JP6317506B2 - Relay device, relay method, and relay program - Google Patents

Description

  The present invention relates to a relay device, a relay method, and a relay program.

  In a corporate LAN or the like, what performs a process of transmitting a request from a client that cannot be directly connected to the Internet to a target server on behalf of the client is generally called a proxy. This proxy is usually installed at the point of contact between the client and the Internet. On the other hand, a proxy is installed at the point of contact between the Internet and the target server, and when a request is made to the server from the outside, a response processing is performed instead of the request destination server, generally called a reverse proxy . This reverse proxy can reduce the burden on the target server by taking over the response that the target server should perform in response to requests from an unspecified number of clients. In addition, by restricting access to the target server with a reverse proxy, the security of the target server can be enhanced.

  As an invention related to a reverse proxy, for example, an apparatus having a function of rewriting a Set-Cookie header so that a cookie set by a server can be used effectively is known (Patent Document 1).

JP 2004-94905 A

  By the way, when a user accesses a Web page, a URI (Uniform Resource Identifier) of a target Web server (hereinafter referred to as “origin server”) is designated to a user agent such as a Web browser. Can be displayed. A method for describing the path from the current current position as a starting point to a target file or folder (hereinafter, this description method is simply referred to as “relative path”) for each content constituting the Web page held by the origin server. )) Is described, but the method of describing all the routes from the highest hierarchy in the device to the target file or folder (hereinafter, this description method is simply referred to as “absolute path”). In some cases, links are described. Here, if a reverse proxy is constructed on behalf of the origin server, and the link to the content in the target web page is described with an absolute path on the origin server, the reverse proxy is set to the origin server. In some cases, the target content cannot be requested. Hereinafter, this problem will be specifically described.

The origin server URI requested from the Web browser includes a character string that is interpreted by either the origin server or the reverse proxy in addition to the origin server that is the destination of the request. For example, assume that the URI of the following description (1) is specified in the Web browser.
http://www.dit.co.jp/origin/company/index.html (1)
To explain with the URI shown in the description (1), the path portion “/origin/company/index.html” corresponds to a character string interpreted by either the origin server or the reverse proxy. In addition, the origin server identifier “/ origin” is embedded in this path part, so the reverse proxy knows which origin server it is requesting and obtains the content requested from the Web browser. The previous origin server can be determined. That is, since “/ origin” in the URI shown in the description (1) is a part that is interpreted by the reverse proxy, the character string directly interpreted by the origin server is “/company/index.html”. Therefore, the reverse proxy that has received the request of (1) from the Web browser sends a request of the following description (2) in which “/ origin” is deleted to the origin server of the request destination, and “index” is sent from the origin server. .html ”is acquired and a response is returned to the Web browser. As a result, “index.html” is displayed on the Web browser.
/company/index.html (2)

Next, it is assumed that the link to the image content described in “index.html” acquired by the request of the description (1) by the Web browser is described with a relative path. For example, it is assumed that there is a link of the following description (3) in “index.html”.
<A HREF=“baz.jpg”> Photo </a> (3)
Assume that the link shown in the description (3) is clicked on the Web page displayed on the Web browser. At this time, since the web browser recognizes “baz.jpg” under “http://www.dit.co.jp/origin/company/” for the image content specified by the above link, The URI request to the reverse proxy has the following description (4).
http://www.dit.co.jp/origin/company/baz.jpg (4)

The reverse proxy that has received the request of the above description (4) from the Web browser sends the request of the following description (5) from which “/ origin” has been deleted to the origin server that is the request destination of the description (4). Then, “baz.jpg” is acquired from the origin server and a response is returned to the Web browser. As a result, the image content “baz.jpg” is displayed on the Web browser.
/company/baz.jpg (5)

On the other hand, it is assumed that the link to the image content described in “index.html” acquired by the Web browser by the URI request shown in the description (1) is described by an absolute path in the origin server. For example, it is assumed that there is a link of the following description (6) in “index.html”.
<A HREF=“/img/baz.jpg”> Photo </a> (6)
Then, it is assumed that the link of the description (6) is clicked by the user on the Web page displayed by the Web browser.

Since the Web browser recognizes that the content exists in “http://www.dit.co.jp/” when making the URI request of the above description (5), the URI request to the reverse proxy is It becomes the following description (7).
http://www.dit.co.jp/img/baz.jpg (7)
The reverse proxy that has received the request of this description (7) does not include the path part for grasping “which origin server it is requested” in, so the request “baz. “jpg” cannot be obtained.

  In other words, if the link of the content held on the origin server side is described with an absolute path, it is impossible to determine which origin server the reverse proxy side is requesting, and the request from the web browser is forwarded correctly. Can not be made.

  The present invention has been made in view of the above-described circumstances, and is a relay device that can be transferred to an origin server even when a content link held on the origin server side is described with an absolute path It is an object to provide a relay method and a relay program.

  In order to solve the above-described problem, one aspect of the present invention relates to a relay device. That is, a relay device that accepts a request from a user agent belonging to an external network and relays the request to one of a plurality of origin servers belonging to the internal network according to the request destination of the user agent. Includes a header describing information given in advance by the relay device to the user agent, or a header describing information indicating the link source from which the user agent made a request to the origin server. A determination unit that determines whether or not to rewrite the URI included in the request from the user agent by comparing the URI included in the request from the agent with the URI included in the header, and according to the determination result of the determination unit And a processing unit that rewrites the URI included in the request from the user agent. The features.

  Further, in the relay device described above, the determination unit, according to the type of request transmitted from the user agent, a header in which information that the relay device has previously given to the user agent before the request is described, and It is possible to determine whether or not to rewrite the URI included in the request from the user agent by referring to any of the headers describing information indicating the link source that the user agent has requested immediately before the request. .

  In addition, in any of the relay devices described above, when there is no header in which the information given in advance by the relay device is included in the request from the user agent, when the response of the request is transmitted to the user agent, A header adding unit that assigns a Set-Cookie header is provided, and the processing unit receives a request from a user agent that includes a Cookie header that includes the Cookie name specified by the Set-Cookie header assigned by the header adding unit. If the URI description included in the request does not start with the description of the cookie value corresponding to the cookie name included in the cookie header of the request, the URI included in the request from the user agent is not included. Rewriting can be performed.

  Also, in any of the above-described relay apparatuses, the determination unit refers to a Referer header that describes information indicating a link source from which the user agent has made a request to the origin server immediately before the request from the user agent. Then, the URI included in the request from the user agent can be rewritten by comparing the URI included in the request from the user agent with a part of the URI included in the Referer header and if the two do not match. .

  One aspect of the present invention relates to a relay method. In other words, the relay method of the present invention receives a request from a user agent belonging to an external network, and relays the request to one of a plurality of origin servers belonging to the internal network according to the request destination of the user agent. In the relay method, the request includes a header in which information previously given by the relay device to the user agent is described, or information indicating a link source from which the user agent makes a request to the origin server. A determination step of determining whether to rewrite the URI included in the request from the user agent by comparing the URI included in the request from the user agent with the URI included in the header; The UR included in the request from the user agent according to the determination result of the determination step Those having a processing step of rewriting the.

  One aspect of the present invention relates to a relay program. In other words, the relay program of the present invention receives a request from a user agent belonging to an external network, and requests a computer to one of a plurality of origin servers belonging to the internal network according to the request destination of the user agent. Is a relay program for functioning as a relay device that relays the request, and the request includes a header in which information previously given by the relay device to the user agent is described, or the user agent makes a request to the origin server Included in the request from the user agent by comparing the URI contained in the request from the user agent with the URI contained in the header. Determining means for determining whether or not to rewrite the URI, It is intended to function as processing means for rewriting the URI included in the request from the user agent in response to the means of determination results.

  According to the present invention, a relay device, a relay method, and a relay method capable of causing a URI request from a user agent to be transferred to an origin server even when a link of content held on the origin server side is described with an absolute path. A relay program can be provided.

It is a figure showing the whole personal platform system composition containing reverse proxy server 10 concerning a 1st embodiment of the present invention. It is a block diagram which shows the functional structure of the reverse proxy server 10 shown in FIG. It is a flowchart which shows an example of the relay process which the reverse proxy server 10 shown in FIG. 1 performs. It is a flowchart which shows an example of the relay process which the reverse proxy server 10 shown in FIG. 1 performs. It is a block diagram which shows the functional structure of 10 A of reverse proxy servers which concern on 2nd Embodiment of this invention. It is a flowchart which shows an example of the relay process which 10 A of reverse proxy servers shown in FIG. 5 perform. It is a flowchart which shows an example of the relay process which 10 A of reverse proxy servers shown in FIG. 5 perform. It is a figure which shows an example of the hardware constitutions of the reverse proxy servers 10 and 10A demonstrated in FIG. 1 and FIG.

  Hereinafter, reverse proxy servers 10 and 10A, which are an embodiment of a relay apparatus of the present invention, will be described with reference to the drawings. Note that the relay method and relay program of the present invention will be described by taking the operation of the reverse proxy servers 10 and 10A and the relay program installed in the reverse proxy servers 10 and 10A as examples. However, the embodiment described below is an example embodying the present invention, and does not limit the present invention.

[First embodiment]
FIG. 1 is a diagram showing an overall configuration of a personal platform system including a reverse proxy server 10 according to the first embodiment of the present invention. In this personal platform system, devices 200A, 200B, and 200C connected to a LAN (Local Area Network) 120 in a user's home from the Web browser 100 via the Internet 110 and the reverse proxy server 10 (hereinafter, individually distinguished). When there is no need, it is simply a system that can safely access the “device 200” from the outside. Although illustration is omitted, access to the reverse proxy server 10 from the Internet 110 side is communication to a specific port of the reverse proxy server 10 when the specific IP address is the source. It is assumed that a firewall that is set to permit only communication is built in advance.

  The Web browser 100 is an example of the user agent described in the claims, and is installed in an information processing apparatus (none of which is not shown) such as a notebook PC or a smartphone owned by the user of the personal platform system shown in FIG. Yes. The user activates the web browser 100 when away from home, connects to the device 200 from the outside via the reverse proxy server 10, and contents (for example, image data, video data, Text data, etc.).

  The reverse proxy server 10 is an example of a relay device described in the claims, and is a server having a so-called reverse proxy function. The reverse proxy server 10 is connected to the LAN 120 together with a device 200 to be described later in the user's home, responds to the URI request from the Web browser 100 activated by the user on behalf of the device 200, and responds to the requested device 200. The device transmits a URI request and returns a response from the device 200 to the Web browser 100. In the reverse proxy server 10, a path name for the device 200 to publish content to the outside (hereinafter, this path name is referred to as “public path name”) and a virtual indicated on the directory of the reverse proxy server 10. The relationship information for rewriting with a specific path name (hereinafter, this path name is referred to as “virtual path name”) is set in advance, and can be interpreted by the device 200 in response to a request from the Web browser 100. It is configured so as to be appropriately converted and transferred so as to be in a format.

  The device 200 is, for example, an electronic device such as a network camera, a video recorder, or a NAS (Network Attached Storage), and is connected to a LAN 120 that is built in a user's home using the personal platform system shown in FIG. Each device 200 has its own directory structure and content, and the content is stored under the public path associated with the virtual path described above so that it can be connected from the outside via the reverse proxy server 10. It is configured. In the personal platform system shown in FIG. 1, it is assumed that the user previously executes registration processing for the device 200 with respect to the reverse proxy server 10 from a PC (not shown) connected to the LAN 120.

  FIG. 2 is a block diagram showing a functional configuration of the reverse proxy server 10 shown in FIG. The reverse proxy server 10 includes a filter unit 11 and a reverse proxy unit 12.

  The filter unit 11 receives a request from the Web browser 100 and determines whether to rewrite the URI included in the received request. Further, when the filter unit 11 determines that the URI included in the received request needs to be rewritten, the filter unit 11 rewrites the URI so as to be an appropriate URI and transfers it to the device 200 via the reverse proxy unit 12 described later. To do. In addition, the filter unit 11 transfers the response received from the device 200 via the reverse proxy unit 12 to the Web browser 100. In order to realize these functions, the filter unit 11 includes a determination unit 21, a cookie granting unit 22, a processing unit 23, a request issuing unit 24, and a response issuing unit 25 as shown in FIG. doing.

  The determination unit 21 determines whether the request from the Web browser 100 includes a cookie name assigned by a cookie assignment unit 22 described later. If the determination unit 21 determines that there is a cookie name, the request from the web browser 100 is received at the head of the virtual path name based on the cookie value included in the cookie header in which the cookie name is described. It is further determined whether or not it has started, and in accordance with this determination result, it is determined whether or not to rewrite the URI included in the accepted request.

  For example, the determination unit 21 includes the cookie name given by the cookie granting unit 22 described later in the request from the Web browser 100, and the virtual path name from which the URI included in the request is derived from the cookie value If the URI included in this request begins, it is determined that there is no problem even if the URI included in this request is interpreted as it is in the reverse proxy unit 12 which will be described later, and the request issuing unit 24 which will be described later is transferred. Hand it over as it is.

  On the other hand, the determination unit 21 has a cookie name given by a cookie granting unit 22 to be described later in the URI included in the request from the Web browser 100. The virtual path from which the URI included in this request is derived from the cookie value If the name does not begin with the head part of the name, the reverse proxy unit 12 described later interprets the URI included in this request as it is, so it cannot be determined to which device 200 the URI should be transferred. It is determined that rewriting is necessary, and the processing unit 23 described later is instructed to execute URI rewriting processing.

  In addition, when the cookie header is not included in the request from the Web browser 100, the determination unit 21 determines that it is the first connection to the device 200 or that the cookie information has disappeared for some reason. Then, it instructs the Cookie granting unit 22 to be described later to add a Set-Cookie header to the header of the response to this request. As a result, a Cookie name and a Cookie value specified by the Cookie giving unit 22 described later in the Set-Cookie header are given to a request from the Web browser 100 generated after this request.

The cookie adding unit 22 adds a Set-Cookie header to the header of a response to the URI request from the Web browser 100 according to an instruction from the determination unit 21. The Set-Cookie header describes the cookie name and the cookie value given by the reverse proxy server 10. The cookie name and the cookie value are described, for example, as shown in the following description (8). It is preferable that an appropriate character string that does not collide with the cookie name assigned by another system (the possibility of collision is sufficiently low) is set for the cookie name. In the Cookie value, a character string (for example, “/ orijin” or the like) from which the head portion of the virtual path name of the device 200 is obtained is described.
Set-Cookie: Cookie name = Cookie value (8)

  The processing unit 23 executes a process of rewriting the URI request according to the instruction from the determination unit 21 described above. Note that the processing for rewriting the URI included in the request from the Web browser 100 will be described with reference to a flowchart (FIG. 4) described later.

  The request issuing unit 24 transfers the URI request after being rewritten by the processing unit 23 described above or the URI request transferred as it is from the determination unit 21 to the reverse proxy unit 12 described later.

  The response issuing unit 25 transfers the response (corresponding to the response of the URI request from the Web browser 100) from the device 200 delivered from the reverse proxy unit 12 described later to the Web browser 100.

  The reverse proxy unit 12 converts the URI request received from the filter unit 11 into a format suitable for the directory structure on the device 200 side, transfers the converted request to the device 200, and sends a response from the device 200 to the transferred URI request. It has a function of receiving and transferring it to the filter unit 11. In order to realize these functions, the reverse proxy unit 12 includes a request issue transfer unit 31, a response issue transfer unit 32, and a device information registration table 33.

  The request issuance transfer unit 31 transfers the URI request transferred from the filter unit 11 described above into a format suitable for the directory structure on the device 200 side, and transfers the request to the device 200. Specifically, when the URI request delivered from the request issuing unit 24 is a virtual path name, the request issue transfer unit 31 is included in a request from the device information registration table 33 or the Web browser 100 described later. With reference to the Cookie header, the character string corresponding to the head part of the virtual path name of the transfer destination device 200 is removed to convert it to a public path name and transfer it to the target device 200.

  When the response issuance / transfer unit 32 receives a response from the device 200, the response issuance / transfer unit 32 transfers the response to the filter unit 11 described above.

  The device information registration table 33 is a table that stores relationship information for appropriately rewriting the public path name of the device 200 and the virtual path name on the reverse proxy server 10. In the reverse proxy server 10, the device information registration table 33 or the Cookie header included in the request from the Web browser 100 is referred to ascertain which device 200 is the URI request, and the correct transfer destination Can be determined.

(Communication processing)
Next, relay processing executed by the reverse proxy server 10 will be described. In the reverse proxy server 10, when a URI request is transmitted from the Web browser 100, the subsequent processing differs depending on the presence or absence of the cookie name assigned to the URI request by the cookie granting unit 22. A flowchart will be shown and described. 3 and 4 are flowcharts illustrating an example of relay processing executed by the reverse proxy server 10 illustrated in FIG.

  First, with reference to the flowchart shown in FIG. 3, processing when the cookie name assigned in advance by the cookie adding unit 22 is not included in the URI included in the request from the Web browser 100 will be described. The relay processing shown in FIG. 3 is a processing procedure executed when the device 200 is accessed for the first time after the Web browser 100 is activated, for example.

In step S <b> 1 of FIG. 3, the Web browser 100 issues a URI request (9) below, and transmits the URI request to the reverse proxy server 10. In this URI request, “proxy” is set as host information indicating the reverse proxy server 10.
GET /origin/index.html HTTP / 1.1
Host: proxy (9)

  In step S2, when the filter unit 11 of the reverse proxy server 10 receives the URI request (9) described above, the cookie name (FIG. 3), which is information specified by the Cookie granting unit 22 in the URI request by the Set-Cookie header. Then, the cookie name is P). In this example, in step S2, the filter unit 11 determines that there is no cookie name assigned in advance to the URI request transmitted in step S1, and the process proceeds to step S3.

  In step S <b> 3, the filter unit 11 passes the URI request transmitted in step S <b> 1 to the reverse proxy unit 12 as it is without rewriting.

In step S4, the reverse proxy unit 12 converts the URI request shown in the following description (10) obtained by converting the URI request passed from the filter unit 11 in step S3 into a format suitable for the directory structure of the device 200. Issued against. Specifically, a URI request from which “/ origin”, which is a part for determining which device 200 is the reverse proxy server 10 side, is issued to the device 200. Hereinafter, the part that determines which device 200 is the reverse proxy server 10 side will be referred to as a “transfer start point”.
GET /index.html HTTP / 1.1
Host: orijin (10)

  In step S5, the device 200 issues a response to the URI request in step S4. Specifically, data constituting “index.html” is transmitted to the reverse proxy server 10.

  In step S6, when receiving the response from the device 200 transmitted in step S5, the reverse proxy unit 12 passes the response to the filter unit 11.

In step S7, since the response from the device 200 delivered in step S6 is a response to the URI request transmitted by the web browser 100 in step S1, the filter unit 11 sets Set in the response header transmitted to the web browser 100. -Return the one with the Cookie header to the Web browser 100. An example is shown in the following description (11).
HTTP / 1.1 200 OK
Content-Length: xxx
Connection: close
Set-Cookie: p = / origin (11)

  When the Web browser 100 that has received the response shown in the above description (11) transmits a URI request to the target device 200 thereafter, the information specified in the Set-Cookie header is added to the Cookie header. And send it.

  Next, with reference to a flowchart shown in FIG. 4, a relay process when the Cookie name that is information instructed by the cookie giving unit 22 is included in the URI included in the request from the Web browser 100 will be described. In the relay process shown in FIG. 4, after the above-described relay process of FIG. 3 is executed once, the link “index1.html” in the displayed “index.html” Web page is clicked by the user. It will be explained as a thing.

In step S <b> 10, the Web browser 100 transmits a URI request with the following description (12) to the reverse proxy server 10. The URI request shown in step S10 is described with an absolute path. In addition, the URI request shown in step S10 includes a cookie header in which a cookie name (“P”) given in advance by the reverse proxy server 10 and a referer header in which URI information of the link source is described. include.
GET /index1.html HTTP / 1.1
Host: proxy
Cookie: p = / origin
Referer: http: //proxy/origin/index.html (12)

  In step S <b> 11, when the filter unit 11 of the reverse proxy server 10 receives the request transmitted in step S <b> 10, the filter unit 11 determines whether or not there is a cookie name assigned in advance by the cookie granting unit 22 in this request. As described in step S10, since the URI request transmitted in step S10 has “P” as the cookie name assigned in advance, the filter unit 11 further proceeds to the determination in step S12.

In step S12, the filter unit 11 determines whether or not the URI request from the Web browser 100 (that is, the part starting with the GET method) starts at the head part of the virtual path name derived from the Cookie value described in the Cookie header. Determine. In this example, the URI request from the Web browser 100 starts with “/index1.html”, whereas the cookie value corresponding to the cookie name “P” described in the cookie header is “/ origin”. The filter unit 11 determines that the URI included in the request from the Web browser 100 does not start with the head part of the virtual path name derived from the Cookie value described in the Cookie header. Then, “/ origin”, which is the head part of the virtual path name derived from the Cookie value described in the Cookie header, is inserted at the head of the above-described URI request. As a result, the filter unit 11 passes the URI request shown in the following description (13) to the reverse proxy unit 12. As a result, the reverse proxy unit 12 can know which device 200 the URI is requested to even if the URI included in the request from the Web browser 100 is described with an absolute path.
GET /origin/index1.html HTTP / 1.1
Host: proxy
Cookie: p = / origin
Referer: http: //proxy/origin/index.html (13)

  The processing of the reverse proxy unit 12 from step S13 to step S16 is the same as the processing of the reverse proxy unit 12 from step S4 to step S7 described in FIG.

(Regarding the effect of the first embodiment)
Since the content that the device 200 has is normally designed on the assumption that it is viewed in the user's home, it is not always designed in consideration of being referred to from the outside. That is, the content storage location may be described with an absolute path based on the directory structure unique to the device 200. For this reason, when browsing the web browser 100 from an outside location, the URI request from the web browser 100 and the response from the device 200 need to be appropriately converted by the reverse proxy server 10 and transferred.

  The reverse proxy server 10 according to the first embodiment accepts a request from a Web browser 100 (user agent) belonging to the Internet 110 (external network), and connects to the LAN 120 (internal network) according to the request destination of the Web browser 100. The request is relayed to any of a plurality of devices 200 (a plurality of origin servers) to which the request belongs, and the cookie granting unit 22 of the reverse proxy server 10 is sent to the Web browser 100 before the request. Includes a cookie name and a cookie value, which are information indicated by the Set-Cookie header, and a virtual derived from the URI included in the request from the web browser 100 and the cookie value included in the cookie header of the request. Path name ahead By comparing the portion, so that determines whether rewrite the URI included in the request from the Web browser 100, rewrites the URI included in the request from the Web browser 100 in accordance with the determination result. Thereby, in this reverse proxy server 10, even if a link described with an absolute path exists in the content of the device 200, the cookie name designated in advance by the Cookie giving unit 22 in the Set-Cookie header is the Web If it is described in the cookie header included in the request from the browser 100, the reverse proxy is inserted after inserting the head portion of the virtual path name derived from the cookie value included in the cookie header into the head portion of the URI request. It is possible to transfer the data to the unit 12 and determine the device 200 to which the reverse proxy unit 12 determines the request. That is, the reverse proxy server 10 can transfer the URI request from the Web browser 100 to the target device 200 even when the link of the content held by the device 200 is described with an absolute path. it can.

  Unlike the present embodiment, when a hyperlink included in a Web page starts with “/”, it is determined that replacement is necessary, and characters to be complemented from the HTTP request that triggered the response. A method of searching for a column (in this embodiment, a character string that identifies which device 200 corresponds to “/ orijin”) and complementing it at the head of the hyperlink character string of the URI request is also possible. Though conceivable, in such a method, it is necessary to rewrite the hyperlink itself included in the Web page. For example, when hyperlinks are dynamically generated using JavaScript, finer control is required. Therefore, when a hyperlink is dynamically generated, it is very difficult to determine whether to replace an absolute path. In this regard, in this embodiment, the operation is performed with reference to the information included in the request to the link destination, rather than referring to the information included in the link source Web page that triggered the response. Therefore, it is not necessary to rewrite the hyperlink itself included in the link source Web page. That is, even if a hyperlink is dynamically generated by the above-described Javascript or the like on the link source Web page, the absolute path replacement determination can be performed without interpreting this.

[Second Embodiment]
FIG. 5 is a block diagram showing a functional configuration of the reverse proxy server 10A according to the second exemplary embodiment of the present invention. Unlike the reverse proxy server 10 according to the first embodiment of the present invention, the reverse proxy server 10A shown in FIG. 5 does not use the Cookie header, but uses a Referer header to request a URI request from the Web browser 100. The data is transferred to the device 200. Note that the Referer header is one of HTTP headers, and describes the address of a web page or resource linked to one web page or one resource on the Internet. In addition, about the function similar to 1st Embodiment, the same symbol is attached | subjected and the one part or all part of the description is abbreviate | omitted.

  As illustrated in FIG. 5, the filter unit 11 </ b> A includes a determination unit 21 </ b> A, a processing unit 23, a request issuing unit 24, a response issuing unit 25, and a device information registration table 26.

  The determination unit 21A determines whether or not the Refer header is included in the URI request from the Web browser 100, and if it is determined that the Refer header is included, the URI request is described in the Refer header. It is further determined whether or not it starts with a path name included in the URI that has been set, and it is determined whether or not to rewrite the accepted URI request according to the determination result.

  For example, when the URI request from the Web browser 100 includes a Referer header and the URI request matches the path name included in the URI described in the Referer header, the determination unit 21A It is determined that the URI request does not need to be rewritten, and is passed to the request issuing unit 24 described later.

  In addition, when the URI request from the Web browser 100 includes a Referer header and the URI request does not match the path name included in the URI described in the Referer header, the determination unit 21A It is determined that the URI request needs to be rewritten, and the processing unit 23 described later is instructed to execute a URI request rewriting process.

  When the Refer header is not included in the URI request from the Web browser 100, the determination unit 21A is the first connection after starting the Web browser, or the user explicitly specifies the URI. Therefore, the URI request is transferred to the reverse proxy unit 12 as it is.

  In the device information registration table 26, relationship information for appropriately rewriting the public path name of the device 200 and the virtual path name on the reverse proxy server 10 (for example, information indicating the correspondence relationship between the transfer origin itself and the device 200 described above). Etc.) are stored in the table. In the reverse proxy server 10A, by referring to the relationship information stored in the device information registration table 26, it is determined whether or not the transfer start point is included in the URI and to which device 200 the URI request is made. It is possible to grasp and determine the correct transfer destination. The device information registration table 26 included in the filter unit 11A and the device information registration table 33 included in the reverse proxy unit 12 hold the same information regarding the device 200, and thus the filter unit 11A and the reverse proxy unit 12 May be constructed independently of each other, or may be constructed in one of them, and may be configured so as to be referred to from the person who does not hold the table to the person who holds the table.

  Note that the processing unit 23, the request issuing unit 24, the response issuing unit 25, and the reverse proxy unit 12 have the same functions as those of the above-described first embodiment, and thus description thereof is omitted.

(Communication processing)
Subsequently, a relay process executed by the reverse proxy server 10A will be described. In the reverse proxy server 10A, when a URI request is transmitted from the Web browser 100, the subsequent processing differs depending on the presence or absence of the Referer header. 6 and 7 are flowcharts illustrating an example of relay processing executed by the reverse proxy server 10A illustrated in FIG.

  First, processing when the Refer header is not included in the URI request from the Web browser 100 will be described with reference to the flowchart shown in FIG. The relay processing shown in FIG. 6 is a processing procedure that is executed, for example, when the Web browser 100 is accessed for the first time after being activated.

In step S21 of FIG. 6, the Web browser 100 transmits a URI request having the following description (14) to the reverse proxy server 10A. In this URI request, “proxy” is set as host information indicating the reverse proxy server 10A.
GET /origin/index.html HTTP / 1.1
Host: proxy (14)

  In step S <b> 22, when the filter unit 11 </ b> A of the reverse proxy server 10 </ b> A receives the URI request (14), the filter unit 11 </ b> A determines whether there is a Referer header in the URI request. In step S22, the filter unit 11A determines that there is no Referer header in the URI request transmitted in step S21, and proceeds to step S23.

  In step S23, the filter unit 11A passes the URI request transmitted in step S21 to the reverse proxy unit 12 as it is without rewriting.

In step S24, the reverse proxy unit 12 converts the URI request shown in the following description (15) obtained by converting the URI request passed from the filter unit 11A in step S23 into a format suitable for the directory structure of the device 200. Issued against. Specifically, a URI request from which / origin, which is a part for determining which device 200 is the reverse proxy server 10A side (that is, the transfer start point described in the first embodiment described above) is removed, is sent to the device 200. Issue.
GET /index.html HTTP / 1.1
Host: orijin (15)

  In step S25, the device 200 issues a response to the URI request in step S24. Specifically, data constituting “index.html” is transmitted to the reverse proxy server 10A.

  In step S26, when receiving the response from the device 200 transmitted in step S25, the reverse proxy unit 12 passes the response to the filter unit 11A.

In step S27, the filter unit 11A returns the response from the device 200 delivered in step S26 to the web browser 100. An example is shown in the following description (16).
HTTP / 1.1 200 OK
Content-Length: xxx
Connection: close (16)

  Next, a relay process when the Refer header is included in the URI request from the Web browser 100 will be described with reference to the flowchart shown in FIG. The relay process shown in FIG. 7 will be described on the assumption that the user clicked on the link “index1.html” in the displayed webpage “index.html”.

In step S30, the Web browser 100 transmits a URI request having the following description (17) to the reverse proxy server 10A. The URI request shown in step S30 is described with an absolute path. Further, the URI request shown in step S30 includes a Referer header in which URI information of the link source is described.
GET /index1.html HTTP / 1.1
Host: proxy
Referer: http: //proxy/origin/index.html (17)

  In step S31, when the filter unit 11A of the reverse proxy server 10A receives the URI request transmitted in step S30, the determination unit 22A determines whether or not there is a Referer header in the URI request. As described in step S30, since the URI request transmitted in step S30 includes a Referer header, the filter unit 11A further proceeds to the determination in step S32.

In step S32, the filter unit 11A determines whether the URI request from the Web browser 100 (that is, the part starting with the GET method) matches the absolute path part of the URI described in the Referer header. . In this example, the URI request from the Web browser 100 is “/index1.html”, whereas the absolute path portion of the URI described in the Referer header is “/ origin”, so the filter unit 11A Determines that the URI included in the request from the Web browser 100 and the absolute path portion of the URI described in the Referer header do not match (that is, the transfer start point of both does not match). In step S33, the filter unit 11A sets the absolute path portion (that is, “/ origin”) of the URI described in the Referer header at the head of the URI included in the request from the Web browser 100 described above. Insert and issue request. As a result, the filter unit 11A passes the URI request of the following description (18) to the reverse proxy unit 12. As a result, the reverse proxy unit 12 can know which device 200 the URI is requested to even if the URI included in the request from the Web browser 100 is described with an absolute path.
GET /origin/index1.html HTTP / 1.1
Host: proxy
Referer: http: //proxy/origin/index.html (18)

  Note that the processing of the reverse proxy unit 12 from step S34 to step S37 is the same as the processing of the reverse proxy unit 12 from step S14 to step S17 described in FIG.

(About effects of the second embodiment)
In the reverse proxy server 10A according to the second embodiment, a request from a web browser 100 (user agent) belonging to the Internet 110 (external network) is accepted, and the request is made to the LAN 120 (internal network) according to the request destination of the web browser 100. The request is relayed to any of a plurality of devices 200 (a plurality of origin servers) to which the web browser 100 belongs, and in the above request, the web browser 100 makes a request to the device 200 immediately before the request from the web browser 100. A Referer header in which information indicating the link source is described is included. The URI included in the request from the Web browser 100 is compared with the URI included in the Referer header. Whether to rewrite the included URI Determined, so that rewrite the URI included in the request from the Web browser 100 in accordance with the determination result. Thereby, the reverse proxy server 10A refers to the URI immediately before the request of the Web browser 100 by referring to the Referer header even if the link described in the absolute path exists in the content of the device 200. The path portion for interpretation by the reverse proxy server 10A can be inserted at the head portion of the URI described by the absolute path. As a result, the reverse proxy unit 12 can determine which device 200 the request is for and transfer the request. That is, even when the content link held by the device 200 is described with an absolute path, the URI request from the Web browser 100 can be transferred to the target device 200.

(Other embodiments)
Each embodiment described above can be variously modified without departing from the gist thereof. For example, it has both the functions of the reverse proxy server 10 in the first embodiment and the reverse proxy server 10A in the second embodiment, and which function is used for determination according to the type of request from the Web browser 100. You may comprise so that it may determine. For example, when a request for the GET method is transmitted from the Web browser 100, the function of the reverse proxy server 10A in the second embodiment is configured to make a URI rewrite determination according to the presence or absence of the Referer header. When a POST method request is transmitted from the Web browser 100, the function of the reverse proxy server 10 in the first embodiment determines whether or not the Cookie names included in the Cookie header match. It may be. With such a configuration, for example, when determination based on the presence or absence of a Referer header is adopted (that is, determination based on the function of the reverse proxy server 10A is adopted), a large amount of data must be uploaded twice. The problem of disappearing can be solved. This is a problem that occurs when it is desired to deal with the response of the reverse proxy server 10A by using a response such as “301 Moved Permanently” because the Referer cannot refer to the grandchild. When the reverse proxy server 10A returns a 301 response to the first HTTP request, the Web browser 100 issues a second HTTP request to the destination specified in the Location line included in the 301 response. Become. Then, a 200 (OK) response is returned as the response from the destination of the second HTTP request, and the request of the Web browser 100 is completed. This operation itself is the same for both the POST method and the GET method. However, the GET method transmits several tens to at most several hundred characters in one request, whereas the POST method transmits more characters. Therefore, it may occur that the web browser 100 sends millions of characters twice, for example, and the amount of transmission data becomes very large. Therefore, in the case of a request for the POST method, determination based on the Cookie header is adopted (that is, determination based on the function of the reverse proxy server 10) so that the response can be made without using the 301 response. In this case, the determination based on the presence or absence of the Referer header can solve the above-described problem that the amount of transmission data becomes very large.

  In the first embodiment described above, since the reverse proxy server 10 assigns a cookie name to the response header to the Web browser 100, in the case where the device 200 side (origin server side) also assigns a cookie. Therefore, it is necessary to select an appropriate character string so that name collision does not occur between these cookies. If there is such a possibility, for example, a rule may be provided so that the cookie name assigned on the reverse proxy server 10 side or the cookie name assigned on the device 200 side can be uniquely identified.

  In addition, in step S2 of the flowchart shown in FIG. 3 described above, the request URI that does not have a cookie name is normally considered to be a situation in which the reverse proxy server 10 is not accessed before this request. In addition, there is a possibility that the user clicked a link displayed in the web page after the cookie was discarded due to expiration of the cookie. If the cookie is discarded in the Web browser 100, the reverse proxy server 10 tries to refer to its own folder specified by the absolute path indicated in the request URI, but typically 404 that the resource does not exist. An error (Not Found) occurs. When access authentication is performed for each top level of the virtual path set in the reverse proxy server 10, a 401 error (Unauthorized) occurs. In such a case, by making another URI request to the link source on the Web browser 100 side, the reverse proxy server 10 determines that the request has a cookie name, so that it is possible to access the correct link destination thereafter. Become.

Further, in step S3 of the flowchart shown in FIG. 3 described above, if the virtual path name on the reverse proxy server 10 side and the public path name on the device 200 side are the same, the determination may not be made correctly. For example, it is assumed that the device 200 has a folder structure and file “/foo/img/baz.jpg”, and an absolute path in the device 200 is specified there. Here, in the reverse proxy server 10
http: //proxy/foo/bar.html
When there is a request to
http: //origin/bar.html
To “bar.html”
<A HREF=/foo/img/baz.jpg>
When there is a hyperlink (a),
<A HREF=./img/baz.jpg>
(B), a request such as the following description (19) is issued from the Web browser 100.
GET /foo/img/baz.jpg HTTP / 1.1
HOST: proxy
Cookie: p = / foo (19)
On the reverse proxy server 10 side, it cannot be distinguished whether the request is (a) or (b) described above, but (a) should cause a replacement, and (b) caused a replacement. Should not be allowed to In order to avoid such an event, the virtual path name may have a collision such as a GUID so that the virtual path name in the reverse proxy server 10 and the public path name on the device 200 side do not collide. It is preferable to use a character string having a low. Similarly, for the public path name, it is preferable to use a character string having a low possibility of collision such as GUID.

Note that the request header line in the above-described example is in the absolute path (abs_path) format referred to in RFC 2616 and RFC 2396, but there is a possibility of an absolute URI (absolute URI) in the standard. That is,
GET http: //proxy/foo/img/baz.jpg HTTP1.1 (20)
Or GET http: // proxy: 80 / foo / img / baz.jpg HTTP1.1 (21)
In the case of the above description (20) and description (21), there is no difference in that (a) and (b) cannot be distinguished. In the determination of the reverse proxy server 10, normally (a) is determined, but when the virtual path is valid in the reverse proxy server 10, an optional function for determining (b) is provided. May be. By providing such an optional function, even if both (a) and (b) are valid resources, only one of them is referred to, but it is possible to select which one to refer to.

In addition, there may be a case where the file name specified in the URI request in step S12 in the flowchart shown in FIG. 4 is omitted.
For example, as shown in the description (22) below, GET /dir/HTTP1.1
HOST: proxy
Cookie: p = / foo (22)
In the case of the request, it is preferable to rewrite the filter unit 11 of the reverse proxy server 10 so as to refer to “/ foo / dir”.

Also, as shown in the description (23) below, GET / HTTP1.1
HOST: proxy
Cookie: p = / foo (23)
In the case of the request, it is preferable to rewrite so that “/ foo /” is referred to by the filter unit 11 of the reverse proxy server 10.

Further, when the URI request is an absolute URI (absolute URI), for example, as shown in the following description (24),
GET http: //proxy/img/baz.jpg HTTP1.1
HOST: proxy
Cookie: p = / foo (24)
In the case of the request, it is preferable to rewrite so that “http: //proxy/foo/img/baz.jpg” is referred to by the filter unit 11 of the reverse proxy server 10.

If the absolute path (abs_path) portion is empty, for example, as shown in the following description (25),
GET http: // proxy HTTP1.1
HOST: proxy
Cookie: p = / foo (25)
In the case of the request, it is preferable to rewrite so that “http: // proxy / foo” is referred to by the filter unit 11 of the reverse proxy server 10.

Further, the reverse proxy server 10A in the second embodiment described above obtains a public path to the device 200 by analyzing the Referer header. However, in addition to the absolute URI (abusolute URI) referred to in RFC 2616 / RFC 2396 being specified as the Referer header, there is a possibility that it becomes a relative URI (relative URI) in terms of specifications. For example, the following descriptions (26), (27), and (28) are included.
Referer: //proxy/foo/bar.html (26)
Referer: /foo/bar.html (27)
Referer: bar.html (28)
In the case of the above descriptions (26), (27), and (28), it is preferable that the reverse proxy server 10 </ b> A makes a determination by supplementing the lacking part from other information included in the request URI from the Web browser 100.

In addition, the reverse proxy server 10A in the second embodiment described above may not be able to make a determination correctly. For example, consider a case where “/bar.html” of the device 200 has a hyperlink to “/baz.html” and “/baz.html” has a hyperlink to “/img/qux.jpg”. When the Web browser 100 referring to “/bar.html” requests “/baz.html”, the URI request to the reverse proxy server 10A has the following description (29).
GET /baz.html HTTP1.1
HOST: proxy
Referer: http: //proxy/foo/bar.html (29)
When the reverse proxy server 10A rewrites the request URI from the Web browser 100, the following description (30) is obtained.
GET /foo/baz.html HTTP1.1
HOST: orijin
Referer: http: //proxy/foo/bar.html (30)
The reverse proxy server 10A transmits the request of the above description (30) to the device 200 and acquires “/baz.html”. Then, when the web browser 100 refers to “qux.jpg” by following the hyperlink described by the absolute path “/img/qux.jpg” included in “/baz.html”, the web browser From 100, a request of the following description (31) is generated.
GET /img/qux.jpg HTTP1.1
HOST: proxy
Referer: http: //proxy/baz.html (31)
As shown in the Referer header of the above description (31), since “/ foo” is not included, the determination described in step S32 of FIG. 7 described above does not correspond to the transfer origin of the request URI. Since the comparison requirement as a premise for executing this determination is not satisfied, rewriting is not performed. This is because, when referring to “/baz.html”, it is determined that the Web browser 100 has traced from “http: //proxy/baz.html”. Therefore, in order for the reverse proxy server 10A to operate correctly, “/baz.html” specified in the request URI for the Web browser 100 is actually “http: // proxy / foo / baz”. You must be informed that it is in .html ”.

In HTTP, a response code (for example, 301 (Moved Permanently), 302 (Found), 305 (Use Proxy), 307 (Temporary Redirect) for notifying the Web browser 100 that the URI has changed is shown in HTTP. ) Etc.) are defined. Since there is a difference in which response code can be supported by the HTTP version and the Web browser 100, the response is returned by appropriately branching which response code is returned by looking at the HTTP version of the request source and the UserAgent line. Must be set. For any response code, the changed URI is written in the Location line of the response header. For example, a response header as shown in the following description (32) is generated.
HTTP / 1.1 307 Temporary Redirect
Date: Fri, 18 May 2012 01:05:07 GMT
Location: http: //proxy/foo/img/baz.jpg
Connection: Close (32)

  In addition, the reverse proxy servers 10 and 10A described in the above-described embodiment and all or a part of the processes executed in these modified examples are configured by a program, and the program is recorded in an executable state. It can also be installed directly on a computer or on a computer remotely over a network.

  FIG. 8 is a diagram illustrating an example of a hardware configuration of the reverse proxy servers 10 and 10A described in FIG. 1 and FIG. In the computer constituted by the above-described reverse proxy servers 10 and 10A and modifications thereof, the CPU 301 transfers, for example, the relay program stored in the storage unit 308 to the RAM 103 via the input / output interface 304 and the bus 305. By loading and executing, the above-described series of processing is performed.

  The relay program executed by the computer (CPU 301) is, for example, a magnetic disk (including a flexible disk), an optical disk (CD-ROM (Compact Disc-Read Only Memory), DVD (Digital Versatile Disc), etc.), a magneto-optical disk, or It is recorded on a removable medium 311 that is a package medium made of a semiconductor memory or the like, or provided via a wired or wireless transmission medium such as a local area network, the Internet, or digital satellite broadcasting.

  The relay program can be installed in the storage unit 308 via the input / output interface 304 by attaching the removable medium 311 to the drive 310. The relay program can be received by the communication unit 309 via a wired or wireless transmission medium and installed in the storage unit 308. In addition, the program can be installed in the ROM 302 or the storage unit 308 in advance.

  Note that the relay program executed by the computer may be a program that is processed in time series in the order described in this specification, or in parallel or at a necessary timing such as when a call is made. It may be a program in which processing is performed. Furthermore, the relay program may be transferred to a remote computer and executed. Further, as hardware for executing the relay program, in addition to the general-purpose computer, if the reverse proxy servers 10 and 10A described above and the functions of these modifications are provided, they are configured on other electronic devices. It may be a thing.

  DESCRIPTION OF SYMBOLS 10,10A ... Reverse proxy server (an example of a relay apparatus) 11, 11, A ... Filter part, 12 ... Reverse proxy part, 21, 21A ... Determination part, 22 ... Cookie grant part, 23 ... Processing part, 24 ... Request issue part 25 ... Response issuing unit, 26 ... Device information registration table, 31 ... Request issue transfer unit, 32 ... Response issue transfer unit, 33 ... Device information registration table, 100 ... Web browser (an example of user agent), 110 ... Internet, 200, 200A, 200B, 2000C ... equipment (an example of an origin server)

Claims (4)

In response to a request from a user agent belonging to an external network, a relay device that relays the request to one of a plurality of devices belonging to the internal network and forwards a response from the relayed device to the user agent. There,
The relay device is
A determination unit for determining whether to rewrite the request;
A processing unit that rewrites the request when it is determined that the determination unit needs to be rewritten;
A device information registration table in which relation information between a public path name for publishing contents possessed by the plurality of devices and a virtual path name indicated on the directory of the relay device is stored;
Referring to the virtual path name in the device information registration table, the request destination device and the public path name of the device are specified, and the request is converted into a format suitable for the directory structure on the device side. A request issuance transfer unit for transferring to the device;
A response issuance transfer unit that, when receiving a response from the requested device, transfers the response to the filter unit;
A relay apparatus comprising:
The relay device according to claim 1,
The request includes a Referer header in which information indicating a link source from which the user agent makes a request to the device is described.
The determination unit
A relay apparatus comprising: comparing a URI included in the request with a part of a URI included in the Referer header, and determining that the request is to be rewritten when the two do not match.
In response to a request from a user agent belonging to an external network, a relay device relays the request to any of a plurality of devices belonging to the internal network and forwards a response from the relayed device to the user agent. A relay method to execute,
A determination step of determining whether to rewrite the request;
A rewriting step for rewriting the request when it is determined that it is necessary to rewrite in the determination step;
Refer to the device information registration table in which the relationship information between the public path name for publishing the contents possessed by the plurality of devices and the virtual path name indicated on the directory of the relay device is stored. The requesting device and the public path name of the device are identified from the virtual path name included in the request rewritten by the rewriting step, and the request is converted into a format suitable for the directory structure on the device side. And a request issuance transfer step for transferring to the requesting device.
A response issuance transfer step for transferring the response to the filter unit when a response is received from the requested device;
A relay method characterized by the above.
In response to a request from a user agent belonging to an external network, the computer relays the request to one of a plurality of devices belonging to the internal network, and forwards a response from the relayed device to the user agent. A program for functioning as a relay device,
The computer,
A determination unit for determining whether to rewrite the request;
A processing unit that rewrites the request when it is determined that the determination unit needs to be rewritten;
A device information registration table in which relation information between a public path name for publishing contents possessed by the plurality of devices and a virtual path name indicated on the directory of the relay device is stored;
Referring to the virtual path name in the device information registration table, the request destination device and the public path name of the device are specified, and the request is converted into a format suitable for the directory structure on the device side. A request issuance transfer unit for transferring to the device;
A response issuance transfer unit that, when receiving a response from the requested device, transfers the response to the filter unit;
Program to make it function.

Images