JP2000031989A - Information repeater - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the reliability of information communication by label switch routing. SOLUTION: In a label switch router 100 that conducts label switch routing, plurality of data relay control sections 101 consisting of an active system data relay control section 101-a and standby system data relay control section 101-b connect to an asynchronous transfer mode ATM switch 102, control information and control logic to switch the active system data relay control section 101-a into the standby system data relay control section 101-b are given to an ATM switch control section 104 of an ATM switch 102 and the reliability of the label switch router 100 due to redundancy of the data relay control section 101 connecting to the ATM switch 102 is enhanced. Furthermore, the utilizing efficiency of the communication path is enhanced by assigning dynamically logical connection information of a VPI/VCI or the like to plurality of the data relay control section 101 to operate them in parallel.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to information relay technology, and more particularly, to IP (Internet Protocol).
l) Among network devices used for network construction, ATM (Asynchronous Tran)
The present invention relates to a technology that is effective when applied to a label switch router or the like that performs cut-through transfer of an IP datagram using a sfer mode (switch) switch or the like.

[0002]

2. Description of the Related Art WWW (World Wide Wed)
With the spread of applications such as b), the amount of information flowing on the Internet continues to increase, and its communication performance,
In particular, the transfer performance of a router, which is a network device that transfers IP datagrams between adjacent networks, has been regarded as a problem as a performance bottleneck. This is because the transfer process requires a process of confirming the transfer destination for each of the IP datagrams.

As means for solving the problem, Japanese Patent Laid-Open No.
As described in Japanese Patent Application No. 1255692, a method for realizing high-speed packet transfer utilizing an ATM switch has been devised. If such a reference technique is simply and concretely shown, it is a combination of an ATM switch as shown in FIG. 8 and a control device for transferring and controlling an IP datagram.

In FIG. 8, reference numeral 901 denotes a data relay control unit for transferring and controlling an IP datagram, detecting an IP datagram which satisfies a specific condition in the transferred IP datagram, and controlling an ATM switch based on the result. , 902 are ATM switches for switching ATM cells, and 903 is an adjacent node connected via an ATM line.

[0005] Here, the adjacent node 903 and the data relay control unit 901 send a standard VC (virtual channel) route (default VC route) for exchanging IP datagrams to the ATM switch 902 at an arbitrary VP (virtual path). Set it on the route. Then, the IP datagram is usually sent on the default VC route relayed by the data relay control unit 901. Thereafter, upon detecting an IP datagram satisfying a certain condition, the data relay control unit 901 causes the IP
A VC path (cut-through VC path) for directly connecting a datagram transfer source and transfer destination adjacent node 903 via an ATM switch 902 is set. Then, the data relay control unit 901 exchanges IP datagrams between the adjacent nodes 903 on the cut-through VC route that is not relayed.
As a result, the individual I
Communication can be performed without confirming the transfer destination of the P datagram and without performing the transfer process, thereby achieving a dramatic improvement in the transfer performance of the IP datagram.

[0006] As a specific method, IETF
(Internet Engineering Tas
k Force) RFCs (Requests for
Comments (Comments) 2098.
Switch Router) and Tag Switching described in RFC2105.

The communication path in the ATM is set using the VC path and the VP path that bundles a plurality of VC paths as described above. Also, originally, the adjacent node 903 is another network that can communicate only via the data relay control unit 901, but since IP datagrams are directly exchanged between the adjacent nodes 903 via cut-through VC routes, In the node 903, a mechanism for encapsulating an original IP datagram into an IP datagram for the adjacent node 903 and sending the data is mounted.

[0008]

According to the configuration of the prior art described above, if a failure occurs in the data relay control unit connected to the ATM switch, even if there is no problem in the ATM switch, data cannot be transmitted between adjacent nodes. There was a technical problem that communication between the members became impossible.

An object of the present invention is to provide an information relay technology capable of stably relaying communication data by quick recovery from a failure.

Another object of the present invention is to provide an information relay technology capable of improving the reliability of information communication by label switch routing.

Another object of the present invention is to provide an information relay technology capable of improving the use efficiency of a communication path in information communication by label switch routing.

[0012]

According to the present invention, a first relay means for relaying first communication data in a first information communication network, and a plurality of first relay means connected to the first relay means. A second relay unit that relays second communication data carried by the first communication data on a logical second information communication network constructed on the first information communication network configured by the relay unit; Relaying means that satisfies a predetermined condition in the first communication data without passing through the second relaying means.
In the information relay apparatus relaying only by the relay means, a plurality of second relay means and at least one of the plurality of second relay means are selected and logically or physically connected to the first relay means. An operation of relaying the second communication data on the second information communication network as a connection by connection, an operation of monitoring the operation of the second relay means in use, a second operation of the second use
When the second relay means fails, the second relay means is selected and logically or physically connected to the first relay means to relay the second communication data in the second information communication network. And a redundancy control means for executing an operation to be continued.

Further, according to the present invention, a first relay means for relaying first communication data in a first information communication network,
A logical second network connected to the first relay unit and constructed on a first information communication network composed of a plurality of first relay units.
And second relay means for relaying the second communication data carried by the first communication data on the information communication network, wherein the first communication data satisfying a predetermined condition is included in the first communication data. In an information relay device relaying only by the first relay unit without passing through the second relay unit, a plurality of second relay units connected to the first relay unit and the first relay unit Redundancy control for operating a plurality of second relay means in parallel by dynamically allocating a plurality of logical communication paths set on the first information communication network to individual second relay means. Means.

More specifically, as an example, a plurality of data relay control units (label switch routing controllers) for relaying IP datagrams can be connected to an ATM switch, and a data relay control unit is provided. Storage means for setting and storing connection information such as the ATM port of the ATM switch to which is connected and the used VC, and an operation for investigating the data relay control unit actually connected from the information stored in the storage means. Any one of the plurality of data relay control units detected as a result of this investigation
The operation of setting the status of the actual system to the active system and the status of the other to the standby system, monitoring the operation of the data relay control unit set and operating in the active system by this setting operation,
When the failure is detected, all VCs in the ATM switch set by the failed data relay control unit are deleted, and another standby data relay control unit replaces the new active data relay control unit with one. And a control logic for executing an operation of selecting a platform.

Further, a means for dynamically allocating a plurality of VPIs to a plurality of data relay controllers connected to the ATM switch is provided, and control logic for operating the plurality of data relay controllers in parallel is provided.

According to the present invention, a data relay control unit for performing cut-through control of a transferred IP datagram can be made redundant in a label switch router or the like. Can be continued. Therefore, a user using the adjacent node is not affected by the failure of the data relay control unit. Further, the reliability can be improved in proportion to the number of data relay control units connected to the ATM switch.

Furthermore, since a virtual route to a plurality of data relay control units can be set on a communication line between adjacent nodes, the communication route including the data relay control unit is made redundant, Efficient use becomes possible.

[0018]

Embodiments of the present invention will be described below in detail with reference to the drawings.

(Embodiment 1) FIG. 1 is a conceptual diagram showing an example of a configuration of a label switch router which is an embodiment of an information relay device of the present invention.

In FIG. 1 of the present embodiment, as an example of an information relay device, a configuration of a label switch router 100 in which a control unit for performing cut-through control of a transferred IP datagram is made redundant, and its components 2 shows an example of the internal structure of the ATM switch.

In FIG. 1, a label switch router 10
0, 101-a and 101-b are for transferring and controlling IP datagrams, and for transferring IP datagrams.
A data relay control unit functioning as a label switch routing controller for detecting a datagram that satisfies a specific condition and controlling an ATM switch (such as cut-through transfer control) based on the detection result;
Reference numeral 102 denotes an ATM switch for switching ATM cells.

Among the elements constituting the ATM switch 102, reference numeral 103 denotes an ATM switching mechanism for transferring cell data actually flowing at high speed by hardware, and 104 denotes control of the ATM switch 102 or data relay control. An ATM switch control unit 105 for processing an instruction from the unit 101 is a line I / F accommodating a connection line 105a to the data relay control unit 101 or another network device.
It is.

The connection line 105a is, for example, a conductor,
An arbitrary medium such as an optical fiber and a radio wave can be used.

Here, as exemplified in FIG. 1, the ATM switch 102 of the present embodiment has a plurality of line I / Fs.
105 can be connected to and accommodated in the ATM switching mechanism 103.

Further, as an element constituting the ATM switch control unit 104, a CPU (Centr) 106 that executes control software for performing ATM switch control
alProcessing Unit), 107 is CP
ROM (Read Only Mem) for fixedly storing control software and initial information operated in U106
ory) 108 is a RAM (R) for dynamically storing the work information of the control software and a table to be described later.
An MMI (Man-Mac) 109 for providing a user with setting means such as an operation state of an ATM switch and contents of a table described later.
hine Interface).

Although FIG. 1 shows a redundant configuration in which two data relay control units 101 are connected for the sake of simplicity of illustration, the present invention does not limit the number of connected units and three or more units are connected. It is also applicable to a redundant configuration, and the following processing is performed for three or more data relay control units 10.
1 is assumed. Further, in each data relay control unit 101 assuming a redundant configuration, it is a premise of redundancy that the same information is set in Internet-related setting information such as an IP address and a routing protocol.

Next, the RA of the ATM switch control unit 104
An example of control information set on M108 and used in the label switch router 100 of the present embodiment will be described with reference to FIG.

FIG. 2 shows the M of the ATM switch controller 104.
FIG. 3 is a diagram showing a configuration of a data relay control unit connection port management table 200 which stores a port number of an ATM switch connecting the data relay control unit 101 set via the MI 107 and various information at the time of connection.

In FIG. 2, reference numeral 201 denotes a main table 200.
, An item number field for storing a number for identifying and managing information for each data relay control unit 101 connected to the ATM switch 102, a port number field 202 for storing a port number for connecting the data relay control unit 101,
An operation status field 203 stores the status of the data relay control unit 101 connected to the port stored in the port number field 202, and a VP or VC 204 used by the connected data relay control unit 101. Information (VPI: Virtual Path Identi)
fier and VCI: Virtual Channel
This is a connection information field for storing various connection information such as an Identifier.

Here, among the values of the data relay control unit connection port management table 200, the serial number in ascending order is stored in the item number field 201 at the time of generation. And
Port number field 202 and connection information field 20
4 stores a value input by the user using the MMI 109 of the ATM switch 102. Further, the operation status field 203 includes a data relay control unit detection process 60 described below.
At 0, an appropriate value is stored. Further, in fields other than the item number field 201, when connection information is not registered, "NULL" indicating non-storage is stored.

Next, a processing procedure for detecting the data relay control unit 101 connected to the ATM switch 102 using the data relay control unit connection port management table 200 will be described with reference to the flowchart of FIG.

The ATM switch 102 executes the data relay control unit detection processing 600 after restarting by turning on the power supply and pressing the reset switch. In the data relay control unit detection processing 600, first, information registered in the data relay control unit connection port management table 200 is stored in the port number field 202 corresponding to the item number indicated by the item number field 201 in ascending order. (Step 601). Then, the data link of the relevant port is established, and the data relay control unit 10 usable on the relevant port is communicated with the data relay control unit 101.
It is checked whether 1 exists (step 602). As a result, when the data relay control unit 101 is detected in the corresponding port, the port
It is checked whether or not the port has detected (step 603).
If so, the "working" state information is stored in the corresponding operation state field 203 of the data relay control unit connection port management table 200 (step 604).

On the other hand, the data relay control unit 10
If no “1” is detected, “unconnected” state information is stored in the corresponding operation state field 203 (step 6).
05). If the presence of the data relay control unit 101 is detected but the port is not the first detection port, the “active” state information is stored in the corresponding operation state field 203 (step 606). Then, it is determined whether information is stored in the port number field 202 corresponding to the next item number (a value other than “NULL”) (step 607).
If another information is stored, the above-mentioned steps 601-6
Step 06 is repeated. If there is no stored information, the process ends.

With the above processing, the ATM switch 102 selects one of the connected data relay control units 101 as the active system that actually operates and the other as the standby system.
It becomes manageable. Then, the ATM switch 102 receives only the processing request from the data relay control unit 101 which has become the active system by the above processing, sets the communication path such as VP / VC, and sets the communication path from the other standby system data relay control unit 101-b. Is ignored or the communication is rejected, and only the active data relay control unit 101-a can operate normally.

Further, in the following, referring to FIGS. 4 and 7, when a failure occurs in the data relay control unit operating as the active system, one of the standby data relay control units 101-b is used by the active system. The processing procedure up to the operation as a system will be described.

FIG. 4 shows V set in the information communication system including the label switch router 100 of the present embodiment.
FIG. 7 is a conceptual diagram showing an example of a P route and a default VC route. FIG. 7 performs failure detection of the active data relay control unit 101-a and switching of the active system to another standby data relay control unit 101-b. Active system data relay controller failure processing 7
9 is a flowchart illustrating an example of the 00.

When the active data relay control unit 101-a is operating normally, as shown in FIG. 4A, for example, the label switch router 1 shown in FIG.
Adjacent node 80 composed of a relay device or the like having the same configuration as 00
0 and the VP of the port connected to the active data relay control unit 101-a.
A default VC route is set only for the route, and a state is set in which communication between adjacent nodes 800 is possible only through the active data relay control unit 101-a.

Here, the ATM switch control unit 104 of the ATM switch 102 periodically executes the active data relay control unit failure processing 700 to monitor the status of the active data relay control unit 101-a. An example of the processing procedure will be described below.

In the active system data relay control unit failure processing 700, first, the active system data relay control unit 101-a registered in the data relay control unit connection port management table 200 is periodically exchanged with an alive signal or the like. It is checked whether or not it is operating normally (step 70).
1). As a result, when a failure of the active data relay control unit 101-a is detected, “standby” is displayed in the operation state field 203 of the data relay control unit connection port management table 200.
Another data relay control unit 10 in which status information is stored
1, that is, whether or not there is a standby data relay control unit 101-b (step 702).

Then, the standby data relay control unit 101-
If b exists, the value of the operation state field 203 in the data relay control unit connection port management table 200 corresponding to the active data relay control unit 101-a is changed to “not connected” (step 703), and then a failure occurs. The communication route (VC route) in the ATM switch 102 set in response to the request from the working data relay control unit 101-a is deleted (step 704). After that, any one of the standby data relay control units 101-b in the “standby” operation state registered in the data relay control unit connection port management table 200 is selected.
A table is selected (for example, the one found first in the search from the top of the table 200), and the value stored in the corresponding operation state field 203 is changed to the "current" operation state (step 705).

On the other hand, if no failure has occurred in the active data relay control unit 101-a in the check in step 701, and a failure has occurred in the active data relay control unit 101-a in the check in step 702. If the standby data relay control unit 101-b is not connected to another port, the process 700 ends without performing any processing.

By the above-described processing 700, the ATM switch 102 sets the active data relay control unit 10 in which the failure has been detected.
The operation status of 1-a can be changed from "working" to "not connected", and the working data relay control unit 101-a in which a failure has occurred.
It is possible to protect itself from malfunctions due to Also,
One of the standby data relay control units 101-b connected to the ATM switch 102 can be selected and managed as the active system thereafter. Here, if it is desired to more strictly disconnect the faulty data relay control unit, the ATM switch 102
May process to close the port to which the fault data relay control unit is connected. Also, the data relay control unit 101
The active / standby state and the port closing may be managed in cooperation (that is, the active connection port is opened and the standby connection port is closed). After that, the ATM switch 102 receives only the processing request from the data relay control unit 101 which has become the active system by the above processing, and sets a communication path such as VP / VC. On the other hand, the data relay control unit 101 changed from the standby system to the active system can accept the request for setting the communication path, which has been rejected until now, and sets the communication path such as the default VC. As a result, as illustrated in FIG. 4B, a request from another standby data relay control unit 101-b other than the selected one standby data relay control unit 101-b is ignored or communication is performed. , Only the selected one standby data relay control unit 101-b can normally operate as the new active data relay control unit 101.

According to the above processing, when a failure occurs in the active data relay control unit 101-a, as shown in FIG. 4B, an arbitrary VP route of each port connected to the adjacent node 800 is established. A default VC route is set for the VP route of the port to which the standby data relay control unit 101-b (currently the active system) is connected, and the adjacent node is set via the new active data relay control unit 101-a. 800 is set in a state where communication is possible between adjacent nodes 80.
Communication between 0 can be continued.

As described above, the AT to which the present embodiment is applied
When the M switch 102 is used, a plurality of data relay control units 101 are connected to the ATM switch 102 in advance, so that even when a failure occurs in the active data relay control unit 101-a, the standby data relay control unit 101-a is connected. Control unit 101-b
And the communication can be continued between the adjacent nodes 800 with the ATM switch 102 interposed therebetween.
The switching of the data relay control unit 101 is repeated by the number of connected standby data relay control units 101-b.

As described above, by using the ATM switch 102 of the present embodiment, the label switch router 1
Since the data relay control unit 101 that performs cut-through control of the transfer IP datagram can be made redundant at 00,
Even when a failure occurs in the data relay control unit 101, the transfer processing of the IP datagram can be continued. Therefore, the user using the adjacent node 800 is not affected by the failure of the data relay control unit 101 in the label switch router 100. By increasing the number of data relay control units 101 connected to the ATM switch 102, the data relay control unit 10
It is possible to further improve the reliability of the label switch router 100 in proportion to the number of additions of 1.

(Embodiment 2) Next, a label switch router 100A as another embodiment of the information relay apparatus of the present invention will be described with reference to FIGS. In the following description of the label switch router 100A, portions different from the first embodiment described above will be described, and the description of the same components will be omitted.

FIG. 3 shows the M of the ATM switch control unit 104.
FIG. 5 shows another example of the configuration of the data relay control unit connection port management table 300 which stores the port number of the ATM switch connecting the data relay control unit 101 set via the MI 107 and various information at the time of connection. FIG. 14 is a diagram showing a VP route and a default VC route set in a system according to another embodiment of the present invention.

First, the RA of the ATM switch control unit 104
Data relay control unit connection port management table 3 as control information used in the second embodiment on M108
00 will be described with reference to FIG.

In FIG. 3, reference numeral 301 denotes the main table 300.
, An item number field for storing a number for identifying and managing information for each data relay control unit 101 connected to the ATM switch 102A;
A port number field for storing a port number for connecting to a VPI; a VPI offset field 303 for storing a value to be added to a VPI specified when a communication path is actually set in a communication path setting request from the data relay control unit 101; 304 is a connected data relay control unit 1
01 (VPI: Virtu)
al PathIdentifier and VCI: Vir
This is a connection information field in which various connection information such as a dual channel identifier (Tual Channel Identifier) is stored.

Here, among the values of the table 300, the serial number in ascending order is stored in the item number field 301 at the time of generation. Other fields are ATM switch 102
The value input by the user using the MMI 109 of A is stored. The ATM switch 102A is connected to the table 3
When processing a communication path setting request from the data relay control unit 101 in accordance with the value stored in the VPI offset field 303 of 00, the data relay control unit 101
In the actual setting for the port to which no connection is made, the setting is made using the value obtained by adding the value of the VPI offset field 303 to the requested VPI value.

As a result, as shown in FIG. 5A, the (system 1) data relay control unit 101-a and the (system 2) data relay control unit 101-b perform communication between the adjacent nodes 800. By using another VPI on the line, it becomes possible to share the communication line and operate simultaneously. Therefore,
The route between adjacent nodes 800 is subjected to Internet routing control. When a failure occurs in the (1st) data relay control unit 101-a, the failed (1st) data relay control unit 101-a is separated from the routing path by the Internet routing control technology, and the (2nd system) ) The communication between the adjacent nodes 800 is continued only via the data relay control unit 101-b.

As described above, A to which the second embodiment is applied
When the TM switch 102A is used, a plurality of communication paths between adjacent nodes 800 can be provided by connecting a plurality of data relay control units 101 to the ATM switch 102A.

That is, ATM using the second embodiment
In the switch 102A, it becomes possible to set a virtual route to the plurality of data relay control units 101 on the connection line 105a between the adjacent nodes 800.
In addition to the redundancy of the communication path including the communication path 1, the communication path can be used efficiently.

In the above two embodiments, the plurality of data relay control units 101 are connected to the A through the line I / F 105.
In the above description, an example of a configuration externally connected to the TM switch 102 (102A) has been described.
ATM with a configuration built in M switch 102 (102A)
Also in the case of a switch, the setting value of the port number field 202 or 302 of the data relay control unit connection port management table 200 or the data relay control unit connection port management table 300 is information indicating the internal mounting position (for example, slot number). ) Can be applied simply by changing and managing.

Although the invention made by the inventor has been specifically described based on the embodiments, the present invention is not limited to the above-described embodiments, and can be variously modified without departing from the gist thereof. Needless to say, there is.

For example, the first information communication network is not limited to the ATM communication network exemplified in the above embodiment, and any data link technology such as frame relay, SONET, and Ethernet can be used. .

[0057]

According to the information relay apparatus of the present invention, there is obtained an effect that communication data can be stably relayed by quick failure recovery.

Also, the effect that the reliability of information communication by label switch routing can be improved can be obtained.

Further, there is an effect that the use efficiency of the communication path in the information communication by the label switch routing can be improved.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram showing an example of a configuration of a label switch router as an embodiment of an information relay device of the present invention.

FIG. 2 is a conceptual diagram illustrating an example of a configuration of control information used in a label switch router according to the first embodiment of the information relay device of the present invention.

FIG. 3 is a conceptual diagram showing an example of a configuration of control information used in a label switch router according to a second embodiment of the information relay device of the present invention.

FIGS. 4A and 4B are conceptual diagrams illustrating an example of an operation of the label switch router according to the first embodiment of the information relay device of the present invention.

FIGS. 5A and 5B are conceptual diagrams illustrating an example of an operation of a label switch router according to a second embodiment of the information relay device of the present invention.

FIG. 6 is a flowchart illustrating an example of an operation of the label switch router according to the first embodiment of the information relay device of the present invention.

FIG. 7 is a flowchart illustrating an example of an operation of the label switch router according to the first embodiment of the information relay device of the present invention.

FIGS. 8A and 8B are conceptual diagrams showing the configuration and operation of the reference technology of the present invention.

[Explanation of symbols]

100, 100A ... label switch router (information relay device), 101 ... data relay control unit (second relay means),
101-a: working system (system 1) data relay control unit, 101
-B: standby system (system 2) data relay control unit, 102, 10
2A: ATM switch (first relay means), 103: A
TM switching mechanism, 104: ATM switch controller, 105: line I / F, 105a: connection line, 106
... CPU, 107 ... ROM, 108 ... RAM, 200 ...
Data relay control unit connection port management table 201 201 item number field 202 202 port number field 203
... operation state field, 204 ... connection information field,
300: data relay control unit connection port management table, 3
01 ... item number field, 302 ... port number field, 303 ... VPI offset field, 304 ... connection information field, 800 ... adjacent node.

Claims (3)

[Claims] 1. A first relay means for relaying first communication data in a first information communication network; and a first relay means connected to the first relay means and constructed on the first information communication network. A second relaying second communication data carried by the first communication data on a logical second information communication network;
An information relay device that relays, through the second relay unit, only the first communication data that satisfies a predetermined condition, without passing through the second relay unit. A plurality of second relay means, and at least one of the plurality of second relay means is selected and logically or physically connected to the first relay means to thereby make the current working. An operation for relaying the second communication data on the second information communication network, an operation for monitoring the operation of the working second relay means, a failure of the working second relay means, The second relay means in the second information communication network by logically or physically connecting the second relay means to the first relay means.
An information relay device, comprising: redundancy control means for performing an operation of continuing the relay of communication data. 2. A first relay means for relaying first communication data in a first information communication network, and a first relay means connected to the first relay means and constructed on the first information communication network. A second relaying second communication data carried by the first communication data on a logical second information communication network;
An information relay device that relays, through the second relay unit, only the first communication data that satisfies a predetermined condition, without passing through the second relay unit. A plurality of second relay means connected to the first relay means; and a plurality of logical communication paths set on the first information communication network by the first relay means. Said second individual dynamically
An information relay device, comprising: a redundancy control unit that operates a plurality of the second relay units in parallel by assigning to the relay unit. 3. The information relay device according to claim 1, wherein the first information communication network is an ATM (Asynchronous Transfer Mode) communication network, the first relay means is an ATM switch,
The first communication data is an ATM cell, the second information communication network is an IP (Internet Protocol) network, and the second communication data is an IP datagram transferred on the IP network. The second relay means functions as an IP router for processing the IP datagram, and
A label switch routing controller that maps an address to header information of the ATM cell so that data of the ATM cell including an IP datagram of a specific IP address is cut-through transferred only by the ATM switch. Information relay device.