transmitted preferentially. You.

Preferably, in the present invention (3), in the above-mentioned present invention (1), the ISP device 20 is provided, for example, as shown in FIG.
As shown in (1), a contract information management memory for storing contract information of service quality previously determined with a subscriber, and a packet monitoring unit 21 for detecting a source address of an input packet.
And an additional identifier determining unit 25 that determines a quality identifier to be added to the input packet with reference to the contract information in the contract information management memory based on the detected source address, and adds the determined quality identifier to the input packet. Tagging unit 22
And a differentiation unit 23 for differentiating the added packet according to the quality identifier, and a routing unit 24 for routing the packet after the differentiation according to the destination address. Therefore, it is possible to efficiently perform the differentiation control of the IP packet based on the service quality determined in advance with the subscriber.

Preferably, in the present invention (4), in the above-mentioned present invention (1), the ISP device 20 is provided, for example, as shown in FIG.
As shown in the figure, a contract information management memory for storing contract information of a high quality setting packet amount per unit time previously determined with a subscriber, a source address of an input packet is detected, and an input packet for each subscriber is detected. Based on the packet monitoring unit 21 for measuring the amount and the detected source address,
An additional identifier determining unit that compares the measured amount of input packets with the amount of high-quality set packets in the contract information management memory and determines whether to add a high-quality type quality identifier to input packets that do not exceed the amount of high-quality set packets; 26 and a tagging unit 22 for adding the determined quality identifier to the input packet.
And a differentiation unit 23 for differentiating the added packet according to the quality identifier, and a routing unit 24 for routing the packet after the differentiation according to the destination address. Therefore, IP packet differentiation control based on the amount of high quality set packets per unit time previously determined with the subscriber can be efficiently performed.

Preferably, in the present invention (5), in the above-mentioned present invention (1), a server B adjacent to the relay device B is provided as shown in FIG. And the same quality identifier is added to the downlink IP packet corresponding to the uplink IP packet from the server B.

Therefore, even in the case of two-way communication in which data is downloaded on the Web, the user A sends the server B
If the data (request packet) up to this is transmitted as a high quality type, the same priority identifier is added to the response data (response packet) as a high quality type. If one is guaranteed, the other is also guaranteed. A network that guarantees bidirectional data can be easily constructed. Therefore, the upstream packet is given priority and the communication destination B
, But the data in the down direction is
, And the phenomenon that the user A repeats retransmission to the access server B can be effectively avoided.

[0016]

Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Note that
Throughout the drawings, the same reference numerals indicate the same or corresponding parts. FIG. 2 is a configuration diagram of the network system according to the embodiment. This network system adopts a configuration in which a plurality of access networks A to C are mutually connected via an Internet backbone. In the figure, 10 is a user terminal device, ALC is an access line accommodation device,
20 is an ISP (Internet Service Provider) device, 40
Is a relay device, and 80 is a server. This network configuration is the same as that described with reference to FIG. 20, but by configuring the ISP device 20 and the relay device 40 as in the following embodiments, IP packet differentiation according to the present invention is realized.

FIG. 3 is a diagram for explaining an ISP device according to the first embodiment, in which IP packet transmission (pass / discard) is differentiated according to a packet transmission quality contract previously exchanged between the user and the ISP. Is shown. FIG. 3A shows a block diagram of the ISP device 20. In the figure, 21 is a packet monitoring unit, 22 is a tagging unit,
23 is a differentiation processing unit, 24 is a routing unit, 25 is a contract information management unit, and 26 is an additional identifier determination unit.

A contract between the user and the ISP is made for the quality of the packet on the network, and the contents are stored in the contract information management memory of the contract information management unit 25 in advance. FIG.
(B) shows the contents stored in the contract information management memory. For example, user A has a high quality type contract and user B has a low quality type contract.

On the other hand, the additional identifier determining section 26 has an additional identifier determining memory, in which additional identifier (tagging) information corresponding to each quality type is stored. FIG. 3C shows the storage contents of the additional identifier determination memory. For example, the high-quality type tagging information is “0x01”, and the low-quality type tagging information is “0x00”.

With this configuration, the packet monitoring unit 21 detects the source IP address (user) from the IP header of the input packet and notifies the contract information management unit 25 of this.
The contract information management unit 25 detects the quality type corresponding to the transmission source IP address (user) with reference to the contract information management memory, and notifies the additional identifier determination unit 25 of the quality type. The additional identifier determining unit 26 reads the tagging information “0x01” for the high-quality type and the tagging information “0x00” for the low-quality type, and notifies the tagging unit 22 with reference to the additional identifier determining memory. The tagging unit 22 adds the tagging information to the TOS (Type of Service) field of the packet, and runs the information in parallel with the packet as priority information (information indicating that the packet is a priority-added packet according to the present invention). Notify the differentiation processing unit 23.

FIG. 5A shows the format of the IP header. The service type (TOS; Type of Service) is composed of 8 bits and is defined to indicate the service quality of an IP packet. FIG. 5B shows the bit format of the TOS field. For writing the tagging information, for example, bits 0 to 2 indicating the priority of the IP packet or unused bits 7 can be used. Conventionally, the TOS field has been set and managed exclusively by the user's application program.
The difference is that the data is automatically added by the P device 20.

Returning to FIG. 3A, the differentiation processing unit 23
Normally, the input priority-added IP packet is passed as it is. However, for example, when the processing load of the ISP device 20 becomes large (packet is congested) and the packet has to be discarded, the notified priority is given. The quality of the input packet (TOS field) is checked based on the degree information, and the packet is discarded in the case of the low quality type. The packet may be passed / discarded only with the notified priority information (that is, without referring to the TOS field of the IP packet). Further, the routing processing unit 24 performs routing on the IP packet after the differentiation processing according to the destination. By incorporating such differentiation control into the ISP device 20 or a relay device 40 described later, communication considering quality can be realized end-to-end.

FIG. 3D shows an example of the configuration of the differentiation processing section 23. The differentiation processing unit 23 is, for example, as shown in FIG.
As shown in (1), a high-priority buffer and a low-priority buffer are provided to handle the quality of packets (priority information), and input packets are sorted and stored in a high-priority (quality) buffer or a low-priority (quality) buffer according to the priority information. When the ISP device 20 is congested, the IP packet in the high-priority buffer is read with priority.

Alternatively, as shown in FIGS. 3D and 3B, the IP
A common buffer for packets is provided, and a threshold is provided in the common buffer. If the amount of packets accumulated in the buffer does not exceed the threshold, all input packets are written to the buffer. When the number of packets stored in the buffer reaches the threshold value, control is performed such that high-quality packets can be written, but low-quality packets are discarded, and differentiation according to packet quality is performed.

FIG. 4 is a block diagram of a relay device according to the first embodiment, and shows a configuration of a relay device 40 capable of performing a differentiation operation in cooperation with the ISP device 20 according to the first embodiment. In the figure, reference numeral 41 denotes a priority information detecting unit,
Reference numeral 2 denotes a differentiation processing unit, and reference numeral 43 denotes a routing unit.

The priority information detector 41 detects the T
The value of the OS field is read out, the value is run in parallel with the packet as priority information, and the packet is notified to the differentiation processing unit 42. Normally, the differentiation processing unit 42 allows the input priority-added IP packet to pass as it is, but, for example, when the processing load of the relay device 40 becomes large (packet is congested) and the packet has to be discarded. , The quality of the input packet (TOS field) is checked based on the notified priority information, and the packet is discarded in the case of the low quality type. Note that only the notified priority information (that is, TO
The packet may be passed / discarded (without referring to the S field). Further, the routing processing unit 43 transfers the differentiated IP packet according to the destination IP address.

FIG. 6 is a diagram for explaining an ISP device according to the second embodiment, in which an ATM (Async) is used as an access line.
(Hronous Transfer Mode) shows a case where a network is accommodated in an ISP device. By performing fine control of service quality in the ATM network and performing differentiation processing in cooperation with the ATM access line in the IP network, network communication can be performed in consideration of fine quality for each IP packet.

FIG. 6A is a block diagram of the ISP device 20. In the figure, reference numeral 27 denotes a packet assembling unit that assembles ATM cells into IP packets. Other configurations may be the same as those of the ISP device according to the first embodiment. In this case, each of the terminal devices 10 of the users EG operates in the ATM mode.

FIG. 7A shows a format diagram of an ATM cell, and FIG. 7B shows an image of assembling a packet of the ATM cell. With the ATM access line, strict service quality can be specified with the network side, and the network side performs bandwidth control according to the quality contract. Further, the packet assembling unit 27 assembles a plurality of ATM cells into an IP packet, so that a fine-grained differentiation process for the IP packet can be performed thereafter.

FIG. 6B shows an example of an ATM access network. In the figure, for example, for a user E who seeks a real-time service (high quality service) such as voice, a CBR (Cons) that realizes low delay and low jitter is provided.
tant Bit Rate) Connection is applied to guarantee the bandwidth for each user. For a user F who demands low cost, UBR (Unspecif
ied Bit Rate) connection, a certain bandwidth is shared by multiple users, and the service quality is set low. For a user G who requires a certain level of quality and low cost, a GFR (Guar
anteed Frame Rate) connection. Under such circumstances, the user and the ISP make a quality contract suitable for the ATM quality of the access line, and perform end-to-end consistent quality control. Hereinafter, a specific description will be given.

For example, the user E sets the access line to CBR.
If a contract is made with a PCR (Peak Cell Rate) of 10 Mbps, 10 Mbps data is guaranteed on the access line.
By adding a high-quality type to data (IP packets) equivalent to “its”, data equivalent to 10 Mbps can be communicated as a high-quality packet from end to end.

The user F sets the access line to GFR PCR = 10 Mbps and MCR (Minimum Cell Rat
e) If a contract is made at 5 Mbps, data of MCR = 5 Mbps is guaranteed on the access line.
The P device 20 adds a high-quality type to data (IP packets) equivalent to 5 Mbits per second, and adds a low-quality type to data (IP packets) equivalent to the remaining 5 Mbits per second. 5 Mbps data can be communicated as high quality packets, and the remaining 5 Mbps data can be communicated as low quality packets.

If the user G has contracted the access line with UBR, the UBR band is not guaranteed,
In the ISP device 20, a low quality type is added to a packet that has passed through the UBR connection.

FIG. 8 is a diagram for explaining an ISP device according to the third embodiment. The ISP device 20 monitors the amount of input packets and, according to the amount of input packets, converts user packets into high-quality type / low-quality type. Is set. FIG. 8A is a block diagram of the ISP device 20. In the figure, reference numeral 21 denotes a packet monitoring unit for measuring an input packet amount, and reference numeral 25 denotes a contract information management unit for storing and managing contract data of the input packet amount.

FIG. 8B shows the contents of the contract information management memory in the contract information management unit 25. Subscriber and ISP in advance
And a contract for a high-quality set packet amount indicating how many bytes worth of packets per unit time are to be of the high-quality type, is stored in the contract information memory. For example, the subscriber A has a contract of 10 kbytes per unit time, and the subscriber B has a contract of 32 kbytes per unit time.

With this configuration, the packet monitoring unit 21
The source address of the input packet is detected and notified to the contract information management unit 25, and the input packet amount per unit time for each subscriber (source IP address) is measured and this is added to the additional identifier determination unit 26. Notify. The contract information management unit 25 reads the corresponding high-quality set packet amount based on the source address (subscriber) of the packet, and notifies the additional identifier determination unit 26 of this. Additional identifier determining unit 26
In the above, the measured input packet amount is compared with the read high quality setting packet amount, and tagging information for adding a high quality type is generated for input packets not exceeding the high quality setting packet amount, In addition, tagging information for adding a low-quality type to input packets exceeding the high-quality set packet amount is generated, and this is notified to the tagging unit 22. The tagging unit 22 performs necessary tagging on the TOS field of the input packet according to the notified tagging information. Therefore, IP packet differentiation control according to the input data amount per unit time can be efficiently performed.

By the way, some users connect to the Internet mainly at night or on holidays, and do not connect during the daytime on weekdays, or vice versa. Therefore, a differentiating process is performed by adding a high-quality type or a low-quality type to a packet according to a time zone in which communication is performed. This will be described below.

FIG. 9 is a diagram for explaining an ISP device according to the fourth embodiment. The ISP device 20 adds a high-quality type / low-quality type to a user packet according to quality setting (contract) information for each time zone. Is shown. FIG.
FIG. 2A is a block diagram of the ISP device 20. In the figure, reference numeral 28 denotes a clock section for counting the actual time. Further, the contract information management unit 25 includes a contract information management memory as shown in FIG. In this case, the user must
You can sign a quality type for each time zone with.

FIG. 9B shows the contents of the contract information management memory. (A) of the figure shows the contents of the quality call table by time zone, in which the address of the quality table to be called by time zone is recorded. (B) of the figure shows the contents of a quality table provided for each time zone, in which quality types set for each subscriber are recorded.

With this configuration, the packet monitoring unit 21
The source address (subscriber) is detected from the input packet and notified to the contract information management unit 25. In the contract information management unit 25, the time zone (for example, 8: 00 ≦
t <18:00), and reads the corresponding table address a from the time zone-specific quality call table. Further, based on the source address (for example, A) of the input packet, the corresponding high quality type is read from the quality table a, and this is notified to the additional identifier determination unit 26. Additional identifier determining unit 26
Generates the tagging information corresponding to the notified high quality type 2 and notifies the tagging unit 22 of the generated tagging information. Tagging unit 22
Performs necessary tagging on the TOS field of the packet according to the notified tagging information. The same applies to the case where the quality information read from the quality table is of a low quality type. Thus, quality control for each time zone can be performed efficiently.

FIG. 10 is a diagram for explaining an ISP device according to the fifth embodiment, and shows a case where a user packet is set to a high quality type / low quality type according to a time zone and an input packet amount. FIG. 10A shows the ISP device 2
0 shows a block diagram. In the figure, the contract information management unit 25 includes a contract information management memory as shown in FIG. 10B, and in this case, the user can contract with the ISP in advance for the guaranteed input packet amount for each time zone.

FIG. 10B shows the contents of the contract information management memory. (A) of the figure shows the contents of the packet amount calling table for each time zone, in which the address of the high quality setting packet amount table to be called for each time zone is recorded.
(B) of the figure shows the contents of the high quality setting packet amount table provided for each time zone, in which the high quality setting packet amount per unit time set for each subscriber is recorded.

With this configuration, the packet monitoring unit 21
The source address (user) of the input packet is detected, and this is notified to the contract information management unit 25. It also measures the amount of input packets per unit time for each subscriber (source IP address), and notifies the additional identifier determination unit 26 of the amount of input packets corresponding to the source address 2 of the input packet. On the other hand, the contract information management unit 25 reads the corresponding table address a from the hourly packet amount call table based on the time zone to which the current time belongs (for example, 8:00 to 18:00), and further reads the source address of the input packet ( For example, based on A), read the corresponding high quality setting packet amount (10 kbytes) from the high quality setting packet amount table a,
This is notified to the additional identifier determination unit 26.

The additional identifier determining unit 26 compares the measured amount of input packets with the read amount of high quality set packets, and determines that the amount of input packets not exceeding the high quality set packet amount is high. It generates tagging information of a quality type, and generates tagging information of a low quality type for an input packet exceeding the high quality set packet amount, and notifies the tagging unit 22 of this. Tagging unit 22
Then, according to the notified tagging information, necessary tagging is performed on the TOS field of the packet. Thus,
Differentiation control according to the input packet amount for each time zone can be efficiently performed.

FIG. 11 is a diagram for explaining an ISP device according to the sixth embodiment. By adding a quality identifier according to the CLP bit of an ATM cell to an IP packet generated by the ISP device, it is possible to cooperate with an ATM access line. It shows a case where a network is constructed. A in FIG. 7A
In the TM cell format, cell loss priority indication (C
In LP (Cell Loss Priority), bit “0” indicates priority, the cell is hardly discarded, and bit “1” indicates non-priority, and the cell is easily discarded. Therefore, ATM
By adding a quality identifier according to the CLP bit of the cell (head cell) to the assembled packet, a network cooperating with the ATM access line is constructed.

FIG. 11A is a block diagram of the ISP device 20. In the figure, reference numeral 29 denotes a cell monitoring unit for detecting a CLP bit of an input cell. FIG. 11B shows the contents of the CPL information translation memory in the additional identifier determination unit 26. Here, the "high quality type" corresponds to the CLP bit "0" and the "low quality type" corresponds to the CLP bit "1". Is recorded respectively.

With such a configuration, the cell monitoring unit 29
The CLP bit of the M cell (preferably, the head cell for each user) is detected, and this is notified to the additional identifier determination unit 26.
The packet assembler 27 converts a series of ATM cells for each user into I
Assemble the P packet 2. On the other hand, the additional identifier determining unit 26
Is the CLP bit notified from the cell monitoring unit 29 =
If "0", generate high-quality type tagging information,
If the CLP bit is “1”, the tagging section 22 generates low-quality type tagging information and notifies the tagging section 22 of this.
In the tagging unit 22, according to the notified tagging information,
The necessary tagging is performed on the TOS field of the assembly packet.

A contract information management unit 25 is provided between the cell monitoring unit 29 and the additional identifier determination unit 26, and the contract information management unit 25 has a CP as shown in FIG.
An L information translation memory may be provided to convert the CLP bit to the quality type. Thus, quality control according to the CLP bit of the ATM cell can be efficiently performed.

FIG. 12 is a diagram for explaining an access operation to a server according to the embodiment. FIG. 12A shows a case where a user A accesses a certain server B with a high quality / low quality type. The relay device B adjacent to the server B
This figure shows a case where the same (high quality / low quality) type quality identifier is automatically set for a response (download) packet from the server B. In the figure, when a user A tries to download data of a server B, a quality identifier is added by the ISP device A according to the present invention to data in the upward direction (packet) from the user A to the server B, Finally, the process is routed to the server B via the relay device B adjacent to the server B. In the present embodiment, in order to guarantee the same quality of service for downlink data (packets) from server B to user A, data (packets) transmitted from server B to user A Repeater B adds the same quality identifier.

FIG. 13 is a diagram for explaining a relay device according to the second embodiment, and shows a case where it is suitable for application to the network system of FIG. FIG. 13A is a block diagram of the relay device 40. In the figure, 44
Is a packet monitoring unit that recognizes header information of an input packet, 45 is a priority identifier tagging unit that adds a required priority identifier to a response packet from a server (adjacent server) adjacent to the relay device, and 46 is the IP of the adjacent server. An adjacent server address management unit for managing addresses, a header information management unit 47 for storing predetermined header information for a packet (uplink packet) passing through the relay device, and an addition 48 for generating additional identifier information for an input (response) packet An identifier determining unit. Other configurations may be the same as those described with reference to FIG.

FIG. 13B shows the contents of the adjacent server management memory. The entry number 1 of the adjacent server includes the IP address “11.10.10.
0.100 "is recorded. If there are other adjacent servers, they are registered after entry number 2. FIG. 13 (C)
Shows the format of the header information memory. here,
The predetermined header information of the upstream packet accessing the adjacent server is recorded. With such a configuration, the quality identifier adding process in the relay device will be described below.

FIG. 14 is a flow chart of the quality identifier adding process in the relay device. When the packet monitoring unit 44 detects the IP address IPAD of the input packet, it inputs to this process. In step S11, the adjacent server address management unit 46 refers to the adjacent server management memory based on the destination address of the input packet, and determines whether or not the destination address corresponds to the adjacent address of the relay device 40. "Destination address of input packet" = "adjacent server address"
In this case, it indicates that the packet is a request (uplink) packet to an adjacent server, and a header information write enable signal W is output to the header information management unit 47. In step S18, the header information management unit 47 determines whether or not the quality identifier of the input packet is a high quality type. If the quality identifier is the high quality type, predetermined header information is recorded in the header information memory in step S19.

Referring to the header information memory of FIG. 13C, here, for example, the IP address “11.10.10.100” of the destination (adjacent server) B of the upstream packet,
The source (user A) address “10.10.10.
10 ", port number" 20 ", high quality setting start number (A
CK number) "5000", window size "800"
0 "and a high quality setting end number (ACK number + window size)" 13000 "are recorded.

Then, the input packet is differentiated and relayed to an adjacent server. Further, in this embodiment, if the quality identifier is not of the high quality type in the determination in step S18, it is not necessary to change the response packet to the high quality type, so the process of step S19 is skipped.

If "destination address of input packet" is not equal to "adjacent address" in step S11, the adjacent server management memory is referred to by the source address of the input packet in step S12. Is determined to correspond to the adjacent address of the relay device 40. If “source address of input packet” = “adjacent address”, it indicates that the packet is a response (downlink) packet from the server, and a header information read enable signal R is sent to the header information management unit 47. Step S1
In 3, the header information management unit 47 reads, from the header information memory, predetermined header information of the upstream packet corresponding to the pair of the source address (server) and the destination address (user) of the downstream packet.

Referring again to the header information memory of FIG. 13C, here, for example, the IP address “11.10.10.10” of the transmission source (adjacent server) B of the downlink packet
0 ”and its destination (user A) address“ 10.10.1 ”
The predetermined record information on the uplink packet corresponding to the pair of “0.10”, ie, port number “20”, high quality setting start number (ACK number) “5000”, window size “8000”, high quality setting end number (ACK number + window size) “13000” is read out.

In step S14, the additional identifier determining unit 4
8 determines whether or not there is registered (corresponding) data in the header information memory, and if so, compares the header information extracted from the downlink packet with the header information read from the header information memory, and in step S15, It is determined whether the numbers match. If they match, it is further determined in step S16 whether or not the sequence number is within (high quality setting start number ≦ downstream sequence number <high quality setting end number). If all of these conditions are satisfied, a high quality type priority identifier is added to the downlink packet in step S17. Otherwise, the process exits without adding the high quality type priority identifier. Thereafter, the downstream IP packet is subjected to a differentiation process and relayed to the user.

FIG. 15 shows the upstream / downstream I input to the relay device.
FIG. 15A shows an example of a P packet, and FIG. 15B shows a downlink packet before adding a quality identifier. Here, the “frame length” is the TCP / I of the packet.
Indicates the payload length excluding the P header (total 40 bytes). "Sequence number" indicates the position in the byte stream of the sender in the segment. The “acknowledgement (ACK) number” indicates the number of the octet that the start point of the communication expects to receive next. The "window field (size)" indicates how much data the packet source terminal can receive in a burst, and the destination terminal requested for data divides the data into a plurality of blocks (packets) if necessary. Divide and send to the source terminal.

FIG. 12B is a sequence diagram of TCP communication between the user A and the server B. For example, if the response confirmation number of user A is 2000 and its window size is 1000, server B sends a packet with sequence number 2000 and a frame length of 500, and a packet with sequence number 2499 and a frame length of 500. Is transmitted two times in total. At this time, the upstream (request) packet from the user A to the server B is sent to the ISP
Apparatus A performs a process of adding a quality identifier, and relay apparatus B performs a process of recording information in a header information memory. For a downstream (response) packet from server B to user A, relay apparatus B performs a process of adding a quality identifier corresponding to the upstream packet. The same applies hereinafter.

Next, the operation of the server access according to the embodiment will be specifically described with reference to FIG. FIG.
When an upstream packet as shown in (A) is input to the relay device 40, its source IP address “10.10.1”
0.10 ", destination IP address" 11, 10, 10, 1 "
00 ”is extracted. Further, the destination IP address “11.10.10.100” of the input packet matches the adjacent server address in the adjacent server management memory, and the quality type TOS = 1 of the input packet is “high quality type” and is higher than IP. If the layer protocol is TCP (protocol = 6), predetermined header information of the input packet is written in the header information memory. Then, the IP packet is differentiated and relayed to an adjacent server.

Thereafter, when a downstream packet as shown in FIG. 15B is input to the relay device 40, the source IP address "11.10.10.100" matches the adjacent server address in the adjacent server management memory. . After confirming the match, the header information memory is searched based on the source (adjacent server) address of the downlink packet and its destination (user) address, and if information matching the two addresses is written, the recorded content is Compare. This comparison includes a match for port number "20", a match for protocol (6),
Sequence number “10000” and frame length “150”
It is determined whether the value obtained by adding "0" does not exceed the high quality setting end number in the header information memory. If these conditions are satisfied, the quality identifier "1" of the high quality type is added to the TOS field of the downlink packet. Is written. Thus, upstream and downstream packets can be communicated with the same quality type.

FIG. 16 is a diagram for explaining an ISP device according to the seventh embodiment. In the two-way communication quality control of FIG. 12, according to the amount of response packets per unit time required by an input (upstream) packet. The figure shows a case where the quality of an uplink packet (and, consequently, its response packet) is set to a high quality type / low quality type. FIG. 16A shows this IS
1 shows a block diagram of a P device 20. A contract is made in advance between the subscriber and the ISP for the amount of high quality setting response packets per unit time required by the upstream packet, and this is stored in the contract information management memory of the contract information management unit 25.

With this configuration, when bidirectional communication is performed, the packet monitoring unit 21 monitors the amount of response packets per unit time requested by the access server for uplink packets for each subscriber. The amount of response packets per unit time is determined by the packet monitoring unit 21 as T
It can be easily measured by monitoring the information in the window field (size) of the CP header. Note that FIG.
(B) shows the format of the TCP header.

The additional identifier determining unit 26 compares the measured response packet amount with the high quality setting response packet amount read from the contract information management memory, and determines the response amount of the data amount not exceeding the high quality setting response packet amount. High quality type identification information is generated for an input (upstream) packet requesting a packet, and low for an input (upstream) packet requesting a response packet having a data amount exceeding the high quality set response packet amount. The identification information of the quality type is generated, and this is notified to the tagging unit 22. The tagging unit 22 adds the notified tagging information to the TOS field of the input (up) packet.

Therefore, in cooperation with the relay apparatus shown in FIG. 13, the high quality of the input (uplink) packet (and, consequently, the response packet) which does not exceed the high quality set response packet amount per unit time in the bidirectional communication. Service can be performed efficiently.

FIG. 17 is a diagram for explaining an ISP device according to the eighth embodiment. In the bidirectional communication quality control shown in FIG. 12, an input (upstream) packet amount per unit time and an input (upstream) packet are required. In this case, the quality of the upstream packet (and thus the response packet) is set to a high quality type / low quality type according to the response packet amount per unit time. FIG. 17 shows a block diagram of this ISP device. The subscriber and the ISP have previously contracted the amount of high quality setting input packets per unit time and the amount of high quality setting response packets per unit time required by input (upstream) packets, and the contract information management unit 25 contract management information memories.

With this configuration, when bidirectional communication is performed, the packet monitoring unit 21 determines whether the number of input (upstream) packets per unit time for each subscriber (source IP address) and the number of upstream packets for each subscriber are And the amount of response packets requested per unit time. The amount of response packets per unit time can be easily measured by the packet monitoring unit 21 monitoring the information of the window field (size) of the TCP header in the uplink IP packet.

The additional identifier determining unit 26 compares the measured input packet amount and response packet amount with the high quality setting input packet amount and the high quality setting response packet amount read from the contract information management memory, and sets the high quality setting value. Generates tagging information that adds a low-quality type quality identifier to input packets that exceed the amount of input packets or input packets that request response data with a data amount that exceeds the amount of high-quality set response packets, and that both exceed Tagging information for adding a high quality type quality identifier is generated for an input packet that has not been input, and this is notified to the tagging unit 22. The tagging unit 22 adds the notified tagging information to the TOS field of the input (up) packet.

Therefore, in cooperation with the relay apparatus shown in FIG. 13, the input (up) packet (up) packet which does not exceed the high quality setting input packet amount and the high quality setting response packet amount per unit time in the bidirectional communication. As a result, the high quality service of the response packet can be efficiently performed.

FIG. 18 is a diagram for explaining an ISP apparatus according to the ninth embodiment. An asymmetrical access line for realizing an asymmetrical access line in the up / down direction is used as the access line for performing the bidirectional communication quality control of FIG. Digital subscriber line (A
DSL (Asymmetric Digital Subscriber Line) is used. In such a configuration, for example, if ADSL of 224 kbps in the uplink direction and 512 kbps in the downlink direction is used, the bandwidth of the response packet to the uplink (request) packet is wide, so that the data download in the access network can be performed without difficulty and efficiently.

By using an ATM line as an access line instead of using ADSL, various settings in uplink / downlink can be set according to the purpose of use of the user. For example, it is possible to set the PCR of the ATM connection or the MCR for the amount of high quality setting response packets with the ISP, and perform communication in cooperation with the down direction of the access line. Can be.

FIG. 19 shows an ISP according to the tenth embodiment.
FIG. 4 is a diagram illustrating an apparatus, and illustrates a case where an input packet is set to a high quality type / low quality type according to an upper layer protocol of an IP packet. FIG. 19A shows a block diagram of this ISP device. In addition, FIG.
This shows the format of the P header.

The packet monitor 21 detects the value of the protocol field in the IP header and notifies the additional identifier determiner 26 of the value. The additional identifier determining unit 26 determines a quality additional identifier according to the notified value of the protocol field, and notifies the tagging unit 22 of this. At this time, for example, when an unreliable UDP packet is to be communicated as a high quality type, the additional identifier determining unit 26
Determines the high quality type when the protocol field is UDP "17", and determines the low quality type when the protocol field is TCP "6".

Then, the tagging section 22 adds the tiling information to the packet, and preferably, a high-priority buffer as shown in FIG. 19C is provided in the differentiating section 23 to add the high quality type. Write the IP packet to the high priority buffer. Then, at the time of buffer reading, control is performed such that packets are written to other buffers with absolute priority, and packets are read from the high-priority buffer. In this regard, the same processing is also performed in the differentiation processing unit 42 of the relay device 40. Thus,
A UDP / IP packet that cannot be discarded is transmitted preferentially by adding a high quality type quality identifier.

Although a plurality of preferred embodiments of the present invention have been described, it is to be understood that various changes in the configuration, control, processing, and combinations thereof can be made without departing from the spirit of the present invention. Not even.

(Supplementary Note 1) In a network system including an ISP device accommodating a subscriber's access network and a relay device for relaying an IP packet of the ISP device on an Internet backbone, an ISP receives an IP packet from the subscriber. A network system, wherein a high quality / low quality type quality identifier is added by a device, and differentiation control according to a quality identifier of an IP packet is performed on a network including a relay device.

(Appendix 2) TCP / IP packet and UD
Mix P / IP packets on the same network,
2. The network system according to claim 1, wherein the ISP device adds a high quality type quality identifier to the subscriber's UDP / IP packet.

(Supplementary Note 3) The ISP device includes a contract information management memory for storing contract information of service quality previously determined with the subscriber, a packet monitoring unit for detecting a source address of an input packet, and a ISP device. An additional identifier determining unit that determines a quality identifier to be added to the input packet by referring to the contract information in the contract information management memory based on the source address obtained, a tagging unit that adds the determined quality identifier to the input packet, 2. The network system according to claim 1, further comprising: a differentiation processing unit that differentiates a subsequent packet according to a quality identifier; and a routing unit that routes the differentiated packet according to a destination address.

(Supplementary Note 4) The network system according to supplementary note 1, wherein the ISP device 20 includes a packet assembling unit 27 for assembling an ATM cell from a subscriber into an IP packet, as shown in FIG. 6, for example. Therefore, A
An access network based on a TM network can be accommodated.

(Supplementary Note 5) The ISP device is characterized by adding a high quality type quality identifier to an IP packet corresponding to a PCR (Peak Cell Rate) of a CBR (Constant Bit Rate) connection in an ATM access network. 4. The network system according to claim 4, wherein Therefore, the data for the PCR guaranteed by the CBR connection can also be guaranteed on the IP network.

(Supplementary Note 6) The network according to supplementary note 4, wherein the ISP device adds a low quality type quality identifier to an IP packet corresponding to a UBR (Unspecified Bit Rate) connection in the ATM access network. system. Therefore, U which is not guaranteed by ATM
BR connection data is not guaranteed even in an IP network.

(Supplementary Note 7) The ISP apparatus adds a high quality type quality identifier to an MCR (Minimum Cell Rate) IP packet of a GFR (Guaranteed Frame Rate) connection in the ATM access network. 5. The network system according to 4. Therefore, data for the MCR guaranteed by the GFR connection can also be guaranteed on the IP network.

(Supplementary Note 8) The ISP device detects a contract information management memory for storing contract information of a high-quality set packet amount per unit time previously determined with the subscriber, and a source address of the input packet. A packet monitoring unit that measures the amount of input packets for each subscriber, and based on the detected source address, compares the measured amount of input packets with the amount of high-quality set packets in the contract information management memory,
An additional identifier determining unit that determines the addition of a high quality type quality identifier to an input packet that does not exceed the high quality set packet amount; a tagging unit that adds the determined quality identifier to the input packet; 2. The network system according to claim 1, further comprising: a differentiation processing unit that performs a differentiation process according to the quality identifier; and a routing unit that routes a packet after the differentiation process according to a destination address.

(Supplementary Note 9) For example, as shown in FIG.
The P device includes a contract information management memory for storing contract information of service quality for each time zone prearranged with the subscriber, a packet monitoring unit for detecting a source address of an input packet, a detected source. An additional identifier determining unit for determining a quality identifier to be added to the current input packet by referring to the contract information for each time zone in the contract information management memory based on the address, and a tagging unit for adding the determined quality identifier to the input packet The network according to claim 1, further comprising: a differentiation processing unit that differentiates the packet after the addition according to the quality identifier; and a routing unit that routes the packet after the differentiation according to the destination address. system. Therefore, the IP packet differentiation control based on the service quality for each time zone prearranged with the subscriber can be efficiently performed.

(Supplementary Note 10) For example, as shown in FIG.
The ISP device includes a contract information management memory for storing contract information of a high-quality set packet amount per unit time for each time zone previously determined with the subscriber, a source address of the input packet, A packet monitoring unit that measures the amount of input packets for each, and, based on the detected source address, compares the measured amount of input packets with the amount of high-quality set packets for each time zone in the contract information management memory. An additional identifier determining unit that determines the addition of a high quality type quality identifier to an input packet that does not exceed the high quality set packet amount; a tagging unit that adds the determined quality identifier to the input packet; A differentiation processing unit that performs a differentiation process according to the identifier; and a routing unit that routes a packet after the differentiation process according to a destination address. The network system according to Supplementary Note 1, symptoms. Therefore, IP packet differentiation control based on the amount of high-quality set packets per unit time for each time zone predetermined with the subscriber can be efficiently performed.

(Supplementary Note 11) As shown in FIG. 11, for example, the ISP device 20 can store a CLP (Cell Loss Prism) of an ATM cell.
iority) corresponding to the high priority / low priority of the bit.
5. The network system according to claim 4, wherein a high quality / low quality type quality identifier is added to the P packet. Therefore, the IP corresponding to the ATM cell differentiation processing
Packet differentiation control can be performed efficiently.

(Supplementary Note 12) A server adjacent to the relay device is provided, the relay device stores the quality identifier added to the upstream IP packet to the server, and stores the quality identifier in the downstream IP packet corresponding to the upstream IP packet from the server. The network system according to claim 1, wherein the same quality identifier is added to the network system.

(Supplementary Note 13) The relay device is, for example, as shown in FIG.
As shown in the figure, an adjacent server management memory that stores the IP address of a server adjacent to the own station, a packet monitoring unit 44 that monitors header information of an input packet, and a destination address of a first input packet that is stored in an adjacent server management memory. A header information memory for storing predetermined header information of the packet by matching the adjacent server address, and a header information memory for storing the source address of the second input packet by matching the adjacent server address of the adjacent server management memory. From the corresponding header information, and determines an additional identifier to be added to the packet.
8, a tagging unit 45 that adds the determined quality identifier to the second input packet, a differentiation processing unit 42 that differentiates the added packet according to the quality identifier, and a packet that has been subjected to the differentiation processing. Supplementary note 1 characterized by comprising a routing unit 43 for routing according to the destination address.
3. The network system according to 2. Therefore, quality control (quality equalization control) of the second downlink packet with respect to the first uplink packet can be performed efficiently.

(Supplementary Note 14) The ISP device includes a contract information management memory for storing information on the amount of high quality setting response packets per unit time previously determined with the subscriber, a source address and a destination address of the input packet. And a packet monitoring unit for measuring a response packet amount per unit time required by an input packet for each subscriber, and a response packet amount measured for the destination address based on the detected source address. And a high quality setting response packet amount of the contract information management memory, and an additional identifier determining unit for determining the addition of a high quality type quality identifier to an input packet of a request not exceeding the high quality setting response packet amount, A tagging unit that adds the determined quality identifier to the input packet, and discrimination that differentiates the added packet according to the quality identifier 13. The network system according to claim 12, further comprising: a conversion processing unit; and a routing unit configured to route the packet after the differentiation processing according to the destination address.

Here, the amount of response packets per unit time requested by the input (uplink) packet is determined by the window field (size) of each uplink IP packet by the packet monitoring unit.
It can be easily measured by monitoring the information of. Therefore, it is possible to efficiently provide high-quality services for response packets that do not exceed the high-quality set response packet amount per unit time.

(Supplementary Note 15) The ISP device includes a contract information management memory for storing information on the amount of high quality setting input packets and the amount of high quality setting response packets per unit time previously determined with the subscriber; A packet monitoring unit for measuring the amount of input packets per unit time of each user and the amount of response packets requested by the input packets per unit time, the measured amount of input packets, the amount of response packets, and a contract information management memory Comparing the high quality setting input packet quantity and the high quality setting response packet quantity of
An additional identifier determining unit that determines the addition of a high quality type quality identifier to an input packet that does not exceed the set amount, a tagging unit that adds the determined quality identifier to the input packet, and a packet after the addition as a quality identifier. 13. The network system according to claim 12, further comprising: a differentiating processing unit that performs a differentiating process in accordance with the packet; and a routing unit that routes the differentiated packet according to a destination address.

(Supplementary Note 16) Asymmetric Digital Subscriber Line (ADSL: Asymmetric Digital Subscr
13. The network system according to supplementary note 12, wherein the network system uses iber line. Therefore, for example, an ADS of 224 kbps in the up direction and 512 kbps in the down direction
When L is used, since the bandwidth of the response packet to the request packet is wide, the data download from the server, which is performed under the differentiation control of the present invention, can be performed without difficulty and efficiently.

(Supplementary note 17) The network system according to supplementary note 12, wherein an ATM line capable of setting different communication speeds and qualities for uplink / downlink lines of the access network is used.

(Supplementary Note 18) The ISP device receives the input IP
A packet monitoring unit for detecting an upper layer protocol in a protocol field of the packet; a tagging unit for adding a high quality type quality identifier to the packet when the UDP / IP protocol is detected; 3. The network system according to claim 2, further comprising: a differentiation processing unit that performs a differentiation process according to the identifier; and a routing unit that routes a packet after the differentiation process according to a destination address.

[0095]

As described above, according to the present invention, the ISP
The device adds a quality identifier to the IP packet and performs differentiation according to the quality identifier on the network, thereby
Network communication can be performed in consideration of quality for each P packet. In the case of two-way communication in which the user downloads data on the Web, the same quality identifier as the data from the subscriber to the server is added to the response data by the relay device adjacent to the server, so that the upstream packet is transmitted. Although the packet arrives with priority and arrives at the server, the downstream (response) packet does not reach the user, so that the user can effectively avoid repetition of retransmission to the access server.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a configuration diagram of a network system according to an embodiment.

FIG. 3 is a diagram illustrating an ISP device according to the first embodiment.

FIG. 4 is a block diagram of a relay device according to the first embodiment.

FIG. 5 is a format diagram of an IP packet.

FIG. 6 is a diagram illustrating an ISP device according to a second embodiment.

FIG. 7 is a format diagram of an ATM cell.

FIG. 8 is a diagram illustrating an ISP device according to a third embodiment.

FIG. 9 is a diagram illustrating an ISP device according to a fourth embodiment.

FIG. 10 is a diagram illustrating an ISP device according to a fifth embodiment.

FIG. 11 is a diagram illustrating an ISP device according to a sixth embodiment.

FIG. 12 is a diagram illustrating an operation of accessing a server according to the embodiment.

FIG. 13 is a diagram illustrating a relay device according to a second embodiment.

FIG. 14 is a flowchart of a quality identifier adding process in the relay device.

FIG. 15 is a diagram illustrating an example of an uplink / downlink packet input to a relay device.

FIG. 16 is a diagram illustrating an ISP device according to a seventh embodiment.

FIG. 17 is a diagram illustrating an ISP device according to an eighth embodiment.

FIG. 18 is a diagram illustrating an ISP device according to a ninth embodiment.

FIG. 19 is a diagram illustrating an ISP device according to a tenth embodiment.

FIG. 20 is a diagram illustrating a conventional technique.

[Explanation of symbols]

 Reference Signs List 10 user terminal 20 ISP device 21 packet monitoring unit 22 tagging unit 23 differentiation processing unit 24 routing unit 25 contract information management unit 26 additional identifier determination unit 27 packet assembling unit 28 clock unit 29 cell monitoring unit 40 relay device 41 priority information detection Unit 42 differentiation processing unit 43 routing processing unit 44 packet monitoring unit 45 priority identifier tagging unit 46 adjacent server address management unit 47 header information management unit 48 additional identifier determination unit 80 server ALC access line accommodation device

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Kazuhiro Fukamachi 3-22-8 Hakata-ekimae, Hakata-ku, Fukuoka City, Fukuoka Prefecture Within Fujitsu Kyushu Digital Technology Co., Ltd. (72) Inventor Ryu Kazuya Fukuoka City, Fukuoka Prefecture Fujitsu Kyushu Digital Technology Co., Ltd. (72) Inventor Masashi Shibata 3-22-8 Hakata Ekimae 3-chome, Hakata-ku, Fukuoka City, Fukuoka Prefecture ) Inventor Shinya Yonemoto 3-22-8 Hakata-ekimae, Hakata-ku, Fukuoka City, Fukuoka Prefecture Inside Fujitsu Kyushu Digital Technology Co., Ltd. (72) Inventor Nobuyuki Umeda 4-1-1 Kamidadanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture Fujitsu Limited (72) Inventor Hiroyuki Sato 4 Kamikadanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture No. 1 No. 1 Fujitsu Limited in the F-term (reference) 5K030 GA03 HA08 HC01 HD03 HD05 KA01 KA05 KA13 KX29 LB05 MB10 5K034 AA01 BB06 FF06 HH01 HH02 HH06 HH18 MM21

Claims (5)

[Claims] In a network system including an ISP device accommodating a subscriber's access network and a relay device for relaying an IP packet of the ISP device on an Internet backbone, an ISP device responds to an IP packet from the subscriber. A network system characterized by adding a quality identifier of a high quality / low quality type and performing differentiation control according to a quality identifier of an IP packet on a network including a relay device. 2. The method according to claim 1, wherein the TCP / IP packet and the UDP / IP packet are mixed on the same network, and the ISP device adds a high quality type quality identifier to the subscriber's UDP / IP packet. 2. The network system according to 1. 3. The ISP device includes: a contract information management memory for storing contract information of service quality previously determined with a subscriber; a packet monitoring unit for detecting a source address of an input packet; An additional identifier determining unit for determining a quality identifier to be added to the input packet by referring to the contract information in the contract information management memory based on the original address; a tagging unit for adding the determined quality identifier to the input packet; The network system according to claim 1, further comprising: a differentiation processing unit that differentiates the packet according to the quality identifier; and a routing unit that routes the packet after the differentiation processing according to the destination address. 4. An ISP device includes a contract information management memory for storing contract information of a high-quality set packet amount per unit time previously determined with a subscriber, and detects a source address of an input packet and joins. A packet monitoring unit that measures the amount of input packets for each user; and, based on the detected source address, compares the measured amount of input packets with the amount of high quality set packets in the contract information management memory, and An additional identifier determining unit that determines the addition of a high-quality type quality identifier to an input packet that does not exceed the packet amount, a tagging unit that adds the determined quality identifier to the input packet, and a packet after addition according to the quality identifier. And a routing unit for routing the differentiated packet according to the destination address. Network system according to claim 1. 5. A server adjacent to the relay device, wherein the relay device stores a quality identifier added to an upstream IP packet to the server, and stores a quality identifier corresponding to the upstream IP packet from the server. The network system according to claim 1, wherein the same quality identifier is added.