JP5234391B2 - COMMUNICATION DEVICE, PROGRAM, AND NETWORK SELECTION METHOD - Google Patents

Description

  The present invention relates to a routing technique between a communication terminal and a network, and more particularly to a routing technique in a system in which various communication formats are applied.

  In recent years, with the spread of wireless networks such as mobile phone networks and wireless local area networks (LANs), use cases for connecting mobile terminal devices to a variety of networks are increasing. Examples of the communication method include a cellular method, wireless LAN, PHS (Personal Handy-phone System) connection, and Bluetooth. As a network to be connected, there are a cellular network, the Internet, a corporate intranet, a home LAN, and a VPN (Virtual Private Network) connection to these.

  When connecting a terminal to multiple networks using IP (Internet Protocol), multiple routes are used. In this case, it is necessary to determine a route (routing) so that a packet from the terminal is sent to an appropriate network.

  As a technique related to routing, for example, there is one described in Patent Document 1 described later. This document describes a method in which a router connected to a plurality of terminals captures a source address from a received packet from the terminal and sends the received packet to a logical channel previously associated with the source address.

Japanese Patent Application No. 2006-109253 discloses a routing technique in a system that can use a plurality of communication means. In Japanese Patent Application No. 2006-109253, a communication terminal selects an optimal network by using a communication application to be executed and information on a connection destination, and an application characteristic database and a network characteristic database having information set in advance for the communication application. The method is described.
JP-A-10-173703

  However, in the method of selecting a packet route based on the source address of a packet as in the method described in Patent Document 1, a case where a terminal executes a plurality of applications using a plurality of networks. It is difficult to deal with. This is because, in such a case, the router of the same document controls to send packets having the same source address to the same channel.

  On the other hand, according to the method described in Japanese Patent Application No. 2006-109253, the above case can be dealt with. However, since the terminal for executing the application is provided with a network selection means for the application, the terminal software is modified. There is a need to. Further, it is necessary to perform network selection processing in parallel with the execution of the application in the terminal. Therefore, there arises a disadvantage that it is difficult to reduce the processing load on the terminal.

  The present invention has been made in view of the above problems, and an object thereof is to provide a technique for connecting a communication flow from a terminal to an appropriate network without modifying the software of the terminal. .

The communication apparatus according to the present invention detects communication information for relaying a communication flow between a plurality of terminals and a plurality of networks, and feature information for identifying the characteristics of the communication flow to which the packet belongs from the packet from each terminal. A flow identification means, a recording means for recording information on a detection result by the flow identification means and information on a network to which the packet is transmitted, in a database, and a network satisfying a predetermined condition relating to packet characteristic information from each terminal. Network selection means for searching the database and selecting a network to which the packet is to be transmitted from the search results, and the feature information includes at least information based on characteristics of the communication application of the transmission source and the terminal of the transmission source. includes information based on the characteristics, the network selection means, If the database contains an entry for the same connection destination as the connection destination included in the communication flow characteristic information, the database and the communication flow characteristic information are linked to the entry. If the average communication amount of the communication flow is estimated and the entry for the same connection destination is not included in the database, information other than the connection destination included in the feature information of the communication flow The average traffic associated with the matching entry is estimated as the average traffic of the communication flow, and the network that satisfies the bandwidth with respect to the estimated average traffic is selected .

The network selection method according to the present invention is information for identifying characteristics of a communication flow to which a packet belongs from a packet received from the terminal when relaying a communication flow between the plurality of terminals and the plurality of networks. Information based on characteristics of the communication application of the transmission source and feature information that is information based on characteristics of the terminal of the transmission source are detected, information of the detection result is recorded in a database, and a predetermined condition related to the characteristic information is satisfied The network is searched from the database, and the network to which the received packet is to be transmitted is selected from the search results, the received packet is transmitted to the selected network, and information on the network from which the received packet is transmitted It was recorded in the database and the selection of the network In this case, the database and the communication flow feature information are compared, and if the database includes an entry for the same connection destination as the connection destination included in the communication flow feature information, the database is associated with the entry. If the average communication amount attached is estimated as the average communication amount of the communication flow and an entry for the same connection destination is not included in the database, the connection destination included in the feature information of the communication flow is not included The average communication amount associated with the entries having the same information is estimated as the average communication amount of the communication flow, and the network that satisfies the bandwidth with respect to the estimated average communication amount is selected .

  According to the present invention, it is not necessary for a terminal that executes a communication application to have a function of selecting a network that propagates the communication flow. Thereby, it is not necessary to modify the terminal software, and the processing load on the terminal can be reduced. Furthermore, even when the terminal uses a plurality of networks in parallel by executing a plurality of applications, a network suitable for the communication flow of each application can be selected.

[Example 1]
FIG. 1 shows a configuration of a communication device 10 according to the first embodiment of the present invention. The communication device 10 is a device that relays a communication flow between a plurality of terminal devices 20 that execute the communication application 21 and a plurality of networks 30, respectively.

  The terminal device 20 is an arbitrary communication terminal and is connected to the communication device 10 by wire or wireless. The terminal device 20 uses a communication service provided by a server device (not shown) connected to the network 30 by executing the communication application 21. Examples of the communication application 21 include a web browser, a mail client, a file transfer program, a VoIP (Voice over IP) communication application, a video / audio streaming playback application, an online game, and the like.

  The network 30 provides a communication path used when the terminal device 20 executes the communication application 21, and specifically includes a wired LAN, a wireless LAN, a public telephone network, a mobile phone network, a PHS (Personal Handy -phone System) network.

  As shown in FIG. 1, the communication device 10 includes a communication unit 11, a flow identification unit 12, a recording unit 13, a flow characteristic database 14, a network characteristic database 15, and a network selection unit 16.

  The communication unit 11 is responsible for connection between the terminal device 20 and the network 30. The protocol used for the connection is mainly TCP / IP, but may be a virtual network connection by VPN (Virtual Private Network). The flow identification means 12 detects feature information for identifying the characteristics of the communication flow to which the packet belongs from the packet received from the terminal device 20. The recording unit 13 records the detection result of the flow identification unit 12 and information on the network 30 that has transmitted the packet from the terminal device 20 in the flow characteristic database 14. The network characteristic database 15 stores information such as the communication bandwidth and communication charge of each network 30.

  The network selection means 16 selects the network 30 to which the packet from the terminal device 20 should be transmitted. In selecting the network 30, the network selection means 16 searches the network characteristic database 15 for a network 30 that satisfies a predetermined condition based on the packet feature information, and selects an optimal network 30 from the search results.

  FIG. 2 shows an example of information held in the flow characteristic database 14. Information representing the characteristics of each communication flow is recorded in the flow characteristic database 14. The communication flow has different characteristics depending on the communication application 21 and the terminal device 20 that executes the communication application 21. In the example shown in the figure, information of the transmission source host and the connection destination host, the protocol type, the average communication amount, the time stamp indicating the latest communication time, and the type of the network 30 used are recorded. The recording unit 13 learns and updates information in the flow characteristic database 14 based on the detection result of the flow identification unit 12 and the selection result of the network selection unit 16.

  FIG. 3 shows an example of information held in the network characteristic database 15. The network characteristic database 15 stores information related to predetermined communication characteristics of the network 30. In the illustrated example, the type of network 30, maximum bandwidth, security level, and billing system are set.

  The operation of the flow identification unit 12 will be described with reference to the flowchart shown in FIG. When the communication unit 11 receives a packet from the terminal device 20, the flow identification unit 12 detects characteristic information such as a source address and a port number for identifying the protocol type (step S1). When the flow identification means 12 can identify the characteristics of the communication flow to which the packet belongs from the detected source address and port number (step S2: Yes), the flow identification means 12 outputs the characteristics of the communication flow (step S5).

  On the other hand, if it is determined that the characteristics of the communication flow cannot be identified by the address and the port number (step S2: No), it is determined whether or not the packet is of the same flow as the previously received packet (step S3). Specifically, for example, when a packet indicating the start of a new session is received within a predetermined time after receiving a packet indicating the end of a session, the latter communication flow belongs to the same communication flow as the former Is determined. In this case, the flow identification means 12 outputs that it is the same as the previous identification content (step S5). Note that the reception interval of packets of the same flow can be determined from the time stamp in the flow characteristic database 14.

  Further, when the protocol is HTTP, there is a tendency that consecutive requests are generated under the same or similar conditions such as acquisition of an image file or acquisition of a file from a link destination. Therefore, using such a tendency, it may be determined whether or not a series of packets belong to the same flow.

  If the flow identification unit 12 cannot identify the flow characteristics by the source address or the port number and the current packet does not belong to the same communication flow as the previous time (step S3: No), the feature information is included in the packet. It is determined whether or not other information that can be described is described (step S4).

  The above determination is applied, for example, when the communication flow protocol is HTTP. In the case of HTTP, the communication device 10 operates as a transparent proxy, but for example, the situation where the same terminal device 20 is executing different communication applications 21 may be recognized from the source address and port number. Can not. Therefore, the flow identification unit 12 refers to terminal information such as User-Agent in the header of the received HTTP request, determines the HTTP communication application 21 executed by the terminal device 20, and identifies the communication flow. Also, by referring to the Server header of the HTTP request and determining whether it is a response of a specific server, or by determining whether or not there is a specific description in the request body, the communication flow is identified. Also good.

  When the information such as the above-mentioned User-Agent is obtained (step S4: Yes), the flow identification unit 12 outputs the characteristics of the communication flow determined based on the information (step S5). If the above information cannot be obtained (step S4: No), a message indicating that the characteristics of the communication flow cannot be identified is output (step S6).

  The operation of the network selection means 16 will be described with reference to the flowchart shown in FIG. The network selection unit 16 selects the network 30 to which the packet received this time is to be transmitted based on the information output from the flow identification unit 12 regarding the communication flow. Here, a case will be described in which communication security is ensured as a selection condition and a network 30 with a lower charge is selected.

  The network selection means 16 optimizes from the network characteristics database 15 (FIG. 3) using the information of the flow characteristics database 14 (FIG. 2) when the communication flow characteristics are identified by the flow identification means 12 (step S11: Yes). A network 30 is selected (step S12). As an example, assuming that the communication application 21 of the terminal device 20 communicates with “streaming.media.com” using “RTSP”, the network selection unit 16 first determines the communication flow from the information of FIG. Is estimated to be “1024 kbps”.

  If there is no entry related to the target connection destination in the flow characteristic database 14, the one closest to the condition is selected. For example, when “terminal A” tries to connect to “www.example.com” by “HTTPS”, in the example shown in FIG. 2, “www.example.com” does not exist at the connection destination. In this case, the network selection unit 16 searches for an entry having a combination of the current source “terminal A” and the protocol “HTTPS”. As a result, the entry of the connection destination “www.yahoo.com” is hit, and the average traffic is estimated to be “32 kbps”.

  Next, the network selection means 16 selects a network satisfying the current condition from the network characteristic database 15 using the estimated average traffic as a key. As an example, a case is assumed in which the average traffic is estimated to be “1024 kbps” by the flow characteristic database 14.

  In the above case, referring to the network characteristic database 15 of FIG. 3, the “radio spot” and the “intranet” satisfy the bandwidth for the average communication amount “1024 kbps”. Among them, “Intranet” is cheaper. Furthermore, regarding security guarantee, the security of the “wireless spot” is “low” and the “intranet” is “high”. The network selection means 16 finally selects “Intranet” in view of the above-mentioned condition that security is ensured and a network 30 with a lower charge is selected.

  When the network selection unit 16 determines the network 30 as described above (step S13), the network selection unit 16 notifies the communication unit 11 of the determined information on the network 30 (step S14). If the flow identification unit 12 determines that the current received packet belongs to the same communication flow as the previous received packet, the same network as that applied to the previous packet is notified.

  On the other hand, if the characteristics of the communication flow are not identified by the flow identifying means 12 for the packet received this time (step S11: No), a preset standard network is notified to the communication means 11 (step S15). .

  The communication unit 11 sends the received packet this time to the network 30 notified from the network selection unit 16. The recording means 13 records the information of the network 30 to which the packet is transmitted and the time stamp of the communication time in the network characteristic database 15.

  Note that the network to be notified from the network selection unit 16 to the communication unit 11 may be notified to a plurality of ranked network candidates instead of being limited to one as in the above procedure. Specifically, in the above example, “Intranet” that meets the most conditions is set as the first candidate, and “Radio spot” is set as the second candidate. Thereby, even if a failure occurs in the first candidate during packet transmission, it is possible to control to switch to the second candidate.

  Also, depending on the communication application 21, security may be indispensable, so a network that uses a low-security protocol such as the above-mentioned “wireless spot” may be set in advance so that it is not listed as a candidate. . Here, the low-security protocol refers to a protocol that does not encrypt passwords and confidential information, such as POP (Post Office Protocol), SMTP, and telnet.

  Next, the operation of the recording means 13 will be described. The recording unit 13 analyzes the contents output by the flow identification unit 12 regarding the communication flow of the received packet and the selection result of the network selection unit 16 to obtain statistics, and updates the information of the flow characteristic database 14 using the statistics. . For example, if the terminal device 20 performs “2048 kbps” communication this time with “download.media.com”, this history is recorded in the flow characteristic database 14. Thereby, the bandwidth required by this flow is recorded. In addition, the time stamp of the flow characteristic database 14 is updated at the time of connection to the network 30, and it is possible to determine whether or not a series of packets belong to the same communication flow.

  According to the present embodiment, an optimal network can be selected for each communication flow generated by the execution of the communication application 21 by the terminal device 20. In addition, since the communication device 10 between the terminal device 20 and the network 30 performs selection of the network, the processing load on the terminal device 20 can be reduced, and further, software modification in the terminal device 20 becomes unnecessary.

  In the system as shown in FIG. 1, a plurality of terminal devices 20 may request connection to the same connection destination, or a single terminal device 20 may execute a plurality of communication sessions in parallel. In these cases, the communication device 10 individually selects the network 30 for each communication session. For example, when the Web browser accesses a plurality of servers in parallel, the communication device 10 performs network determination processing and learning processing for each session.

  For protocols that use different ports for control and data, such as FTP and VoIP, different conditions may be set for each port. For example, the control packet is a flat-rate system and uses a PHS with a wide communication area, and the fastest network available at that time is selected for the data packet. As a result, a standby fee is charged for a fixed amount, and a pay-per-use network is used only when communication occurs, so communication charges can be reduced.

Further, by analyzing the communication content of the control packet and predicting the flow type of the data packet to be executed in the future, the control packet and the data packet may be controlled to flow on separate paths. .
[Example 2]

  FIG. 6 shows the configuration of the communication device 10 according to the second embodiment of the present invention. The configuration of the communication device 10 of this embodiment is the same as that obtained by adding the candidate notification means 17 to the configuration of the above-described first embodiment (FIG. 1).

  The candidate notification means 17 notifies the terminal device 20 of the network candidates searched by the network selection means 16, and functions to recognize what the user has selected from the candidates. The network candidates indicate a plurality of network candidates ranked as described above.

  With reference to the flowchart shown in FIG. 7, the operation of the network selection means 16 of this embodiment will be described. When the communication flow cannot be identified by the flow identifying means 12 (step S21: Yes), a preset standard network is notified to the communication means 11 as in the above-described embodiment (step S25).

  On the other hand, when the communication flow can be identified (step S21: Yes), the network selection unit 16 searches for a network using the flow characteristic database 14 and the network characteristic database 15 in the same manner as in the above-described embodiment. Then, the candidate obtained from the search result is output to the candidate notifying means 17 (step S22). Using the above example, the first candidate “intranet” and the second candidate “wireless spot” are output.

  Candidate notifying means 17 notifies network candidates to terminal device 20 by means of communication means 11, and then receives candidates designated by terminal device 20 and outputs them to network selecting means 16.

  When the network selection means 16 recognizes the network candidate designated from the terminal device 20 (step S23), it notifies the communication means 11 of it (step S24). The communication unit 11 sends the packet to the designated network 30.

  According to the present embodiment, it is possible to control to select a network that meets the user's request from the candidates searched by the network selection means 16. Thereby, since the user can be given an opportunity to select the network used for executing the communication application 21, the convenience of the user is improved.

FIG. 8 shows how the recording means 13 of this embodiment corrects information in the flow characteristic database 14. In the example shown in the figure, the communication application 21 of the connection destination “terminal A” connects to “www.yahoo.com” using “HTTPS” as the “intranet” that has been used in the same type of flow until then. Instead of “”, the recording is performed when “cellular” is designated by the user. In addition, the time stamp of the communication is updated to “20:00:59”.
[Example 3]

  FIG. 9 shows the configuration of the communication apparatus 10 according to the third embodiment of the present invention. The configuration of the communication apparatus 10 according to the present embodiment is equivalent to the configuration in which the address conversion means 18 is added to the configuration of the first embodiment (FIG. 1). The address conversion means 18 functions to rewrite the IP address and port number used in the terminal device 20 to the IP address and port number used on the network 30 side. Note that this embodiment may be combined with the above-described second embodiment in which the candidate notification means 17 is provided.

  With reference to FIG. 10, the address conversion process in the present embodiment will be described. Here, it is assumed that local addresses “A”, “B”, “C”, and “D” are assigned to the four terminal devices 20. The communication device 10 is connected to four types of networks 30 to which global addresses “W”, “X”, “Y”, and “Z” are assigned. The local address is an address arbitrarily assigned to the terminal device 20 by the communication device 10, and the global address is assigned by each network 30.

  From FIG. 10, for example, when connecting from the terminal device 20 with the local address “A” to the network 30 (PHS) with the global address “Z”, the communication device 10 receives the local address “A” in the received packet from the terminal device 20. To the global address “Z”. In addition, when the communication apparatus 10 also includes the above-described candidate notification unit 17 and the network 30 of the global address “X” (WiMax) is designated by the user, the address conversion unit 18 changes the global address to be rewritten from “Z” to “ Change to "X".

  According to the present embodiment, for example, even when a plurality of terminal devices 20 are simultaneously connected to a plurality of networks 30, it is possible to smoothly relay packets. The local address given to the terminal device 20 can also be used for flow identification.

  The present invention is not limited to being implemented as the communication device 10 that relays between the terminal device 20 and the network 30 as in the above embodiment, but may be implemented as a program corresponding to the operation of the communication device 10.

It is a block diagram which shows the structure of the communication apparatus in Example 1 of this invention. It is explanatory drawing of the information stored in the flow characteristic database of an Example. It is explanatory drawing of the information stored in the network characteristic database of an Example. 4 is a flowchart illustrating an operation of a flow identification unit according to the first embodiment. 4 is a flowchart illustrating an operation of a network selection unit according to the first exemplary embodiment. It is a block diagram which shows the structure of the communication apparatus in Example 2 of this invention. It is a flowchart which shows operation | movement of the network selection means of Example 2. It is explanatory drawing regarding the update of the flow characteristic database by the recording means of Example 2. FIG. It is a block diagram which shows the structure of the communication apparatus in Example 3 of this invention. It is explanatory drawing regarding the address conversion in Example 3. FIG.

Explanation of symbols

10 Communication equipment
11: Communication means, 12: Flow identification means, 13: Recording means, 14: Flow characteristic database, 15: Network characteristic database, 16: Network selection means, 17: Candidate notification means, 18: Address conversion means
20 Terminal equipment
21 Communication applications
30 network

Claims (13)

A communication means for relaying a communication flow between a plurality of terminals and a plurality of networks;
Flow identification means for detecting feature information for identifying characteristics of a communication flow to which the packet belongs from a packet from each terminal;
A recording unit for recording information on a detection result by the flow identification unit and information on a network to which the packet is transmitted in a database;
Network selection means for searching a network satisfying a predetermined condition related to feature information of a packet from each terminal from the database and selecting a network to which the packet is to be sent out from the search results;
The feature information includes at least information based on characteristics of the communication application of the transmission source and information based on characteristics of the terminal of the transmission source,
The network selection unit compares the database and the feature information of the communication flow, and if the database includes an entry for the same connection destination as the connection destination included in the feature information of the communication flow, the network selection unit When the average communication amount associated with the entry is estimated as the average communication amount of the communication flow, and the entry for the same connection destination is not included in the database, it is included in the feature information of the communication flow Estimating the average traffic associated with the entry with the matching information other than the connection destination as the average traffic of the communication flow, and selecting the network satisfying the bandwidth with respect to the estimated average traffic It said communication apparatus. In the database, a fee generated when a network is used is recorded for each network,
The network selection means is a network that meets the bandwidth for average traffic which the estimated, and, according to claim 1, characterized in that the most the rates in the network that satisfies selects the cheapest network The communication device described. In the database, the network security level is recorded for each network,
The network selecting means selects a network that satisfies a bandwidth with respect to the estimated average traffic and that has the highest security level in the satisfied network. The communication apparatus according to 1 .   The network selection means determines that a series of packets having common feature information and a reception interval within a predetermined time belong to the same communication flow, and applies the same network to transmission of each packet belonging to the same communication flow The communication apparatus according to claim 1, wherein: The communication apparatus according to any one of claims 1 to 4 , wherein the flow identification unit detects characteristic information of a packet representing an HTTP request from a header of the packet. The communication apparatus according to claim 5 , wherein the flow identification unit detects terminal information from the header of the packet as the characteristic information of the packet representing the HTTP request. It said recording means, a communication device according to any one of claims 1 to 6, wherein the updating by the learning process the information in the database. In addition, a candidate notification means for externally outputting a candidate for a network to which a packet is transmitted and recognizing a selection result input with respect to the candidate,
The network selection unit notifies the candidate notification unit of the search result of the database as the candidate, and applies the network as the selection result recognized by the candidate notification unit to the transmission of the packet. communication apparatus according to any one of 1 to 7. Further, in any one of claims 1 to 8, characterized in that it comprises an address conversion means for converting in response the address of the packet in the communication flow to be relayed by the communication means to the network to which the network selection means selects The communication device described. A program that causes a computer to function as the communication device according to any one of claims 1 to 9 . When relaying communication flows between multiple terminals and multiple networks,
Information for identifying the characteristics of the communication flow to which the packet belongs from the packet received from the terminal, and at least information based on the characteristics of the communication application of the transmission source and characteristic information that is information based on the characteristics of the terminal of the transmission source Detect
Record the information of the detection results in a database;
Search the network for a predetermined condition related to the feature information from the database, and select a network to which the received packet is to be transmitted from the search results;
Sending the received packet to the selected network;
Record the network information to which the received packet was sent in the database ;
When selecting the network, the database and the communication flow characteristic information are compared, and the database includes an entry for the same connection destination as the connection destination included in the communication flow characteristic information. If the average communication amount associated with the entry is estimated as the average communication amount of the communication flow, and the entry for the same connection destination is not included in the database, the feature information of the communication flow The average traffic associated with the entry that matches the information other than the connection destination included in is estimated as the average traffic of the communication flow, and the network satisfying the bandwidth with respect to the estimated average traffic is selected. features and to Rene Ttowaku selection method to. In the database, a fee generated when a network is used is recorded for each network,
When the network is selected, a network satisfying a bandwidth with respect to the estimated average traffic is selected, and a network with the lowest charge is selected in the satisfied network. network selection method according to claim 1 1, wherein. In the database, the network security level is recorded for each network,
When the network is selected, a network satisfying a bandwidth with respect to the estimated average traffic is selected, and a network having the highest security level among the satisfied networks is selected. network selection method according to claim 1 1 wherein the.

Images