JP3558908B2 - ACCESS POINT SELECTION CONNECTION SYSTEM AND STORAGE MEDIUM RECORDING THE PROGRAM - Google Patents

Description

【００７３】
実際のインターネットでは、上記のように複雑なネットワークによって構成されており、どの経路を通過するかは、それぞれのルーターに設定されているルーティングテーブルによって決定される。従って、端末から宛先は、ＲｏｕｔｅｒＡ→ＲｏｕｔｅｒＣ→ＲｏｕｔｅｒＥという経路を通過するとは限らない。また、端末から宛先までの経路が一意に決定されたとしても、その経路上にない装置からは通常、トラフィックや混雑状況の計測はできない。
【００７４】
そこで、ｔｒａｃｅ ｒｏｕｔｅを用いて、センターから各ドメイン（宛先）までの経路を取得し、その間のネットワーク機器（Ｒｏｕｔｅｒ）のアドレスを入手する。
【００７５】
（例）
宛先１……ＲｏｕｔｅｒＦ（１０．１０．２６．１）→ＲｏｕｔｅｒＢ（１０．１０．２１．１）→ＲｏｕｔｅｒＤ（１０．１０．２３．１）→ＲｏｕｔｅｒＥ（１０．１０．２４．１）→宛先１（１０．２０．１．１）
宛先２……ＲｏｕｔｅｒＦ（１０．１０．２６．１）→ＲｏｕｔｅｒＢ（１０．１０．２１．１）→ＲｏｕｔｅｒＤ（１０．１０．２３．１）→宛先２（１０．２０．２．１）
また、センターからプロバイダーのアクセス装置までの経路も同様に入手できる。
【００７６】
（例）アクセスサーバー……ＲｏｕｔｅｒＦ（１０．１０．２６．１）→ＲｏｕｔｅｒＡ（１０．１０．２０．１）→アクセスサーバー（１０．１０．１０．１）
これと同時に、各ルーター等の間の遅延時間が取得できる。情報を整理すると以下のようになる（／＝なし、＊＝不明、−＝接続無し）。
【００７７】
【表２】

【００７８】
上記の結果から、不明な部分はＲｏｕｔｅｒＡ←→ＲｏｕｔｅｒＢ、ＲｏｕｔｅｒＢ←→ＲｏｕｔｅｒＣである。
【００７９】
このような状態で空白となる部分が発生するが、センター側で複数の異なるプロバイダー等のアクセスポイントを利用契約しておくことで、同一の宛先に対しても異なる経路を経由させることが可能となり、この結果、得た情報によって大部分を埋めることが可能となる。
【００８０】
宛先１へアクセスする場合の経路が下記の２通りある場合、端末がアクセスするアクセスサーバーは次のようにして選択される。
【００８１】
経路１：アクセスサーバー１→ＲｏｕｔｅｒＡ→ＲｏｕｔｅｒＣ→ＲｏｕｔｅｒＥ→宛先１
経路２：アクセスサーバー２→ＲｏｕｔｅｒＢ→ＲｏｕｔｅｒＤ→ＲｏｕｔｅｒＥ→宛先１
前述したルーター間延滞時間より、経路１は５ｍｓ＋５ｍｓ＋５ｍｓ＝１５ｍｓとなり、経路２は５ｍｓ＋１０ｍｓ＋３５ｍｓ＝５０ｍｓとなる。これにより、経路１を選択した方がより早くアクセスすることが可能であるため、センターはアクセスサーバー１への接続をユーザー端末に通知する。
【００８２】
センター側で必要なＤＢの構成は、主として以下のようなものがある。物理的な配置は負荷等を考慮して分散しても良い。
・トラフィック情報キャッシュテーブル
（リクエストに応じて検索されたトラフィック情報に重み付けを行って一定期間キャッシュする。）
ＵＲＬ、ＩＰアドレス、スループット、プロバイダー通番、アクセスポイント番号、作成日時、生存時間、レスポンス時間、論理ボトルネック速度、アクセス回数
・トラフィック情報レスポンステーブル
（エージェントサーバーから一定のノードまでのレスポンスを測定し、蓄積する。）
プロバイダー通番、アクセスポイント番号、宛先ノード通番、レスポンス時間、計測日時、被参照回数
・ノードマップ管理テーブル
ノード通番、ノード名、インターフェース数、ノード種別
・ノードマップ回線情報テーブル
ノード通番、インターフェース通番、インターフェースＩＰアドレス、回線速度、回線使用率
・ルーティング情報管理テーブル
ノード通番、Ｄｉｓｔｎａｔｉｏｎ数、ルーティングプロトコル種別（Ｓｔａｔｉｃ、ＢＧＰ、ＲＩＰ、ＩＧＲＰ、……複合）
・ルーティング情報テーブル
ノード通番、Ｄｉｓｔｎａｔｉｏｎアドレス、Ｃｏｓｔ／Ｍｅｔｒｉｃ種別、Ｃｏｓｔ／Ｍｅｔｒｉｃ値
・プロバイダー情報テーブル
プロバイダーＩＤ、プロバイダー名、ドメイン名、アクセスポイント数、補正係数
・プロバイダーアクセスポイントテーブル
プロバイダー通番、アクセスポイント番号、名称、電話番号、アクセスサーバーノード番号、補正係数
・ユーザー情報テーブル
ユーザーＩＤ、ユーザーステータス、その他ユーザー情報（住所、氏名、決済情報等）
・ユーザー課金テーブル
ユーザーＩＤ、対象日、利用回数
・ユーザー決済テーブル
ユーザーＩＤ、対象年月、利用回数、金額、補正金額、決済ステータス
・ユーザー登録情報テーブル
ユーザーＩＤ、（対象プロバイダーＩＤ、対象アクセスポイント番号）×ｎ
（対象とする接続先数を限定しない場合、テーブル数は複数となる）
・ユーザー認証テーブル
ユーザーＩＤ、登録電話番号、パスワード
・課金管理テーブル
基本料金、問い合わせ単価
・月別課金管理テーブル
補正金額
この他、プロバイダー毎や宛先毎等の問い合わせ状況を集計する統計管理テーブルや、システム／サービスのステータスを有するシステム管理テーブル、エージェントサーバーのスケジューリングを行うエージェント制御テーブル等がある。
【００８３】
【発明の効果】
以上説明したように、本発明によれば、
・エンドユーザーから見たスループットが向上する、
・エンドユーザーは、接続してから遅いことに気付いて接続先を変える場合、最低で１０円のコストがかかっていたが、この仕組みならば、パケットの料金＋サービス料金＝数円／回のコストで確実に速度の出る接続先を選択することができる、
・インターネット全体の負荷を分散できる、
・小規模プロバイダーが活躍できる分野を創出できる、
・ルーティング情報提供センターを設立して事業化できる、
・複数プロバイダーを束ねた形での契約形態を提供することで、ユーザーからは１契約で効率的な接続を行うことができ、プロバイダー側では回線や設備の効率的な利用、溢れることで失っていた顧客と接続料金を得ることができる、
・話中等で混雑時の利用を諦めていたユーザーが接続可能となることによって、トラフィック収入の増加を見込める、
等である。
【００８４】
また、複数のプロバイダーだけではなく、単一プロバイダーや、会社でのリモートアクセスでも、アクセス回線の話中状況やトラフィックに応じた応答をセンターが返す仕組みとすれば適用可能であり、特に大規模プロバイダーでは計画的かつ効率的な設備投資及び運用が可能となる。
【００８５】
さらに本発明はＩＳＤＮのＤチャネルパケットだけではなく、Ｂチャネルパケットや、ＡＴＭ、ＦＲ等の多重化が可能なサービスに対しても適用可能である。
【図面の簡単な説明】
【図１】本発明のシステム全体の概要図
【図２】本発明システムにおける端末接続装置の動作の概要を示すシーケンス図
【図３】端末接続装置の実施の形態の一例を示す構成図
【図４】端末接続装置の実施の形態の他の例を示す構成図
【図５】端末接続装置の実施の形態のさらに他の例を示す構成図
【図６】ルーティング情報提供センターの実施の形態の一例を示す構成図
【図７】本発明システムにおけるルーティング情報提供センターの動作の概要を示すシーケンス図
【図８】ルーティング情報提供センターの実施の形態の他の例を示す構成図
【符号の説明】
１：インターネット、２，２０：ユーザー端末、３：端末接続装置、４：ルーティング情報提供センター、２１：アプリケーション、２２：ソケット、２３：ネットワークプロトコル、２４：ネットワークドライバ、２５：ハードウェアドライバ、２６：ＥＰＲ用ソケット、２７：ＮＡＴ機能部、２８：ＥＰＲ制御ドライバ、３０，４０：ターミナルアダプタ、３１，４１：既存機能部、３２，４２：設定情報ＤＢ、４３：端末制御機能部、４４：ＮＡＴ機能部、４５：ＰＰＰ／ＩＰ−シリアル変換部、４６：ＥＰＲ制御機能部、５０：ルーター、５１：ＬＡＮ側インタフェース（Ｉ／Ｆ）、５２：ルーティング機能部、５３：ＴＡ機能部、５４：ＮＡＴ機能部、５５：設定情報ＤＢ（データベース）、５６：ＥＰＲ制御機能部、６１，６１’：レスポンス調査エージェントサーバー、６２，６２’：ＤＢ（データベース）／エンジンサーバー、６３：通信サーバー、６４：ユーザー管理サーバー、６５，６５’：管理端末。[0001]
TECHNICAL FIELD OF THE INVENTION
According to the present invention, when connecting to a certain network, of a plurality of access points available to a user for the network, a congestion state to a communication partner (destination destination) in the network and a line state to the plurality of access points The present invention relates to an access point selection and connection system for selecting and connecting an optimum access point according to the above, and a storage medium storing the program.
[0002]
[Prior art]
Currently, when a user attempts to access a specific domain by dial-up connection to the Internet, the user is required to reach a contracted Internet Service Provider (ISP) (hereinafter, abbreviated as a provider) (or a provider thereof). A connection was made via a telephone line (up to an access point), and then a specific domain was accessed via the Internet by specifying a domain name (address).
[0003]
Also, regarding dial-up connection to an intranet of a company or the like, a user connects to the access point via a telephone line, and then designates a name (address) of a target server or the like to connect to the intranet via the Internet. Had access to the intended server.
[0004]
[Problems to be solved by the invention]
However, regarding the congestion status of the network beyond the access point and the status of the telephone line (access line) to the access point itself (whether busy or idle), until the user actually connects or performs the connection operation. I couldn't know. For this reason, conventionally, even if the user has signed up with multiple providers or knows the phone numbers of multiple access points, the connection or connection operation is performed at least once, and the transfer speed is low. If you do not recognize that you are busy, you will not be able to change the phone number of the provider or access point and reconnect, and you will be forced to perform complicated operations. However, there is a problem that extra cost is required.
[0005]
In terms of providers, the oligopoly situation of large-scale providers who have thick lines in various areas is progressing, and small providers share a large number of lines even if they have thick lines only in specific areas. There was a problem that it was difficult to obtain, and it was being eliminated.
[0006]
Furthermore, large-scale providers may make capital investments at the peak hours to reduce the busy rate, and may have multiple access points in the city or adjacent areas. Since the change requires a complicated operation on the user side, it is difficult for the user to efficiently use the equipment, and as a result, the utilization rate of the equipment is reduced.
[0007]
An object of the present invention is to enable a user to select an optimal access point among a plurality of available access points and connect to a network without burdening the user with complicated operations and extra costs. It is an object of the present invention to provide an access point selection / connection system and a storage medium storing the program.
[0008]
[Means for Solving the Problems]
According to the present invention, there is provided an access point selection / connection system for selecting one of a plurality of access points available to a user for a certain network and connecting a user terminal to the network to achieve the above object. The addresses of the above access points, servers, domains, nodes, etc., and the congestion status between them are periodically collected and stored, and in response to an inquiry request including the address of the communication partner from the user, the storage is performed. A routing information providing center for predicting a route and a throughput connecting each access point and a communication partner based on the information, and selecting and responding to an optimum access point among a plurality of access points available to the user according to the information, Receives a connection request including the destination address from the user terminal At this time, a terminal for sending to the routing information providing center as an inquiry request including the address of the communication partner from the user, receiving a response from the routing information providing center, and connecting the user terminal to the network via the selected access point The communication between the routing information providing center and the terminal connection device is performed using a logically multiplexable line such as an ISDN line D-channel packet service.
[0009]
According to the present invention, the optimum access point can be recognized before the actual connection, so that it is unnecessary to connect the terminal to another access point by noticing that the transfer speed is low during communication. Since no real cost is required and the actual connection can be performed by the terminal connection device, the user is not forced to perform complicated operations.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
FIG. 1 shows an outline of the whole system when the present invention is applied to the Internet.
[0011]
This system (End Point Routing: hereinafter, abbreviated as EPR) is roughly divided into any one of a plurality of access points (contracted providers A, B, and C) that can be used by a user for a network (Internet) 1. And the terminal connection device 3 for connecting the user terminal 2 by selecting the address, congestion status of each provider, server, node, etc. on the Internet 1 and the congestion status. A routing information providing center 4 for predicting a route and throughput for accessing the server a or b), selecting an optimal provider, and informing the terminal connection device 3 of them, and communication between them is provided by the D of the ISDN line. Using circuits that can be logically multiplexed such as channel packet services Divide.
[0012]
The terminal connection device 3 (hereinafter, simply referred to as a connection device), as described later,
1. Implement software or hardware that operates under the OS on the user terminal 2;
2. Adding software or hardware to a terminal adapter (TA) connected to a serial interface or the like of the user terminal 2;
3. By using software or hardware (existingly implemented Address Translation (NAT) function) for a WAN connection router such as a dial-up connected to the user terminal 2 via a LAN, a newly added portion is minimized. ),
This is achieved by:
[0013]
In any method, in order to minimize the influence on the user interface and the application in the user terminal 2, a proxy response to the terminal (or the application) 2 and an address conversion function after connection are provided, and the routing information providing center ( Hereinafter, it is simply referred to as a center.) The system is not made aware of the connection with the center 4.
[0014]
FIG. 2 shows the outline of the operation in the connection device 3 including the exchange with the user terminal (or its application) 2, the center 3, and the provider.
[0015]
(1) Upon receiving a connection request (destination address, URL, etc.) from the user terminal 2, the connection device 3 returns a connection response including the address of the dummy (temporary) access point to the user terminal 2, and Then, a connection is made using a D-channel packet or the like, and an inquiry request including the information is made.
[0016]
(2) The connection device 3 receives from the center 4 a response of an inquiry result including an address (telephone number) and speed information of an optimal provider (access point: connection destination) among contracted providers.
[0017]
At this time, it is also possible to determine whether the received response is appropriate by comparing the result response with the setting information such as information on the contracted provider registered and accumulated in the database in advance. good.
[0018]
(3) The connection device 3 makes an actual line connection using a B channel or the like to the address of the selected provider, and when the connection is established, notifies the user terminal 2 of the completion of the connection in the same manner as in normal communication.
[0019]
In the line connection, a connection to an access point is made based on acquired information using a channel occupied such as a B channel or a line capable of logical multiplexing such as a D channel.
[0020]
(4) Thereafter, the connection device 3 relays the actual data communication between the user terminal 2 and the communication partner via the provider. At this time, the temporary access point and the actual access point (provider) Address conversion.
[0021]
When the address of the access point is dynamically changed, normally the application on the user terminal needs to be restarted, but in the present invention, it has been used for one-to-N connection. By implementing NAT (address translation function) on a one-to-one basis such as a normal terminal adapter or software on a user terminal, it is possible to use it without being aware of the user interface on the user terminal.
[0022]
Next, a configuration example of the connection device 3 will be described. Here, an example in which an ISDN line is mainly used is shown. However, even in a connection device such as an FR (frame relay) or an ATM, a low-speed logical path is provided in place of the D channel, and one physical line is provided according to information. The same effect can be obtained by controlling the logical paths and the like for a plurality of connection destinations.
[0023]
FIG. 3 shows an example of an embodiment of a connection device, here an example realized by software mounted on a user terminal. In the figure, reference numeral 20 denotes a user terminal incorporating the terminal connection device, and 30 denotes a normal TA. It is.
[0024]
The user terminal 20 includes an existing application 21, an existing socket 22, an existing network protocol (such as TCP / IP) 23, an existing network driver (such as PPP / NDIS) 24, and an existing hardware driver (such as serial / LAN). 25, an EPR socket 26, a NAT function unit 27, and an EPR control driver 28, and the TA 30 includes an existing function unit 31 and a setting information DB (database) 32 (however, not related to the present invention). The part is omitted.). In the drawing, solid arrows indicate control information before the IP connection is established, dotted arrows indicate control information after the IP connection is established, and double arrows indicate data flows after the connection is established.
[0025]
The existing application 21 requests the socket 26 for a request such as an IP address or a URL. The socket 26 returns a connected state to the application 21 and provides the EPR control driver 28 with request information such as an IP address and a URL. The NAT function unit 27 assigns a dummy IP address, performs mutual conversion with the IP address assigned on the ISDN line side, and does not make the application or driver aware of dial-up itself.
[0026]
The EPR control driver 28 monitors the connection status such as TA, and if the connection is not established / can be changed, obtains the destination IP address or the like from the application 21 via the socket 26 and inquires the center side. , Decrypt the response, and make a PPP / IP connection to the optimal access point on the D channel or the B channel. When data is directly exchanged between the drivers 24 and 25 without using the NAT function, the existing driver 24 is notified / recognized of the telephone number / IP address of the connection destination assigned by each provider, and the IP address is changed. (At this time, if the change of the IP address can be recognized like the current PPP driver, it is only necessary to dynamically notify the PPP driver of the change of the telephone number of the connection destination.)
[0027]
The existing function unit 31 of the TA 30 also performs standard processing such as control at the time of connection, response, and asynchronous / synchronous PPP conversion, but basically, data between the line side and a serial interface (I / F). Perform the relay. The setting information DB (database) 32 stores a user name (ID), a password, and the like registered in advance.
[0028]
FIG. 4 shows another example of the embodiment of the terminal connection device, here an example realized by additionally mounting software or hardware on a terminal adapter connected to a serial interface or the like of a user terminal. Reference numeral 40 denotes a terminal adapter incorporating a terminal connection device.
[0029]
The terminal adapter 40 includes an existing function unit 41, a setting information DB (database) 42, a terminal control function unit 43, a NAT function unit 44, a PPP / IP-serial conversion unit 45, and an EPR control function unit 46 (however, Parts not related to the present invention are omitted.)
[0030]
The existing function unit 41 also performs standard processing such as connection control, response, asynchronous / synchronous PPP conversion, etc., but basically relays data between the line side and a serial interface (I / F). Do. The setting information DB (database) 42 stores, in addition to a pre-registered user name and password, a pre-registered center, a contracted provider and a telephone number of the access point, and the like.
[0031]
The terminal control function unit 43 returns a dummy connection completion response to the connection request from the user terminal transmitted via the existing function unit 41, and assigns a dummy IP address to the NAT function unit 44. The NAT function unit 44 performs mutual conversion between an IP address assigned on the ISDN line side and a dummy IP address assigned on the serial side.
[0032]
The EPR control function unit 46 obtains the destination IP address using the NAT function unit 44, makes an inquiry to the center side, decodes the response, and establishes the PPP / IP connection to the optimal access point via the D channel or the B channel. Do.
[0033]
The user terminal 2 is a normal terminal, and controls the TA 40 such as calling using an AT command or the like via a serial interface or the like, and after establishing a connection, decodes all data and performs processing as IP.
[0034]
FIG. 5 shows still another example of the embodiment of the terminal connection device, here, an example in which software or hardware is additionally mounted on a router with a NAT function connected to a user terminal via a LAN, and is realized. , 50 are routers incorporating terminal connection devices.
[0035]
The router 50 includes an existing LAN interface (I / F) 51, an existing routing function unit 52, an existing TA function unit 53, an existing NAT function unit 54, a setting information DB (database) 55, and an EPR control function unit 56. (However, parts not related to the present invention are omitted.) In the drawing, solid arrows indicate control information before the IP connection is established, double arrows indicate data flows after the connection is established, and dotted arrows indicate flows of information added in correspondence with the EPR.
[0036]
When functioning as a normal router, the routing function unit 52 performs functions such as IP routing, control of routing information, and control of a dial-up destination (DDR) according to a destination. The TA function unit 53 also performs standard processing such as connection control, response, and asynchronous / synchronous PPP conversion, but basically interprets / controls the line-side protocol. The NAT function unit 54 interprets the IP / TCP / UDP header of the packet received from the client (user), and if the request is to the outside, sends the request source IP address to the address assigned to its own WAN side I / F. And sends the packet to the WAN side. For the packet received from the WAN side, the source of the request on the LAN side is determined based on the received port number and the like, and the destination address is converted as a packet addressed to the terminal on the LAN side to convert it to the LAN side. Send to
[0037]
The setting information DB (database) 55 stores pre-registered centers, contracted providers, telephone numbers of their access points, and the like, as well as pre-registered user names and passwords. The EPR control function unit 56 obtains the destination IP address using the NAT function unit 54, makes an inquiry to the center side, decodes the response, and establishes the PPP / IP connection to the optimal access point via the D channel or the B channel. Do.
[0038]
In any of the mounting methods, it is possible to accumulate / statistically process the results of past inquiries and make some judgment without inquiring of the center. In this case, it is more advantageous to implement the software on the user terminal in relation to the storage capacity and processing capacity.
[0039]
Further, based on only the line speed up to the destination, if the user does not need to frequently update the information, the routing information may be filed, downloaded to the user terminal or TA, and the determination may be made by local processing. .
[0040]
By efficiently combining these two types of processing with inquiries to the center, end users can enjoy the merits of this system without incurring unnecessary costs.
[0041]
Here, unless the message exchanged between the connection device and the center is very simple and short, it will adversely affect the response and the load on the center. In addition, when starting a service in a situation where the communication fee is not a flat-rate system, it is necessary to reduce the amount of information as much as possible. Furthermore, it is necessary to perform authentication by issuing a number in order to prevent the injustice that it cannot be determined even if a plurality of devices are set in one account (this information is unnecessary because it is described in the LAPD frame).
[0042]
Therefore, the message items exchanged between the connection device and the center should be at least
(Request) Connection device → Center
User ID, password, sequence ID: terminal unique, request type: normal / retransmission, etc., URL / IP address type, data (URL / IP address)
(Response) Center → Connection device
Request sequence ID, answer type: normal / indeterminable / other ..., answer content: provider ID (number) registered by user
Can be considered necessary.
[0043]
In order to establish with the above data, the user must register at least the provider name (coded) and the ID (about 4 digits) for mapping within his / her TA / router when joining the service It is necessary to register the center number, the provider ID and the telephone number of the access point corresponding to the provider ID in the terminal / TA / router.
[0044]
In the above, it is assumed that all the providers used by the user are registered (encoded), but it is almost impossible for the center to actually investigate the information for thousands and tens of thousands of providers . Further, when it is desired to select from a plurality of access points of one provider, it cannot be determined. In addition, if there is a means such as access via a company, it is impossible to make those comparison targets with the above interface.
[0045]
So, the response code
Answer: Provider registered by user + access point ID (number)
From
Answer content: provider registered by user + access point ID (number) + (expected throughput (kbit / s) x access line usage rate)
By changing to TA, the TA side will respond via the company if a response of 10 kbit / s or less comes in, and the provider that is not supported by the center but the contracted provider is about this speed, etc. The system can be made available.
[0046]
FIG. 6 shows an example of an embodiment of a routing information providing center, here an example corresponding to a service for Internet connection. In the figure, 61 is a response investigation agent server, 62 is a DB (database) / engine server, 63 Is a communication server, 64 is a user management server, and 65 is a management terminal.
[0047]
The response investigation agent server 61 periodically investigates the address of each access point, server, domain, node, etc. on the Internet from a management device on the Internet, and also accesses each access point, server, domain, node, etc. on the Internet. It has a function to periodically check the status of congestion (traffic) between devices on the Internet. In this case, it is important to adjust how often and how many layers are to be investigated in consideration of resource and load conditions.
[0048]
The DB (database) / engine server 62 has a route selection engine (main engine) together with the traffic information / network configuration information DB, the user management DB, and the billing / settlement information DB, and responds to inquiries from the communication server.
[0049]
The communication server 63 receives the investigation (inquiry) request from the user terminal using the D-channel packet, relays the request to the DB server, and sends a response from the relay server to the user terminal. The user management server 64 needs to accept registration (online sign-up) and various entry changes on the Web, and therefore accepts and processes them. The changes are reflected on the DB server.
[0050]
The management terminal 65 performs manual user registration (management), various kinds of statistical processing, creation of billing / settlement information, and the like (functions such as maintenance server monitoring / service control may be implemented here).
[0051]
FIG. 7 shows an outline of the operation of the routing information providing center in the system of the present invention.
[0052]
(1) When the communication server 63 receives a survey (inquiry) request from the user terminal and sends request information such as a destination address and a URL and a caller number to the DB / engine server 62, the DB / engine server 62 sends the request to the user management DB. Authenticate using the stored user information.
[0053]
(2) The DB / engine server 62 predicts a route and a throughput based on the information stored in the traffic information / network configuration information DB, selects an optimal provider (route), and makes the user terminal respond via the communication server 63. .
[0054]
(3) Further, the DB / engine server 62 performs a charging process for the corresponding user based on the charging information stored in the charging / settlement information DB.
[0055]
(4) The response investigation agent server 61 receives an address and a congestion state investigation request including a predetermined schedule and an investigation target from the DB / engine server 62, investigates on the Internet, and reports the result to the traffic of the DB / engine server 62. This is reflected in the information / network configuration information DB.
[0056]
(5) The user management server 64 processes information inquiries such as user authentication and usage status / charge based on the information stored in the user management DB and the charging / settlement information DB of the DB / engine server 62, and registers the user. The result of the reception processing such as / change is reflected in the user management DB.
[0057]
FIG. 8 shows another example of the embodiment of the routing information providing center, here an example corresponding to a single provider / enterprise. In the figure, reference numeral 61 'denotes a response investigation agent server, which periodically investigates the address of each access point, server, node, etc. on the network from a management device on the network. It has a function to periodically check the status of congestion (traffic) between nodes and the like from devices on the network.
[0058]
Reference numeral 62 'denotes a DB (database) / engine server, which is substantially the same as the DB (database) / engine server 62 in FIG. 6 except that it does not have a charging / settlement information DB. A management terminal 65 'is almost the same as the management terminal 65 of FIG. 6 except that charging / settlement information is not created. The communication server 63 and the user management server 64 are almost the same as those in FIG.
[0059]
The operation of this center is almost the same as that described with reference to FIG. 7 except that there is no billing process.
[0060]
When applied to remote access within a single provider or a company, etc., the line usage status of the access server and the like may be obtained from a Simple Network Management Protocol (SNMP) (Message Information Base). There is a limit to the method of measuring the congestion status on the Internet side by the ping method.
[0061]
Therefore, by utilizing the fact that a router on the Internet is an SNMP agent, it is possible to acquire the line type / speed and traffic of the interface.
[0062]
This information and the routing information (route setting) can also be known from the contracted provider by performing Trace Route to a certain destination, so that the center side uses "Internet map with line thickness" as static information. It is possible to create In other words, by acquiring all the necessary information of the SNMP, information other than the Round Trip becomes unnecessary in the Trace Route, and it is also possible to acquire and derive the route information other than the route that passes through the Trace Route.
[0063]
In contrast to the above, more accurate information can be obtained by periodically measuring and adding congestion or the like by a temporal parameter as a weighting element. According to this method, the load on the processing and the network can be reduced as compared with the method of periodically patroling all the domains.
[0064]
Of course, the Internet is also constantly changing routes and lines, so once you survey it, it will not end. Assuming that all domains are 10 million, it takes more than three months to execute Trace Route by ICMP once a second. For this reason, the efficiency can be improved by setting a ranking based on the frequency of inquiries and providing a mechanism for preferentially investigating places where inquiries frequently come.
[0065]
For example, if there is at least one inquiry, it will be reviewed within one month. If more than 10 in a week, it will be reviewed within one week. Schedule re-survey etc. within time.
[0066]
In addition, as a result of the investigation, it is also an issue how to evaluate a place where a response does not return (excluding packet loss due to congestion / collision).
[0067]
The point here is how to efficiently store the "Internet map" in the DB and use it (numerical representation).
[0068]
Using the destination address, a route to be used is determined from the routing information system table, and the logical bottleneck speed is obtained from the node map system table.
[0069]
This is calculated for the provider registered by the user, and the response time and line usage rate, and if the node is an access server, the parameters are added with emphasis on the line busy rate. Is calculated.
[0070]
The means for acquiring traffic on the center side and predicting speed information will be described in more detail.
[0071]
Conventionally, it is assumed that devices for measuring traffic and routes are installed in pairs at the destination and the source, or devices that monitor as agents are installed on the routes between the source and the destination. In order to respond to inquiries from users at different entrances, various types of information are acquired from centers that are not on the route between the user and the destination, and the route from the user to the destination and the congestion state are predicted based on the information.
1. A get is performed on a router on the Internet, which is an SMNP agent, to obtain the number of interfaces, interface types, line speeds, a list of routing tables, and the like.
2. Obtain a domain list from a domain management organization such as InterNIC or JPNIC. It is public and is automatically available via anonymous FTP. This is obtained, and the difference from the previous day is extracted.
3. From the center to each domain with trace route, which is a known means for the server of each newly added domain (for example, generally there is usually www.????.co.jp etc.) At the same time, the Round Trip is measured.
4. When these are combined, the following database is created, and from the request information from the terminal, the route and the bottleneck line portion, the estimated delay and the congestion situation are derived.
[0072]
[Table 1]

[0073]
The actual Internet is composed of a complicated network as described above, and a route to be passed is determined by a routing table set in each router. Therefore, the destination from the terminal does not always pass through the route of RouterA → RouterC → RouterE. Further, even if the route from the terminal to the destination is uniquely determined, it is not usually possible to measure the traffic or the congestion status from a device not on the route.
[0074]
Therefore, the route from the center to each domain (destination) is acquired using trace route, and the address of the network device (Router) between them is acquired.
[0075]
(Example)
Destination 1 ... RouterF (10.10.26.1) → RouterB (10.10.21.1) → RouterD (10.10.23.1) → RouterE (10.10.22.4.1) → Destination 1 (10.20.1.1)
Destination 2 ... RouterF (10.10.26.1)-> RouterB (10.10.21.1)-> RouterD (10.10.23.1)-> Destination 2 (10.0.2.2.1)
Also, a route from the center to the provider's access device can be obtained as well.
[0076]
(Example) Access server ... RouterF (10.10.26.1) → RouterA (10.10.20.1) → Access server (10.10.10.1)
At the same time, the delay time between the routers can be obtained. The information is organized as follows (/ = none, * = unknown,-= no connection).
[0077]
[Table 2]

[0078]
From the above results, the unknown parts are RouterA ← → RouterB and RouterB ← → RouterC.
[0079]
In this situation, blank areas may occur, but it is possible to make the same destination go through different routes by contracting with the access point of multiple different providers etc. at the center side As a result, most of the information can be filled with the obtained information.
[0080]
When there are two routes for accessing the destination 1, the access server to be accessed by the terminal is selected as follows.
[0081]
Route 1: Access Server 1 → Router A → Router C → Router E → Destination 1
Route 2: Access Server 2 → Router B → Router D → Router E → Destination 1
From the above-mentioned inter-router delay time, the route 1 is 5 ms + 5 ms + 5 ms = 15 ms, and the route 2 is 5 ms + 10 ms + 35 ms = 50 ms. Thus, the center can notify the user terminal of the connection to the access server 1 because the access can be made earlier by selecting the route 1.
[0082]
The configuration of the DB required on the center side is mainly as follows. The physical arrangement may be distributed in consideration of the load and the like.
・ Traffic information cache table
(The weighted traffic information retrieved according to the request is cached for a certain period of time.)
URL, IP address, throughput, provider serial number, access point number, creation date and time, survival time, response time, logical bottleneck speed, access count
・ Traffic information response table
(Measures and accumulates the response from the agent server to a certain node.)
Provider serial number, access point number, destination node serial number, response time, measurement date and time, number of references
・ Node map management table
Node serial number, node name, number of interfaces, node type
・ Node map line information table
Node serial number, interface serial number, interface IP address, line speed, line utilization
・ Routing information management table
Node sequence number, number of destinations, type of routing protocol (Static, BGP, RIP, IGRP, ... composite)
・ Routing information table
Node serial number, Destination address, Cost / Metric type, Cost / Metric value
・ Provider information table
Provider ID, provider name, domain name, number of access points, correction factor
・ Provider access point table
Provider serial number, access point number, name, telephone number, access server node number, correction factor
・ User information table
User ID, user status, and other user information (address, name, payment information, etc.)
・ User billing table
User ID, target date, number of uses
・ User payment table
User ID, target year and month, number of uses, amount, correction amount, settlement status
・ User registration information table
User ID, (Target provider ID, Target access point number) × n
(If the number of target connections is not limited, the number of tables will be multiple.)
・ User authentication table
User ID, registered phone number, password
・ Charge management table
Basic fee, inquiry unit price
・ Monthly charge management table
Amendment amount
In addition, there are a statistics management table for totalizing the inquiry status for each provider and each destination, a system management table having system / service status, an agent control table for scheduling agent servers, and the like.
[0083]
【The invention's effect】
As described above, according to the present invention,
・ Improve the throughput seen from the end user,
-If the end user notices that it is slow after connecting and changes the connection destination, it costs at least 10 yen, but with this mechanism, packet fee + service fee = several yen / time cost You can select a connection destination that speeds up reliably,
・ Distributes the load of the entire Internet,
・ Create an area where small-scale providers can be active,
・ A routing information service center can be established and commercialized.
・ By providing a contract form in which multiple providers are bundled, users can make efficient connections with one contract, and providers lose due to efficient use of lines and equipment and overflow. Customers can get connection fees,
・ By allowing users who have given up during busy hours to be able to connect during busy hours etc., we can expect an increase in traffic revenue.
And so on.
[0084]
In addition to multiple providers, it can be applied to a single provider or remote access in a company if the center returns a response according to the busy state of the access line and traffic, especially large-scale providers Thus, planned and efficient capital investment and operation become possible.
[0085]
Further, the present invention is applicable not only to ISDN D-channel packets but also to B-channel packets and services that can be multiplexed such as ATM and FR.
[Brief description of the drawings]
FIG. 1 is a schematic diagram of the entire system of the present invention.
FIG. 2 is a sequence diagram showing an outline of the operation of the terminal connection device in the system of the present invention.
FIG. 3 is a configuration diagram illustrating an example of an embodiment of a terminal connection device;
FIG. 4 is a configuration diagram showing another example of the embodiment of the terminal connection device;
FIG. 5 is a configuration diagram showing still another example of the embodiment of the terminal connection device;
FIG. 6 is a configuration diagram illustrating an example of an embodiment of a routing information providing center.
FIG. 7 is a sequence diagram showing an outline of the operation of the routing information providing center in the system of the present invention.
FIG. 8 is a configuration diagram showing another example of the embodiment of the routing information providing center.
[Explanation of symbols]
1: Internet, 2, 20: user terminal, 3: terminal connection device, 4: routing information providing center, 21: application, 22: socket, 23: network protocol, 24: network driver, 25: hardware driver, 26: EPR socket, 27: NAT function unit, 28: EPR control driver, 30, 40: terminal adapter, 31, 41: existing function unit, 32, 42: setting information DB, 43: terminal control function unit, 44: NAT function , 45: PPP / IP-serial converter, 46: EPR control function, 50: router, 51: LAN side interface (I / F), 52: routing function, 53: TA function, 54: NAT function Unit, 55: setting information DB (database), 56: EPR control function unit, 61, 61 ' : Response investigation agent server, 62, 62 ': DB (database) / engine server, 63: communication server, 64: user management server, 65, 65': management terminal.

Claims (5)

An access point selection and connection system for selecting one of a plurality of access points available to a user for a certain network and connecting a user terminal to the network, comprising: a routing information providing center; and a terminal connection device. An access point selection / connection system for performing communication between the routing information providing center and a terminal connection device using a logically multiplexable line such as an ISDN line D-channel packet service;
The terminal connection device is
Means for receiving a connection request including the destination address from the user terminal, sending the request to the routing information providing center as an inquiry request including the address of the communication partner from the user, and returning the address of the temporary access point to the user terminal;
Means for receiving a response of the address of the selected access point from the routing information providing center, and connecting the user terminal to the network via the selected access point;
An access point selection / connection system , comprising means for converting an address of a connected access point and an address of the temporary access point . When the line from the user terminal to the selected access point cannot be connected due to busy or the like, the inquiry request is sent again to the routing information providing center, a response is received, and this is repeated until connection is possible. The access point selection / connection system according to claim 1. The terminal connection device is realized by software or hardware mounted on a user terminal, or by software or hardware mounted on a terminal adapter or a router connected to the user terminal. The access point selection connection system described. The plurality of access points available to the user means at least one access point at each of two or more Internet service providers to which the user has subscribed, or a plurality of access points at one Internet service provider to which the user has subscribed. The access point selection / connection system according to any one of claims 1 to 3, wherein A storage medium storing a program of an access point selection / connection system for selecting one of a plurality of access points available to a user for a certain network and connecting a user terminal to the network,
When the program is read by a computer,
Means for receiving a connection request including the destination address from the user terminal, sending the request to the routing information providing center as an inquiry request including the address of the communication partner from the user, and returning the address of the temporary access point to the user terminal;
Means for receiving a response of the address of the selected access point from the routing information providing center, and connecting the user terminal to the network via the selected access point;
A storage medium storing a program for an access point selection / connection system, wherein the terminal connection apparatus includes means for converting the address of the connected access point and the address of the temporary access point.

Images