KR100237400B1 - Variable-length address handling apparatus - Google Patents

Abstract

The present invention relates to a variable length address processing apparatus for fast packet routing, comprising: a prefix separating means for separating and outputting a fixed length of a fixed length and a subscriber identifier of a predetermined length from an input destination address; Prefix retrieval means for retrieving from the prefix table storing the corrected prefix information according to the network configuration and outputting the corresponding prefix identifier. If the prefix prefix value is assigned to the corresponding packet switch, the prefix identifier and the subscriber identifier Path information retrieval means for retrieving and outputting corresponding route information by combination, and retrieving and outputting route information with prefix identifier value if prefix identifier value is assigned to other packet switch, all assigned to packet switches constituting network Prefix corrected prefix Control means for converting the fix value into a prefix table in the prefix search means and storing route information corresponding to the prefix assigned to another packet switch and each of the subscribers connected to the packet switch in the route information table in the route information search means. By the configuration, a large-capacity packet switch can be configured, can be applied to large-scale networks, and can be applied to a large-capacity switch requiring high-speed processing.

Description

Variable-length address handling apparatus

The present invention relates to a variable length address processing apparatus for high speed packet routing.

1 is a diagram of a connectionless network using a connectionless server in a conventional ATM network, and illustrates an example of a network configuration for providing a connectionless data service.

Switched Multi-magabit Data Service (hereinafter referred to as SMDS) in the case of a network and packet network, Asynchronous Trsnsfer Mode (hereinafter referred to as ATM), instead of the SMD 30 switch and the connectionless server 20 respectively SMDS switch And packet switch are used and applied to SMDS switch and packet switch.

A connectionless network providing a connectionless data service is composed of virtual connections connecting a connectionless server 20 connected to an ATM exchanger 30 and a connectionless connection between a subscriber and a connectionless server 20.

The subscriber divides the packet, which adds the destination address to the header of the information to be delivered to the destination, into a predetermined length segment in the form of ATM Adaptation Layer type 3/4 (hereinafter referred to as AAL 3/4). After that, the ATM header including the connectionless server and the virtual connection number set is added to make an ATM cell and transmitted to the connectionless server 20.

The connectionless server 20 analyzes the destination address of the cellized packet and delivers the destination address to the subscriber or connectionless server connected to the same server.

When one packet is divided into a plurality of AAL 3/4 cells, the segment type value included in the AAL header of the first cell constituting the packet is a leading segment (hereinafter referred to as BOM). The segment type value of the last cell is set as an end segment (hereinafter referred to as EOM), and the segment type value of cells constituting the middle is set as an intermediate segment (hereinafter referred to as COM).

If the packet consists of one cell, the segment type value is set to a single segment (hereinafter, referred to as SSM).

In addition, when a plurality of packets are divided into cells and multiplexed and transmitted through a single virtual connection, a multiplexing identifier (hereinafter referred to as MID) is used in the AAL header of each cell so that the receiver can recover the original packet from the multiplexed cells. It has a realm.

Multiple cells constituting one packet are assigned the same MID value, and when cells constituting multiple packets are multiplexed, cells constituting different packets are assigned different MID values.

The destination address of the packet is composed of 64 bits as shown in FIG. 2. The destination address area includes a 4-bit address type area indicating whether the packet is a single address packet or a group address to be delivered to a plurality of destinations. It consists of an E.164 address range defined by the International Telecommunication Union (ITU) of up to 15 digits, a combination of four bit combination units, each digit of which is between 0-9. It should only have a value of.

If the E.164 address value is less than 15 digits, fill in from the front and all the bits in the unused area at the back fill each digit with "1111" in hexadecimal "F".

The E.164 address is composed of a country code specifying a country, a region code specifying a region, and a subscriber number as shown in FIG. 2, and the country code and region code may be omitted.

The connectionless server 20 should be able to deliver packets entered with these destination address values to the appropriate destination.

In the conventional technology of processing the routing in the connectionless server 20, the address values of all subscribers receiving the connectionless service to all the connectionless servers and the virtual connection number destined for the destination are stored in the routing table, and the packet is disconnected. When entered into the connected server, the entire destination address value of the packet was retrieved from the routing table and delivered to the destination according to the search result.

This method works well when there are a small number of connectionless service subscribers, but if you have a large number of subscribers, each connectionless server must have a very large table, which makes it difficult to quickly retrieve the destination address from the routing table each time a message arrives. In addition, since a complex search algorithm must be applied for a high speed search, there is a problem in that it is difficult to implement in hardware and has a limitation in speed.

In addition, in a network composed of a plurality of connectionless servers, when a subscriber is added or deleted, the routing table of all connectionless servers must be changed.

In another conventional technique, the destination address of a packet input to the connectionless server 20 is divided into a prefix and a subscriber identifier as shown in FIG. 3, and the prefix and the subscriber identifier are first searched for the prefix and the packet is searched. The connectionless server to be delivered is determined, and the final subscriber is determined using the subscriber number when the destination is a subscriber belonging to the connectionless server.

This decision is made by using the fixed length in front of the destination address area of the 64 bit of the input packet as the prefix and the remaining fixed length as the subscriber identifier, so that the prefix can be easily extracted and retrieved and processed by hardware.

However, since connectionless data service subscribers must be assigned the same prefix value, they must have their own address allocation scheme for connectionless data service.

Since the connectionless data network is a virtual network superimposed on the ATM network, it is preferable to use the addressing system of the ATM network as the address assigned to the subscriber of the connectionless service.

In this case, since the connectionless subscribers will have prefixes of various lengths according to the address allocation scheme of the ATM network, the prior art cannot be applied.

Another conventional technique is to search for and process a fixed value of a prefix.

That is, if the area code of "2" appears in the destination address value configured as shown in FIG. 2, the remaining area is a subscriber number of 7 or 8 digits. If the area code of "412" appears, the remaining area is a 6 digit subscriber number. This means that the prefix can be searched according to the pre-assigned addressing system.

Since this prefix search must use search processing that is specific to the specific address system assigned to it, it cannot flexibly cope with changes in the address system and uses different address systems by country and by operator. There is a problem.

In order to solve the above problems, the present invention provides a packet header when a packet is input to a switch in a network consisting of a plurality of packet switches using an E.164 addressing system, for example, a connectionless server, an SMDS switch, and a packet switch in an ATM network. After extracting the prefix and the subscriber identifier from the destination address, the prefix is first searched in the routing table to determine whether the packet should be delivered to the terminal directly connected to the switch or to another packet switch. Direct output path information is searched for the packet to be delivered, and output path information is searched by the prefix search result and subscriber identifier for the packet to be delivered to the terminal directly connected to the switch. The number of items searched on the network When reduced by having a plurality of terminals, and also to simply configure a packet switch for, and can search the route information of a high speed by a hardware and an object thereof is to facilitate the construction of large-capacity packet switch.

1 is a block diagram of a connectionless network using a connectionless server in a conventional ATM network,

2 is a block diagram of a destination address in the form of a conventional E.164 address;

3 is a configuration diagram of an address form structure processed by a conventional packet switch;

4 is a block diagram of an address processing apparatus to which the present invention is applied;

5 is a diagram illustrating a correspondence between corrected prefix and path information in a prefix search means for routing according to the present invention;

6 is a block diagram illustrating a method of creating a corrected prefix and a subscriber identifier according to the present invention;

7 is a block diagram of a prefix separation unit according to the present invention;

8 is a block diagram of a connectionless server to which the present invention is applied.

<Explanation of symbols for main parts of the drawings>

10: subscriber 20: connectionless server

21: header replica unit 22: delay unit

23: header conversion unit 24: header conversion table

25: address processing section 26: MID generating section

30: ATM exchange

40: destination address in the form of an E.164 address

41,61: Address Type 42,63: Country Code

43,64: area code 44,65: subscriber number

45: unused area 50: destination address processing device

51: prefix separation unit 51a: local prefix area

51b: remote prefix area

51c, 51d, 51e, 51f: Digit selector

51g: priority encoder 52: prefix search unit

53: route information search unit

53a, 53b, 53c, 53d, 53e: zone 1, 2, 3, 4, 5

54 control unit 60 corrected prefix

62: Hexadecimal "F"

In order to achieve the above object, the present invention provides a prefix separation unit for separating and outputting a fixed length of a fixed prefix and a subscriber identifier of a predetermined length from an input destination address, and storing the corrected prefix information when the corrected prefix is input. Prefix search unit for retrieving from the prefix table and outputting the corresponding prefix identifier, and if the input prefix identifier value is assigned to the corresponding packet switch, the corresponding route information is searched and output by the combination of the prefix identifier and the entered subscriber identifier. If the prefix identifier value is assigned to another packet switch, the route information search unit retrieves and outputs route information using the prefix identifier value, and converts all prefixes assigned to the packet switches constituting the network to the corrected prefix value. Prefix frame in prefix search section And stored in a block, characterized by comprising a path information corresponding to each of the subscriber associated with the prefix and the packet switch allocated to the other packet switch to the control unit for storing the path information table in the path search portion.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

4 is a block diagram of a destination address processing apparatus to which the present invention is applied, a prefix separation unit 51 for outputting a fixed length and a subscriber identifier having a predetermined length, and a prefix for outputting a prefix identifier corresponding to information storage of the corrected prefix; The search unit 52, a route information search unit 53 for searching and outputting route information by the input prefix identifier value, and a controller for providing and controlling the corrected prefix information and route information by the prefix search unit and the route information search unit. It consists of 54.

4 is a prefix separating unit 51 which separates and outputs a fixed length of a fixed prefix and a subscriber identifier of a predetermined length from the input destination address, and corrected prefix information according to the network configuration when the corrected prefix is input. A prefix search unit 52 for searching a prefix table for storing the prefix prefix and storing the corresponding prefix identifier, and if the input prefix identifier value is a value assigned to the corresponding packet switch, the corresponding route is a combination of the prefix identifier and the input subscriber identifier. If the prefix identifier value is a value assigned to another packet switch, the route information retrieval unit 53 for retrieving and outputting route information using the prefix identifier value and all prefixes assigned to the switches constituting the network are retrieved. It is converted to the corrected prefix value and converted into the prefix table in the prefix search unit. The controller 54 stores the prefix assigned to the other packet switch and the route information corresponding to each of the subscribers connected to the packet switch in the route information table in the route information retrieval unit.

The controller 54 also converts the prefix value into a corrected prefix value and stores the prefix value in the prefix table in the prefix search unit. However, the controller 54 has a very small frequency and does not require high-speed processing because it is initialized or changed every time the network configuration is changed. There is no need to configure.

5 is a diagram illustrating a correspondence between the corrected prefix in the prefix search unit for routing and the path information according to the present invention, and the corrected prefix in the prefix search unit 51 and the route information table in the path information search unit 53. This example shows the correlation with.

The prefix table in the prefix search unit 51 is divided into a local prefix area 51a in which the correction values of the prefixes assigned to the packet switch are stored, and a remote prefix area 51b in which the correction values of the prefixes assigned to the other packet switches are stored. .

Each of the corrected prefixes stored in the local prefix area 51a corresponds to a portion of the route information table in which subscribers with the prefix value can store route information, and each route information in the region corresponds to each subscriber identifier value. Corresponding.

However, each of the corrected prefixes stored in the remote prefix area 51b corresponds one-to-one with entries in a certain area of the route information table.

This is because a prefix assigned to another packet switch needs only one path information destined for the packet switch to which the prefix is assigned.

FIG. 6 is a diagram illustrating a method of creating a corrected prefix and a subscriber identifier according to the present invention, and illustrates a method of separating the corrected prefix and the subscriber identifier of a predetermined length by the prefix separating unit 51.

That is, the digit of the unused area is filled with the hexadecimal "F" in the 16-digit destination address, and thus the value of the unused area can be known.

Therefore, the fixed length subscriber identifier should be easily extracted, and the prefix of the variable length may be easily corrected by the prefix retrieval unit 52.

In the present invention, the effective address area excluding the address type area is aligned to the right side of the 16-digit address area, and a blank digit between the address type area and the E.164 address area generated by right-alignment is not used in the E.164 address value. In other words, hexadecimal digits are filled with "A" through "F".

In the example of FIG. 6, a corrected destination address value filled with " F " 62 is created, and a lower predetermined length of the corrected destination address value is used as the subscriber identifier and the entire remaining area is used as the corrected prefix.

7 is a configuration diagram of the prefix separation unit according to the present invention, and shows a configuration example of the prefix separation unit 51 which extracts the subscriber identifier and the corrected prefix by making a corrected destination address from the packet's destination address.

The prefix separating unit 51 outputs an input line of the highest priority in which " 1 " is input among 15 input lines, and a priority encoder for outputting a number (input of a high number in FIG. 7 indicates a high priority). 4 input AND for inputting the digits 1 to 14 of the input address value to "1" if the digit value is "F", or "0" if the value is anything other than 2 to 15 of the priority encoder. The gate and the digit selectors 51c, 51d, 51e, and 51f which output one digit indicating a digit select signal value among the 15 input digit values.

In the first input of the priority encoder 51g, a value of "1" is always input.

The digit selector for output digit 1 connects digit 1 to digit 15 to 15 inputs, respectively, but the digit selectors 51c, 51d, 51e, and 51f for output digits 2 to 15 each have one input digit. Input that is shifted and the digit value is not connected always inputs an unused value in the E.164 address value. In the example of FIG.

8 is a configuration diagram of a connectionless server to which the present invention is applied, and the connectionless server processes a message cellified in AAL3 / 4 on a cell basis.

That is, a plurality of cellized messages are multiplexed and inputted cells have a virtual path identifier (BEGIN) of a message including a message address or a single segment message while passing through a header replication unit 21. A virtual path identifier (hereinafter referred to as a VPI), a virtual channel identifier (hereinafter referred to as a VCI), a MID, and a destination address value, and transferring the destination address value to the address processor 25 to which the present invention is applied. The VCI and MID values are stored in the search area of the header conversion table 24, and the VPIo, VCIo corresponding to the destination address output from the destination address processing unit 25, and the MIDo output from the MID generation unit are identically located in the header conversion table. To the output area.

The MID generation unit 26 receives VPIo and VCIo and allocates and outputs MIDo values not used in the virtual connection.

All cells pass through the header replica 21 without being altered and are cut into the delay section 22.

The delay unit 22 temporarily processes the cells until the destination address of the message constituting the cell passing through the delay unit is processed by the address processing unit 25 so that the corresponding VPIo, VCIo and MIDo are stored in the header conversion table 24. Save function.

In other words, the header is delayed until the header converter 23 prepares information for header conversion, and then transferred to the header converter.

The header converter 23 retrieves the VPI, VCI, and MID values of the input cell headers from the search area of the header conversion table 24, and reads the VPIo, VCIo, and MIDo values from the output area of the corresponding position. After replacing the cell header, output the cell.

In addition, if the processed cell is a single segment message, which is the end of the message constituting the end of the message or a single cell message, the header used for the cell header conversion in the header conversion table 24 Remove from conversion table.

The above operation reads the destination address from the cell containing the destination of the message and changes the cell header value to a MIDo value for distinguishing a message multiplexed on the same virtual connection as the VPIo and VCIo, which are virtual connection numbers directed to the destination. Cells constituting the can be delivered to the destination terminal.

As described above, the present invention is applied to a connectionless server, SMDS switch or packet switch constituting a large-scale network, and when a packet is input, fixed length subscriber identifiers and variable length prefixes are fixed in the destination address area of the packet. It is possible to easily configure a large-capacity packet switch by extracting the prefix value and using the corrected prefix and the subscriber identifier, and by searching the route information to which the packet is delivered at high speed.

Also, if the destination is identified as a subscriber connected to the corresponding packet exchange in the first step and the first step of searching for a packet switch to which a packet is delivered with only the corrected prefix, the route information to the subscriber is searched by the prefix identifier and the subscriber identifier value. Since the search process is simple because it consists of a second step, it can be easily applied to large networks.

And since the entire process of extracting the subscriber identifier and the corrected prefix from the destination address value and retrieving the route information from the extracted subscriber identifier and the corrected prefix value can be easily configured by hardware, it can be applied to a large capacity switch requiring high speed processing. There is an advantage to that.

Claims (5)

A packet switch for searching for route information directed to a destination by using a destination address of an input packet used in a network to which subscriber address values of various lengths are allocated and delivering the packet according to the route information, A prefix separator for separating and outputting a fixed length of a fixed length and a subscriber identifier of a predetermined length from the input destination address; A prefix retrieving unit for retrieving a prefix table that stores the corrected prefix information when the corrected prefix is input and outputting a corresponding prefix identifier; If the prefix prefix value is the value assigned to the corresponding packet switch, the corresponding route information is searched and output by the combination of the prefix identifier and the input subscriber identifier. A route information search unit for searching for and outputting information; All prefixes assigned to the packet switches constituting the network are converted into corrected prefix values and stored in the prefix table in the prefix search unit, and the paths corresponding to the prefixes assigned to other packet switches and the subscribers connected to the corresponding packet switches. And a control unit for storing the information in the route information table in the route information retrieval unit. The method of claim 1, wherein the prefix separation unit Four-input ADN for inputting the first two digits from the digit 1 to the digit 14 of the input address value, "1" if the digit is "F", or "0" if the value is other than A gate; A priority encoder for outputting a number of the high priority input line to which "1" of the 15 input lines is input; And a digit selector for outputting one digit value indicated by the digit select signal value among the 15 input digit values. The method of claim 2, In the first input of the priority encoder, a value of “1” is always input. The digit selector for output digit 1 connects digits 1 to 15 to each of the 15 inputs, but the digit selectors for output digits 2 to 15 shift one input digit, respectively. A variable-length address processor as described in that an input with no digit value always inputs an unused value in an E.164 address value. The method of claim 1, wherein the prefix search unit A local prefix area for storing the corrected values of the prefix assigned to the packet switch; And a remote prefix area for storing the corrected values of the prefix assigned to another packet switch. The method of claim 4, wherein And a result of which the searched result is a value in a local prefix area or a value in a remote prefix area.

Images