JP3401478B2 - Location selection device, location selection system, and recording medium - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data network, and more particularly to a device for selecting an optimum server installation site for a user in a short time.

[0002]

2. Description of the Related Art (Necessity of Base Selection Means and Required Conditions) As one means of providing information services on the Internet, there is a method of providing services by a plurality of distributed servers. As an example, it is possible to install a plurality of WEB servers to provide an information distribution service by HTTP. As described above, when a plurality of servers are installed to provide a service, it is necessary to have a mechanism for deciding which of the server groups the user should access when the user makes a service request. Especially when the service to be provided includes a large amount of information transfer such as file distribution and video streaming, it is desirable to select a server installed at a base closer to the user terminal. Further, in order to prevent the user from feeling the deterioration of convenience due to the distribution of servers, it is desirable that the time required for server determination is short. In other words, in order to solve the problem,
It is necessary to have a means for selecting a server at a location near the user terminal in a short time. In order to realize this means, the information for identifying the base accommodating the user terminal is input, and the base where the server capable of providing the service to be provided is installed is set as the candidate base. A device that selects a base close to the user is required. Such a device is called a site selection device.

In the technical field of networks, terms such as "distance" and "length" are used as a measure of perspective between two points on the network. The definition of this measure is a physical length based on the metric system. Not necessarily. For example, it is often defined by the round trip time of information between two points and the number of relay devices on the route between two points. The base means a space in which devices connected to the network are installed. In this specification, a base accommodating a user terminal is called a terminal base. The site selection device receives the information for identifying the terminal site. A specific example of this information is the network address of the user terminal.

(Description of Prior Art) For a route having a certain point on the network as a starting point and another point as an ending point, the identification information of each relay device on the route and the distance from the starting point of the relay device. Information indicating and is defined as route information of the route. A specific example of the identification information of the relay device is the network address of the relay device. There is software called traceroute as software that acquires route information between two points on the Internet. The traceroute receives a certain destination address as an input, and acquires the network addresses of the relay devices on the route from the traceroute execution terminal to the destination address in order from the execution device closest to the relay device. In the route information acquired by the traceroute, the distance is defined based on the number of relay devices on the route between two points (when the end point is a relay device, the relay devices are also counted). That is, the relay device identified by the m-th acquired network address is considered to be separated from the execution terminal by the distance m.

Guyton and Schwartz have published a paper "Loca on how to implement a site selection device using this traceroute."
ting Nearby Copies of Replicated Internet Server
s "(announced at ACM SIGCOM M'95), so it will be described below. For the sake of explanation, assume the network shown in FIG. 9. In FIG. 9, devices beginning with" R "represent relay devices," A device starting with "S" represents a server, a device starting with "D" represents a site selection device, a device starting with "U" represents a user terminal, and a relay device has a maximum of four connection terminals. , The number of each connection terminal is expressed as “0”, “1”, “2”, “3” in the clockwise direction from the 12 o'clock position. The base where the server that can provide the service to be provided to is installed is set as the candidate base, and the base closest to the input user is selected from the candidate bases.

In this specification, the route information of the route from the site selection device to a site p is represented as R (p). Further, the identification information of the relay device separated from the site selection device on the route by the distance m is represented as r (p, m). The symbol U represents information for identifying the terminal location. A symbol Y represents a list of identification information of candidate bases. The element of Y, that is, the identification information of a certain candidate base is represented as y (that is, yεY). Also, the site selection information, which is the result of site selection, is represented by the symbol Z. The element of Z, that is, the identification information of a selected base is z (that is, z ∈
Z).

The proposed site selection device is configured as shown in FIG. The figure shows an operation example when there are four candidate bases A, B, G, and H, that is, when Y = {A, B, G, H}. The traceroute executing unit 41 acquires the route information of the route from the site selection device 40 to the terminal site, that is, R (U) by executing the traceroute. The route information table 42 stores route information regarding each route from the site selection device 40 to each candidate site. The contents of the route information table 42 are as shown in Table 1. In Table 1, "NONE"
Indicates that there is no relay device on the route.

[0008]

[Table 1]

It is assumed that the route information to each base to be held in this table is acquired by using traceroute. The means 43 for obtaining the evaluation value of each candidate base acquires route information R (y) for the candidate base y, which refers to the route information table 42, and acquires this route R (y) and the route R (U to the terminal base. ) Is calculated to obtain an evaluation value D (U, y) of the distance between the candidate base y and the terminal base U. The output means 44 specifies the long / short relationship between the candidate bases regarding the distance to the terminal base based on the magnitude relationship of the evaluation values between the candidate bases, and preferentially selects the candidate base with the short distance (Z). Is output. A flow chart showing the flow of this operation is shown in FIG.

As the definition of the evaluation value, the length of the common part of the route from the site selecting device to the terminal site and the route from the site selecting device to the candidate site is doubled and subtracted from the sum of the two routes. Different values are possible. The technical basis of this definition is shown in Figure 1.
2 is explained. FIG. 12 shows two routes from a certain observation point (0) to a base # 1 and a base # 2. Let n be the length of the common part between the two paths. Let L1 be the length of the part peculiar to the route to the base # 1, and L2 be the length of the part peculiar to the route to the base # 2. The maximum value of the length of the route between the base # 1 and the base # 2 is the sum (L1 + L2) of the lengths of the parts peculiar to each of the observed routes. This is because a route connecting the bases # 1 and # 2 is generated if a route is formed by connecting the unique parts of the route at the branch point of the route. The sum of the lengths of the unique portions of the route is equal to the sum of the two routes minus twice the length of the common portion of the routes (denoted by n).

For example, in the network shown in FIG. 9, the servers that provide a certain service are Sa, Sb, and S.
In the case of g and Sh, the operation of the site selection device D1 selecting a server site for the user terminal U is as follows. First, the traceroute executing means 41 uses {Ra0, Rb0, Rd1, Rh as a route to the user base.
1, Ri2, Rk1, Rm2, Rn1, Ro1} are obtained. Next, the means 43 for obtaining the evaluation value of each candidate base calculates the evaluation value of each candidate base from this information and the route information table 42, and obtains the evaluation value as shown in Table 2.

[0012]

[Table 2]

The output means 44 determines that H is the closest to the terminal base and B is the next closest from the magnitude relation of the evaluation values, and sets the base selection information (Z) to {H, B}. In this example, the output means 44 operates to output the first priority and second priority sites.

In the following, four technical backgrounds that are important for explaining the characteristics of the prior art and the features of the present invention will be described.

The first point is about traceroute. First, the mechanism of traceroute will be described.

In the relay device of the IP network, the number of relay devices through which the packet to be transferred has passed is T in the header of the packet.
If the TL value is exceeded, the packet is not transferred and
Instead, it is provided with means for sending back an ICMP packet as an error notification to the transmission terminal of the packet. Also,
When the terminal receives a packet for a port number that is not waiting for reception, it has means for sending back an ICMP packet as an error notification to the transmitting terminal of the packet. The traceroute utilizes this means to acquire the value of the relay device on the route. The traceroute executing means 41 is configured as shown in FIG. The relay device information acquisition means 46 sets the TTL value in the IP header to a certain value and transmits the IP packet destined for the terminal (x) of the route acquisition target, and the ICMP caused by the packet transmission.
Receive the packet. The TTL value is the tracero
It is determined based on how many times the relay device information is acquired in the execution of ute, and when the relay device information is acquired for the kth time, the TTL value is set to k. When the TTL value of the transmitted packet is k, the source address of the received ICMP packet is the address of the k-th relay device counted from the traceroute executing terminal of the route, that is, r (x, k). Is. Whenever the relay device information acquisition unit 46 acquires the address of the relay device, the continuation determination unit 47 determines whether to continue the execution based on the address. Instruct to acquire relay device information. If the route length is unknown,
In the traceroute execution process, the packet arrives at the terminal for which the route is to be acquired, and the terminal tr sends the ICMP packet.
It is repeated until it is transmitted to the aceroute execution terminal. This state is shown in FIG. tracerout
When the number of relay devices on the route from the e execution terminal to the route acquisition target is N, the relay device information acquisition means 46
The relay device information acquisition is repeated N + 1 times with the TTL value set to 1 to N + 1.

A description will be given of the number of packet transfers that occur as a technical note of the traceroute. Route acquisition target terminal and traceroute
When there are N relay devices on the path between the execution terminals, t
The total number of packet transfers generated by executing the traceroute is (N + 1) (N + 2) in a round trip. If it is approximated that the relay transfer time and load are uniform regardless of the relay device and the relay line, the time required to execute the traceroute,
It can be said that the load is proportional to the square of the length of the target route.

The second point is that the Internet address does not reflect the hierarchical structure of the network.

The hierarchical structure of the network means, for example, a network in a certain area of Musashino City belongs to the network of Musashino City, and a network of Musashino City belongs to the network of Japan as a whole. An address that reflects the hierarchical structure of a network is an address that, when the address is divided into several parts, each part serves as an identifier for a network of a certain tier. For example,
The telephone number is an address that reflects the hierarchical structure of the network. For example, in the telephone number such as "+ 81.422.59.XXXX", "81" indicates "Japan", "422" indicates "Musashino City", and "59" indicates "Musashino City Area". The reason why the Internet addresses are not assigned such addresses is due to the technical characteristics of computer networks. That is, since the computer network has a characteristic that communication with a short duration frequently occurs, a packet including the address of the destination terminal in the header is used instead of connecting the line with the information transfer destination for each information transfer. Transfer information that should be transferred. It is preferable that the relay device uses the address in the header and has a short routing table. However, if the network address is assigned by reflecting the hierarchical structure of the network, it becomes difficult to fix and shorten the address length. This is because it is necessary to divide the address into the expected number of steps and secure a sufficient value space in each part after the division. Therefore, Internet addresses do not reflect the hierarchical structure of the network.

A technical problem caused by the fact that the address does not reflect the hierarchical structure of the network is the distribution of the database using the address as a reference key. For example, if it is a telephone number, it is easy to disperse a database that displays the location of the user from the telephone number. This is because, when a telephone number not registered in the database is given as a reference key, it is possible to construct a database that understands a part of the input telephone number and transfers the reference request to another database. However, it is difficult to distribute a database that uses Internet addresses as reference keys. This is because it is not possible to determine another database to which the request is to be transferred, from a part of the input address.

The third point is the evaluation scale of the base. Distance is not always the optimum evaluation scale. The network of FIG. 9 will be described as an example. The distance between two points is defined as the number of relay devices on the route between two points. When the base is selected by the distance, it means that the base #L is more suitable for the user than the base #H. However, suppose that the relay line of Rk and Ri, the relay line of Ri and Rh is 155 Mbps, and the relay line of Rk and Rl is 512.
If it is kbps, since the base #H can provide the service to the user faster than the base #L,
Selecting the base #H serves the purpose of selecting the base.

The fourth point is that the list of servers providing a certain information service may change with time.

On the Internet, the request frequency for a certain service suddenly increases or decreases. Therefore, it is not efficient to continue the service for a long period of time without changing the number and position of the distributed servers determined at the start of the service. Rather, it is desirable to increase or decrease the number of distributed servers according to the time variation of the request frequency.

[0024]

Five problems of the prior art will be described.

The first point is that since the traceroute is executed and then the route information table is referred to, even when the site can be selected only by the route information to the relay device in the middle of the route, the route to the user site is reached. This is the point where route information is acquired. For this reason, problems such as taking a long time to select a site and generating more network load occur. As a simple measure, it is conceivable to specify the maximum value of the number of hops (the number of relay devices to be acquired) as an option of traceroute. However, because of this measure, it is difficult to identify the number of hops required for site selection before executing traceroute.

The second point is that the size of the route information table to be held increases as the network becomes wider and the number of bases increases. This causes problems such as an increase in storage (memory) resources required for each site selection device and an increase in search time when referring to the route information table. It is desirable to be able to distribute the route information table so that the entire network is divided into several regions and the route information table held in each region is limited to the bases in this region. However, as mentioned as the second point of note in the prior art, it is difficult to determine the site selection device that holds the route information of the site optimal for this user from the Internet address of the user terminal.

The third point is that since only the route information observed from the site selection device is used, it is not possible to accurately grasp the length relationship between the plurality of sites distant from the site selection device and the user site. is there. Therefore, when the user base is far from the base selection device, a base that is not necessarily optimal may be selected. For example, in the above example, it is assumed that there is a server that can provide the service to be provided to the base #L. Since the evaluation value for the base #L is 7, the base #H is always regarded as the optimum base. However, if the number of relay devices on the route is used as the evaluation criterion, the base #L is actually closer to the terminal base. That is, the observation from the base #A cannot accurately grasp the long-short relationship between the distance from the terminal base to the base #H and the distance to the base #L.

In the above paper, as a remedy for this, a method using route information observed from a plurality of points is examined. However, since this method performs traceroute from a plurality of observation bases to the user base, it causes a problem such as a higher network load.

The fourth point is that since the base is selected only from the information in the route information table, factors other than the distance between the bases such as the band of the relay line between the bases and the load on the relay line are used as the evaluation scale of the base. Is not taken into consideration. Therefore, a result that is not necessarily optimal may be selected.

The fifth point is that it is necessary to obtain a list of server sites that can provide services to be provided before executing site selection. If the server list is static,
A list of servers may be set when the service is started. However, as mentioned as the fourth point of note in the prior art, it is desirable to change the server list according to the time variation of the service request frequency in order to efficiently provide the service. When the server list dynamically changes in this way, the site selection device periodically acquires the server list, or acquires the server list each time a service request is received from the user site. It is necessary to take some measures. When the former measure is taken, a problem is that a large network load occurs because the server list is calculated periodically for all the services to be provided. When the latter means is taken, there is a problem that the time required for site selection increases by the time required to acquire the server list.

The present invention solves the above-mentioned problems of the prior art, and compared with the prior art, the required time is shorter, the network load generated is smaller, the amount of storage resources required is smaller, and the server is smaller. It is an object of the present invention to provide a site selection device that can handle a case where a list is dynamically calculated. It is also an object of the present invention to provide a recording medium in which a site selection program that makes the computer device a device corresponding to the site selection device of the present invention is installed by being installed in the computer device.

[0032]

First, the fundamental points of interest leading to the present invention will be described. As described in the prior art, the evaluation value of the site selection device is calculated from the sum of the length of the route to the candidate site and the length of the route to the user site, which is 2 of the length of the common part of the two routes.
It is assumed that it is defined as the value obtained by subtracting double. Of these evaluation values, the length of the route to the user site is commonly added to the evaluation values of all sites. Since the superiority or inferiority of the evaluation value between the candidates is more significant than the evaluation value itself in the site selection, as an alternative evaluation value, "from the length of the route from the site selection device to the candidate site The value obtained by subtracting twice the length of the common part with the path can be used as the evaluation value. In order to obtain this evaluation value, it is not necessary to acquire all parts of the route information to the user base, and if the route information from the site selection device to the point where there is no common part with the routes of all the candidate bases is acquired. It is enough.

FIG. 1 shows the configuration of the site selection device according to the present invention. FIG. 1 also shows an operation example when the list of candidate bases is {A, B, G, H}. The evaluation value updating means 3 corresponds to a means for obtaining the evaluation value of each candidate base in the configuration according to the conventional technique. The relay device information acquisition means 1 corresponds to the relay device information acquisition means of the conventional traceroute execution means, and acquires the identification information of one relay device on the route from the site selection device 30 to the terminal site U. The identification information of the relay device is used as the identification information of the first relay device, and the distance of the identification information of the first relay device from the site selection device 30 is m. That is, the identification information of the first relay device is represented as r (U, m). The route information table 2 stores route information R (y) regarding a route from the site selection device 30 to the candidate site y for the candidate site y, as in the route information table in the configuration according to the related art. The evaluation value table 6 stores the evaluation value information for each candidate base. D is the evaluation value information of candidate base y
Expressed as (U, y). The evaluation value information includes an evaluation value of the distance between the candidate base and the terminal base, or a value to be held in the calculation process in the calculation process of this evaluation value.
Each time the relay device information acquisition unit 1 acquires the identification information r (U, m) of the first relay device, the evaluation value update unit 3 changes the evaluation value information of each candidate base in the route information table to the next value. Update as follows. That is, when updating the evaluation value information D (U, y) of the candidate site y, the identification information of the second relay device included in the route R (y) and separated from the site selection device 30 by m, that is, r (y, m) is acquired from the route information table 2, and the identification information r (U, U) of the first relay device is acquired.
m) and the identification information r (y, m) of the second relay device determine whether or not they indicate the same relay device, and the determination result, m, and the value of the evaluation value information D (U, y) are determined. The evaluation value information D is specified by specifying the length n of the common part of the route to the terminal base and the route to the candidate base y, or the range of values that n can take.
Update (U, y). The continuation determination unit 4 updates the evaluation value table 6 regarding the identification information r (U, m) of a certain first relay device each time the evaluation value updating unit 3 updates the evaluation value table 6.
By referring to the evaluation value table 6, it is judged whether the termination condition is satisfied, and if the termination condition is not satisfied, the relay device is acquired so as to acquire the identification information of the unacquired relay device on the route to the terminal base. The information acquisition means 1 is instructed (continuation instruction), and when the end condition is satisfied, the output means 5 is instructed to output. The output means 5 is an evaluation value table 6
Based on the magnitude relationship of the evaluation value of each candidate base indicated by, the long / short relationship between the candidate bases regarding the distance to the terminal base is specified, the candidate base is selected preferentially from the one with the shortest distance, and the base showing the selected base The selection information (Z) is output.

As a concrete definition of the evaluation value, the length L of the route from the site selecting device 30 to the candidate site is subtracted from twice the length (n) of the common part of the route and the route to the terminal site. Value, that is, L-2n. However, even if the output value of the monotonically increasing function or the monotonically decreasing function with this value as an input is used instead of this value, an equal effect can be obtained.

The end condition is defined so that whether or not the condition is satisfied can be judged based on the contents of the evaluation value table 6. A specific example of the termination condition is that the evaluation values of all the candidate bases have been set, in other words, that the common route length of all the candidate bases has been obtained. Another end condition is that the magnitude relationship of the evaluation values could be specified. Another termination condition is that a predetermined number of evaluation values less than or equal to a predetermined value have been obtained.

One form of the present invention will be described in more detail. The value of the evaluation value information at the time point before the relay device information acquisition unit 1 acquires the identification information of the relay device is set as the initial value of the evaluation value information. The evaluation value table 6 uses this initial value as a value that the evaluation value cannot take. For example, the evaluation value is L
-2n, the route information table 2
The maximum length of the route registered in is used, and a value larger than this maximum length is set as the initial value. Relay device information acquisition means 1
Acquires the identification information of the relay device one by one along the route to the terminal base from the one closest to the base selection device 30. The evaluation value updating means 3 uses the evaluation value information D of the candidate base y.
When updating (U, y), the identification information r (U, m) of the first relay device and the identification information r (y, m) of the second relay device indicate the same relay device. Is it D (U,
When y) is a value other than the initial value, the value of D (U, y) is not rewritten, and r (U, m) and r (y, m) indicate the same relay device and D (U, y). Is an initial value, the distance m ′ of the relay device identified by the relay device identification information previously acquired by the relay device information acquisition means 1 from the site selection device is set as the length n of the common route. An evaluation value is calculated, and this evaluation value is used as the value of D (U, y). One example of the definition of the end condition is that the evaluation value information of all candidate bases in the evaluation value table 6 is other than the initial value.

When the value of m is regularly increased according to the number of times the relay device information acquisition unit 1 acquires the identification information of the relay device according to the definition of the distance, m'can be obtained from m. it can. For example, if the distance is defined by the number of repeaters on the route between two points,
m'is m-1.

The first difference from the conventional method is that an evaluation value table 6 showing the evaluation values of the candidate sites or the data for calculating the evaluation values is added. The second difference is that an end condition that can be determined based on the contents of the evaluation value table 6 is defined, and each time the relay device information acquisition unit 1 acquires a relay device identifier, the evaluation value table 6 is updated and the end condition is satisfied. The change in configuration is such that the determination as to whether is satisfied is executed as a series of operations. As a result, the relay device identifiers for all the routes to the user terminal in the conventional method are not acquired, and the selection process is completed with a smaller number of times of relay device information acquisition than in the conventional method.

The same effect can be obtained for the device configuration in which the evaluation value information stored in the evaluation value table 6 includes a value to be held in the calculation process in the calculation process of the length n of the common path or n. To be In this configuration, the output unit 5 calculates the evaluation value of the candidate base by referring to the common path length n regarding the candidate base stored in the evaluation value table 6 and obtains the magnitude relationship of the evaluation values. .

Points of interest for further improvement will be described. In the process of updating the evaluation value information of the candidate base y, the identification information r (U, m) of the first relay device
And the identification information r (y, m) of the second relay device indicate the same relay device, it is possible to specify that the length n of the common part of the routes is m or more, while r (U, m) and r
When (y, m) indicates a relay device different from each other, the length of the common portion can be specified to be n−d or less, where d is the minimum value of the distance between the relay devices on the network. Using this, the minimum value of the range of evaluation values is defined as the minimum value, and the maximum value is defined as the maximum value, and these values are stored in the evaluation value table 6 as evaluation value information. If the range in which the evaluation value can be taken is narrowed while the value of is updated every time the relay device information acquisition unit 1 acquires the identification information of the relay device, it is not necessary to sequentially acquire the information of the relay device. For example, if the information of the relay devices separated by the average length of the candidate bases is acquired in the first execution process of the relay device information acquisition, the candidate bases can be narrowed down early.

The invention based on the above will be described. First, the evaluation value table 6 includes the maximum value and the minimum value as the evaluation value information. The maximum value of the candidate bases y stored in the evaluation value table 6 is represented by DH (U, y), and the minimum value is represented by DL (U, y). The value of the evaluation value information at the time point before the relay device information acquisition unit 1 acquires the identification information of the relay device is set as the initial value of the evaluation value information. The evaluation value table 6 sets the initial values of DH (U, y) and DL (U, y) by using the values that can be specified by the evaluation value at that time. When updating the evaluation value information of the candidate base y, the evaluation value updating means 3 has the first relay device r (U, m) and the second relay device r out of the maximum value and the minimum value.
The maximum value DH stored in the evaluation value table 6 is the maximum value that can be specified by calculating the identifiable one using the determination result of whether (y, m) are the same and the above-mentioned m.
If it is smaller than (U, y), DH (U, y) can be replaced with the calculated maximum value, and the minimum value can be specified, and the calculated minimum value can be stored in the evaluation value table.
If it is larger than (U, y), DL (U, y) is replaced with the calculated minimum value.

An invention belonging to one mode of the above invention will be described.
The distance between arbitrary relay devices on the network is set to the minimum value d. In the evaluation value table 6, the length of the route to the candidate base y is L, the initial value of the minimum value DL (U, y) of the candidate base y is -L, and the maximum value DH (U, y) is the initial value. Be set to L. Then, the evaluation value updating means 3 uses the first relay device r (U, m) and the second relay device r.
When (y, m) indicates the same relay device, it is determined that the maximum value can be specified, this maximum value is calculated as L-2m, and the first relay device r (U, m) and the second relay device r (U, m) Relay device r
When the relay devices have different (y, m), it is determined that the minimum value can be specified, and this minimum value is calculated as L-2 (m-d).

Even if the output values of the monotonically increasing function or the monotonically decreasing function having these values as inputs are used as the minimum and maximum values instead of the above-mentioned minimum and maximum values, a uniform effect can be obtained.

A specific example of the termination condition in the above embodiment is that the minimum value and the maximum value of each candidate base in the evaluation value table 6 match.

Further, as an invention to be improved, the evaluation value updating means 3 is means for searching the route information table 2 for a route for the identification information of the first relay device obtained by the relay device information obtaining means 1. When a route satisfying the condition is found as a result of the search, the identification information of the relay device for each of the relay devices on the section of the found route from the device to the relay device is the relay device information acquisition unit 1. It is assumed that the apparatus includes means for updating the evaluation value information, which is considered to be acquired by. As a result, when the relay device information acquisition unit 1 acquires information on a remote relay device, the evaluation value information is updated by acquiring the information on the nearby non-acquired relay device from the route information in the route information table 2. ,
It is possible to reduce the number of times relay device information is acquired.
This process is called route information interpolation.

Further, in order to obtain the relay device information in an efficient order, the following improvements are added to the invention. That is,
The relay device information acquisition means 1 inputs a certain distance k and acquires the identification information of the relay device located on the route from the observation base to the terminal base by a distance k from the observation base or in the vicinity of the point. The means to do so shall be included. Then, in the first acquisition of the identification information of the relay device, k is set to the average of the lengths of the routes in the route information table 2. In addition, in the second and subsequent acquisitions of the identification information of the relay device, if there is a route in the route information table 2 that includes the identification information of the relay device that was previously acquired, k is set to the average length of these routes. I will. As a result, it becomes possible to acquire the information of the relay device that is separated by a distance that allows the route information to be interpolated, and reduce the number of times of execution processing of acquiring the relay device information.

Further, as an improvement for solving the fourth problem of the prior art, the site selection device is constructed as shown in FIG. This device performs site selection in two stages: primary selection and secondary selection. As for the candidate base list Y, a candidate base list Y1 for primary selection and a candidate base list Y2 for secondary selection are separately defined. In order to improve the effect, it is preferable that the primary selection candidate base list Y1 includes not only the bases where the information providing devices are installed but also as many bases as possible on the network.
The secondary selection candidate base list Y2 includes candidate bases in the target base selection of the main base selection device, that is, the installation bases of the information providing devices. The primary selection means in this device is the site selection device described above.
The primary selection means 7 is the closest base (z
1) is selected from the primary selection candidate site list Y1 and site selection information Z1 (primary selection result) indicating the selected site is output. For example, in FIG. 2, Z1 = {K} is selected. The secondary selection means 8 uses the site z1 indicated by the primary selection result Z1 and the candidate site y2 for secondary selection (where y2εY).
The evaluation value between (2) and (evaluation value for secondary selection) is acquired by the evaluation value acquisition means 9 for secondary selection, and the base having a good evaluation value is given priority to the base from the candidate base list for secondary selection Y2. To output the site selection information Z indicating the selected site. The secondary selection evaluation value acquisition unit 9 acquires an evaluation value (secondary selection evaluation value) defined based on a scale different from that of the primary selection unit 7 as the secondary selection evaluation value between the two bases. The evaluation value for secondary selection between a certain base a and a certain base b is represented as D2 (a, b).

With this configuration, first, the primary selecting means 7 selects a site on the network closest to the user terminal. Next, the secondary selection evaluation value between this base and each information providing device installation base is acquired by the secondary selection evaluation value acquisition means 9. It is possible to realize more appropriate site selection by reflecting the bandwidth of the line between the sites and the load of the line in the definition of the evaluation value for secondary selection.

As a typical implementation method for implementing the secondary selection evaluation value acquisition means 9, a secondary selection evaluation value table storing secondary selection evaluation values between bases is provided, and the secondary selection is performed. A method is conceivable in which the use evaluation value acquisition means 9 refers to this table to acquire the evaluation value for secondary selection between sites.

When the entire device is regarded as one site selecting device, setting the candidate site list Y of this site selecting device to a certain list means setting the secondary selection candidate site list Y2 to that list. be equivalent to. The site selection information of this device means the information output by the secondary selection means 8.

Further, in order to solve the problems of the second and third points of the prior art, the following improvements are added. That is,
As shown in FIG. 3, the site selection system that executes the site selection process by the site selection device 30 and the activation device 15 is configured, and a communication unit is provided between the site selection process and the activation device 15. Specifically, the activation device 15 includes a request transmission means 12 for transmitting a site selection request message including information U for identifying a terminal site, and a response reception means for receiving a response message including the selected site or other identification information. 13, and the site selection device 3 includes a request receiving means 10 for receiving the site selection request message and the response transmitting means 11
And. The request receiving unit 10 uses the information U for identifying the terminal base indicated by the base selection request message as the input information. It is assumed that the response transmission means 11 generates the response message by including the identification information of the candidate base included in the base selection information output with respect to the input information. Furthermore, site selection device 3
0 is provided with a site selection device table 16 that stores identification information of other site selection devices in association with identification information of the candidate sites. Then, the response transmission means 11 refers to the site selection device table 16 and, when the identification information of the site selection device corresponding to the candidate site indicated by the site selection information is present in the table, returns the identification information as the response. It should be included in the message. Further starter 1
5 is the site selecting device 3 when the response message received by the response receiving means 13 includes the identification information of the site selecting device.
The re-request means 14 for transmitting a site selection request message to 0 will be provided.

FIG. 3 shows the case where Z is output as the site selection information for the input information U, and Z contains the identification information of the candidate site z. N (z) represents the identification information of another site selection device stored in association with the candidate site z in the site selection device table 16.

Conventionally, there is a problem that it is difficult to identify the device to be inquired from the input information even if the databases are distributed because the Internet address has no hierarchical structure. In the present invention, this problem is defined by defining the relationship between the candidate bases and other base selecting devices in the base selecting device table 16, and selecting bases corresponding to the candidate bases selected by the mechanism of the candidate base selecting by the base selecting device. This is overcome by acquiring the identification information of the device from the site selection device table 16. That is, instead of directly obtaining the site selection device of the inquiry destination from the input information, first the candidate site selected for the input information is obtained, and then the site selection device of the inquiry destination corresponding to the selected candidate site is obtained.
Other site selection devices are acquired from the input information in this order. In this method, since it is only necessary to define the correspondence between the candidate bases and other base selection devices, it is applicable even when the input information is not hierarchically configured. As a result, the entire network can be appropriately divided into several areas, a site selection device can be provided for each region, and one selection process can be executed by the cooperation operation of the site selection devices in these multiple regions. Becomes In the route information table 2 of the site selection device in each area, only the route related to the site in the region and the route related to a representative site in another region need to be registered. The route to each site in the other region is registered. No need to register. As a result, the size of the route information table 2 that each site selection device should hold can be reduced, and the memory resources required for the device can be reduced. Further, the present system executes the site selection while gradually transferring the site selection processing to the site selection device closer to the user terminal. Therefore, it is possible to reduce the probability that the perspective relationship between the distant bases is erroneously determined, which has been mentioned as the third problem of the conventional technique.

Further, in order to solve the fifth problem of the prior art, as shown in FIG. 4, a site selection system is constructed by the following means. That is, the site selection system includes the site selection system, and the following expansion is added to the activation device 15 and the site selection device 30 of the system. The request transmission means 20 of the activation device 15 has a means for including information (V) for identifying a service in the request message. The site selection device 30 includes a list S (v) of identifiers of servers that provide a service for information (v) that identifies a service, and a server information acquisition unit 18 that acquires an installation site of each of these servers. I will. The request receiving means 17 acquires a list of server identifiers (S (V)) from the service identification information (V) included in the site selection request message, and obtains a list of installation sites of these servers for the candidate sites. The list (Y) is used. Further, the response transmission means 19 selects the site selection information (Z) from the list S (V) of the server identifiers.
The identifier of the server installed at the site indicated by is included in the response message. In FIG. 4, S (V)
Element is shown in the format "Sa @ A",
The character string before the @ symbol indicates the identifier of the server, and the character string after the @ symbol indicates the identification information of the base where the server is installed.

According to the above invention, it becomes possible to execute the acquisition of the information regarding the server that provides the service in parallel with the selection of the base. Even if the server information providing the service is acquired for each site selection, the server information is acquired only in the area where the site selection device is activated, so the load is less than that of the conventional method. I'm sorry. Further, even when the table indicating the relationship between the service and the server providing the service is held, the content of this table can be distributed in each region, and thus the size of the table can be smaller than that of the conventional method. At the same time, the update load will be low.

Another aspect of the present invention is to install a computer device provided with predetermined hardware and predetermined basic software installed in the hardware, thereby selecting the computer device as the site of the present invention. It is a recording medium on which software is recorded as an apparatus or an apparatus corresponding to a site selection system.

[0057]

BEST MODE FOR CARRYING OUT THE INVENTION The site selection device and site selection system according to the present invention are realized by a computer and software activated on the computer. The route information table, the evaluation value table, the secondary selection evaluation value table, and the site selection device table are realized by a database using a computer memory and a hard disk. In the following description of the embodiments, each table is shown by a table, but in the implementation, it is possible to realize the table shown in the table by using a relational database.
Alternatively, instead of using a database, it is also possible to realize each row of the table as one row and store the file in which the contents of each column of each row of the table are separated by a space (or a comma symbol) in the hard disk. The request transmitting unit, the request receiving unit, the response transmitting unit, and the response receiving unit are realized by communication control software on a computer and a network interface mounted on the computer. Other means are implemented as software subroutines on a computer. The entire site selection device is realized by the main routine that starts these subroutines.

(First Embodiment) The site selecting apparatus of the first embodiment of the present invention will be described. FIG. 9 used in the description of the prior art
Network is assumed. Description will be given focusing on the site selection device D1 in the site A.

The site selection device in this embodiment is as shown in FIG. The site selection device is realized as a main routine program of a computer. This program
It is activated by using the information U for identifying the terminal base and the list Y of the candidate bases as arguments, and outputs the base selection information Z as the execution result. The information U for identifying the terminal base is the network address of the user terminal.

The route information table is shown in Table 1 as in the prior art. That is, each row is realized by a table showing route information to a certain base. In each row of the table, the first column shows the identification information of the base and the second column shows the route information. The route information is a character string listing the network addresses of the relay device interfaces. The length of the route is defined as the number of relay devices on the route. In this case, the minimum value of the distance between the relay devices is 1. In the following description, regarding the route information in the route information storage table, the route length (the number of registered relay device addresses) is L
The address of the i-th relay device on the route to a certain candidate base y is represented as r (y, i). As shown in Table 3, the evaluation value table holds evaluation value information of each candidate base.

[0061]

[Table 3]

In other words, each row is realized by a table showing evaluation value information of a certain base. In each row of the table, the first column shows the identification information of the base, and the second column shows the evaluation value information. The evaluation value information of a certain candidate base y is D (U,
y). An appropriate integer is selected on the condition that it is larger than the maximum length of the route in the route information table, and this value is set as the initial value of each candidate base.

The relay device information acquisition means is a subroutine program, is activated with the TTL values m and U as arguments, and returns the relay device address r (U, m) as a return value. The realization method of the relay device information acquisition means is a conventional tra
This is the same as the relay device information acquisition means of the certout execution means. That is, a packet whose destination is U is TTL
The value is set to m and transmitted, the ICMP packet resulting from this is received, and the source address of the received ICMP packet is set as the return value r (U, m).

Next, a method of realizing the evaluation value updating means will be described.
The evaluation value updating means uses the TTL value m and the relay device address r.
It is a subroutine program started with (U, m) as an argument. Let y be a candidate base and r (U, m) ≠ r
When (y, m), the candidate base y in the evaluation value table y
It is determined whether the evaluation value information D (U, y) of is an initial value. If it is the initial value, the evaluation value is determined by the formula (1), and the evaluation value is rewritten in the evaluation value table. D (U, y) ← L-2 * (m-1) (1)

Next, a method of realizing the continuation judging means will be described. The continuation judging means is a subroutine program and returns 1 as a return value when the end condition is satisfied and 0 when the end condition is not satisfied. A return value 0 means a continuation instruction to the relay device information acquisition means. The end condition is that the evaluation value is rewritten from the initial value in the evaluation value table.

The output means is a subroutine program, which refers to the evaluation value table to generate the site selection information,
The generated site selection information is returned as a return value. The output means selects a base having the lowest evaluated value in the evaluation value table and outputs information on the base. As another example of the output means, a mode in which a list of candidate bases rearranged in order from the lowest confirmed evaluation value is output is also conceivable.

The structure of the main routine program will be described. The main routine program is started with the information U for identifying the terminal base and the candidate base list Y as arguments. The main routine program creates an evaluation value table in which the element Y is a candidate base, and sets the value of the evaluation value information of each candidate base to the initial value. Further, the main routine program stores a variable m representing the TTL value. The initial value of m is 1. The main routine program activates the relay device information acquisition means using U and m as arguments, and relay device address r
Get (U, m). The main routine program that has acquired the relay device address r (U, m) from the relay device information acquisition means activates the evaluation value updating means with m and r (U, m) as arguments. When the execution of the evaluation value updating means is completed, the main routine program activates the continuation judging means. If the return value of the continuation determination means is 0, the main routine increments the value of m by 1 and restarts the relay device information acquisition means. If the return value of the continuation judging means is 1, the main routine activates the output means and outputs the site selection information obtained as the return value from the output means.

An example of execution of site selection when the sites A, B, G, H, and L are server installation sites, that is, candidate sites, and the user terminal belongs to the site U is shown. Table 4 shows the process of updating the evaluation value table.

[0069]

[Table 4]

As can be seen from Table 4, the evaluation values of all the candidate bases are fixed when the fifth acquisition of the relay device information is completed. Therefore, the estimation process is completed at this point.

Since the number of relay devices between the user terminal and the site selection device is nine, the tracerou
When te is executed, the total number of packet transfers that occur is 110. In the present invention, the total number of packet transfers is 3.
It only requires 0 times (twice the sum of natural numbers up to 5). Assuming that the packet transfer load and the transfer processing delay in the relay device are constant regardless of the relay device and the transfer delay on the relay line is constant regardless of the relay line, the time required for estimation according to the present invention, It can be considered that the network load can be reduced to 30% or less.

There is also a method of defining a monotonically increasing function or a monotonically decreasing function f (x) as shown below and rewriting the evaluation value by using the equation (2) instead of the equation (1). Example of monotone increasing function: f (x) = 2x + 1 Example of monotone decreasing function: f (x) =-2x + 1 D (U, y) ← f (L-2 * (m-1)) (2)

(Second Embodiment) Next, a second embodiment obtained by improving the first embodiment will be described. The evaluation value table in the second embodiment holds the minimum value and the maximum value as the evaluation value information. The minimum value in the evaluation value table of a certain candidate base y is represented as DL (U, y), and the maximum value is represented as DH (U, y). The initial values of the minimum value and the maximum value are set based on the length of the route to the candidate base in the route information table. When the path length is L, the minimum value is set to -L and the maximum value is set to L. The initial values of the maximum value DH (U, y) and the minimum value DL (U, y) in the evaluation value table are as shown in Table 5.

[0074]

[Table 5]

In the second embodiment, the relay device information acquisition means takes the TTL value m ′ and the relay device address r (U, m ′) at the time of the previous relay device information acquisition as arguments instead of m,
As a return value, in addition to the relay device address r (U, m), T
Returns the TL value m. The relay device information acquisition means searches the route information table for routes whose m'th relay device address is r (U, m '), and when these routes exist, the average length of these routes is an integer. The converted value is set as the TTL value m. When these routes do not exist, the TTL value is set to m'-1. However, in the first acquisition of the relay device information, it is assumed that the relay device information is activated with m ′ = 0. In this case, the relay device information acquisition means sets the TTL value m as a value obtained by converting the average value of the lengths of the routes in the route information table into an integer. However, the TTL value used in the past is not used again, and for this reason, the relay device information acquisition means stores the TTL value used in the past. The method of acquiring the relay device information is the same as in the first embodiment. That is, the destination address is U,
A packet in which the TTL value is m is transmitted, the ICMP packet is received, and the relay device address r (U, m) is obtained.

Next, a method of realizing the evaluation value updating means will be described.
Similar to the first embodiment, the evaluation value updating means is activated with the TTL value m and the relay device address r (U, m) as arguments.
Let r (y, m) be the address of the relay device corresponding to the TTL value m of the route to the candidate site y in the route table, and set the minimum value DL (U, y) and maximum value DH (U, U) of the candidate site y.
y) is updated by the following method. r (U, m) = r
If it is (y, m), the maximum value is updated by the equation (1). If r (U, m) = r (y, m) is not satisfied,
The minimum value is updated by the equation (2). Where Min
(A, B) indicates the smaller value of A and B, and Ma
It is assumed that the larger value of x (A, B) is shown. Maximum value DH when r (U, m) = r (y, m)
Updating method of (U, y): DH (U, y) ← Min (L-2 * m, DH (U, y)) (3) When r (U, m) = r (y, m) is not satisfied Minimum value DL
Updating method of (U, y): DL (U, y) ← Max (L-2 * (m-1), DL (U, y)) (4)

Next, a method of realizing the continuation judging means will be described.
The continuation judging means is a subroutine program, and its return value is the same as that of the first embodiment. The end condition is that the minimum value and the maximum value match for each candidate base in the evaluation value table. This condition is the same as that there is no candidate base a with DH (U, a) ≠ DL (U, a) in the evaluation value table. As another end condition, it is possible to realize that the magnitude relationship of the evaluation values is fixed. This condition is DL (U, a) in the evaluation value table.
This is the same as the fact that there is no combination of candidate bases a and b that is> DH (U, b). Further, as another termination condition, it is possible to realize that an evaluation value equal to or less than a predetermined value (DT) is found more than a predetermined number (NT). This condition is a candidate base a that satisfies DL (U, a) ≦ DT in the evaluation value table.
Is equal to or more than NT.

Further, a method for adding a means for compensating for an unacquired relay device address by utilizing the route held in the route information table to the evaluation value updating means will be described. It is assumed that the relay device information acquisition means executes the acquisition process with the TTL value set to m and acquires the relay device address r (U, m). The evaluation value updating means searches for a route in the route information table for which the relay device address corresponding to the TTL value m matches r (U, m). When there is a matched route, the route is regarded as T, and r (U, 1) to r (U, m) are considered to be equal to r (T, 1) to r (T, m). Then, if there is a relay device address in ((U, 1) to r (U, m) that has not been acquired yet, it is based on this relay device address as in the case where the relay device information acquisition means acquires it. Update the evaluation value information.

The main routine program in the second embodiment differs from the main routine program in the first embodiment in the method of activating the relay device information acquisition means. In the first activation of the relay device information acquisition means, m ′ = 0, r (U, m ′)
Set the argument as = 0. In the second and subsequent activations of the relay device information acquisition means, m and r (U, m) obtained as return values for the previous activation are set in the arguments m ′ and r (U, m ′).

For example, in the same example as the first embodiment, the initial value of TTL is 3 from the average length of routes in the route table. Rd1 is obtained by the first acquisition of relay device information with the TTL value set to 3. The evaluation value updating means refers to the route information table and searches for a route including Rd1 as the relay device address corresponding to the TTL value of 3, thereby Ra is set as the first three elements of the route to the user terminal.
0, Rb0, Rd1 are obtained. And the evaluation value updating means,
Update when the TTL value is 1 and the relay device address Ra0 is acquired, update when the TTL value is 2 and the relay device address Rb0 is acquired, and when the TTL value is 3 and the relay device address Rd1 is acquired And update. In the route table, the average length of the routes including = Rd1 as the relay device address corresponding to the TTL value of 3 is 5. Therefore, the TTL value is set to 5, and the relay device information is acquired for the second time. Ri2 is obtained by the second acquisition result, but TTL
Since there is no route including Ri2 as the relay device address corresponding to the value 5 in the route information table, the relay device address is not interpolated. The evaluation value updating means executes only the updating when the TTL value is 5 and the relay device address Ri2 is acquired. For the next relay device information acquisition, the relay device address acquired last time was not in the route information table.
Becomes 4. As a result of this acquisition, Rh1 is obtained. The evaluation value table is updated as shown in Table 6 in each acquisition of the relay device information, and the site selection process is completed by the acquisition of the relay device information three times.

[0081]

[Table 6]

The total number of packet transfers that occur in the second embodiment is 24 ((3 + 5 + 4) × 2). In other words, the total number of packet transfers is further reduced by 20% from the first embodiment by the improvement shown in the second embodiment.

As an example of further improvement,
It is conceivable that the temporary storage table that holds the route acquired by the relay device information acquisition is provided, and that the information held in this table is used for the route interpolation performed when the evaluation value is updated in the next estimation process. For example, in the above example, Ra
The routes 0, Rb0, Rd1, Rh1, Ri2 are acquired. If you store this in a temporary storage table, Ri
The time of the estimation process for the user terminal on the route passing through 2 is shortened from the next time.

In addition, in the second embodiment described above, f
A monotonically increasing function or f (x) with (x) = 2x + 1
A method using a monotonically decreasing function of = -2x + 1 can be considered. When using a monotonically increasing function, the evaluation value
Table is the maximum of the evaluation values in the evaluation value table.
The initial value of the large value is set to the value of f (L), and the evaluation value
The initial value of the minimum value is set to the value of f (-L). Also, simply
When using the tone decrease function, the evaluation value table
Is the maximum value of the evaluation values in the evaluation value table.
The initial value is set to f (-L) and the minimum value of the evaluation values is set.
The initial value of is set to the value of f (L).

(Third Embodiment) Further, a plurality of site selection devices are distributedly arranged on the network, and the server providing the service is input by inputting the Internet address of the user terminal and the identifier of the service to be provided to the user. An implementation example of a site selection system that outputs addresses and priorities will be described. This embodiment solves the problems of the second, third, and fifth points of the conventional technology.

As shown in FIG. 5, the entire network is divided into three areas, and a site selecting device and a starting device are installed in each area.

The request transmission means of the activation device requests the site selection in which the internet address or internet host name (U) of the user terminal and the URL (V) indicating the service to be provided to the user are entered in the format shown in FIG. The message is transmitted to the site selection device by the TCP / IP protocol. First, it is assumed that a site selection device that belongs to the same area as the activation device is registered in advance as a site selection device that transmits a site selection request message. As another implementation example, a method of providing the activation device with means for acquiring the address of the site selection device by a network protocol using a broadcast message or the like can be considered. The response receiving means of the activation device receives the response message using the TCP / IP connection used for transmitting the site selection request message. FIG. 7 shows a format example of the response message. First, a status code that can take the values of "OK" and "REDIRECT" is entered at the beginning of the response message. When the status code is "OK", the number of candidate servers is shown after that, and the URLs of the servers are displayed in order from the number of the indicated candidate servers in order of priority. When the status code is "REDIRECT", the internet address of the site selection device in another area is shown. As another implementation example, even when the status code is “REDIRECT”, a method of showing the candidate server list as the second priority candidate in addition to the site selection device in another area is also conceivable. When the status code of the response message is REDIRECT, the re-request means transmits the site selection request message again to the site selection device in another area indicated by the response message.

The activation device is realized by activating a program that causes a computer equipped with the network Internet to perform the above operation. The request receiving means of the site selection device receives the site selection request message by the TCP / IP protocol.

The route information table of the site selection device regards each of the other areas as one site, and holds the route to the area and the address of the site selection device installed in the area as the site information about the area. That is, in this embodiment, the route information table also serves as the site selection device table. Table 7 is a route information table of the site selection device installed in area 1.

[0090]

[Table 7]

Table 8 is a route information table of the site selection device installed in area 2.

[0092]

[Table 8]

The server information acquisition means is activated with the URL (V) indicating the service to be provided as an argument, and returns the URL of the mirror server providing the service and the list of installation bases (S (V)) as a return value. It is a subroutine program. As shown in Tables 9 and 10, the server information acquisition unit holds a server information table in which a URL indicating a service, a URL of a mirror server that provides the service, and an installation base are associated with each other. Table 9 shows the server information table of area 1, and table 10 shows the server information table of area 2.

[0094]

[Table 9]

[0095]

[Table 10]

The URL indicating the service that is an argument
The server information corresponding to is read from the server information table, and the URL of the mirror server and the list of installation bases are returned as return values. As another implementation example of the server information acquisition means, a method of implementing message exchange using multicast between a site selection device and a server in an area, and the site selection device acquires information of a mirror server from a URL indicating a service Can also be considered.

The response transmitting means of the site selection device transmits the response message using the TCP / IP connection used to receive the site selection request message. This means
URL of mirror server and list of installation bases (S (V))
And the site selection information (Z) are used as arguments. Then, the route information table is referred to for the candidate site included in the site selection information Z, and if the address of another site selecting device corresponding to the candidate site is stored, the address of this site selecting device is acquired. To do. If the address of another site selection device is stored, the status code of the response message is "REDIRECT", and if not, the status code is "OK". Status code is REDIRECT
When it becomes T, the address of the acquired other site selection device is entered in the response message. If the status code is OK, the UR of the mirror server accepted as an argument
Of L, the URL of the mirror server installed in the candidate site included in the site selection information is entered in the response message. Then, the generated response message is transmitted.

The site selection device is the main program of the computer. Unlike the first and second embodiments, in the third embodiment, the site selection device is always operating and TCP
/ Waiting for reception of IP connection generation request. Upon receiving the TCP / IP connection generation request, the main program generates the requested TCP / IP connection and performs the following processing. First, the request receiving means receives the site selection request message. Next, obtain the URL (V) indicating the service from the site selection request message,
Using this URL as an argument, the server information acquisition means is activated and the URL of the mirror server and the list of installation bases (S (V))
To get Then, the Internet address of the user terminal included in the site selection request message is (U),
The list of the mirror server installation sites indicated by S (V) is set to the candidate site list (Y), and the site selection process corresponding to the process of the main program of the second embodiment is started. next,
Site selection information Z obtained as an output of this site selection processing
The response transmission means is activated with the URL of the mirror server and the list S (V) of installation bases as arguments.

In the same case as in the first and second embodiments, that is, the server providing the service to be provided is the base A,
FIG. 8 shows an execution example in the case where the user terminals are distributed to G, H, and L, the user terminal belongs to the base U, and the activation device belonging to the base A activates the base selection. Starter is Area 1
The request message is transmitted to the site selection device. Area 1
As a result of the estimation processing, the site selecting device of 1 transmits a response message including the address of the site selecting device of area 2. The activation device, in accordance with the received response message, the area 2
The request message is transmitted to the site selection device. As a result of the estimation, the site selection device in area 2 is U of the server in site L.
Send a response message containing the RL.

In the present embodiment, unlike the first and second embodiments, the base L is selected as the base near the user instead of the base H. Since the distance is evaluated by the number of relay devices on the route, more accurate selection is made. That is, the third problem of the problems of the conventional technology is solved. Further, focusing on the route information table of one site selection device, in the first and second embodiments, 6
In contrast to the entries, the third embodiment has four entries. That is, the sizes of the route information table in the first and second embodiments are reduced. That is, the second problem of the problems of the conventional technology is solved.
Furthermore, the server information table in each area only indicates whether or not there is a server of the target service in other areas. Therefore, it is not necessary to update the server information table in another area even if the number of servers increases or decreases or the location changes in one area. This reduces the load of updating the server information table in each area. That is, the fifth problem of the conventional technology is solved.

(Fourth Embodiment) Further, a fourth embodiment in which the site selection device in the third embodiment is improved in order to improve the problems of the third and fourth points of the prior art will be described. .
In the fourth embodiment, routes related to as many bases as possible are registered in the route storage table of the primary selection means, not limited to the server-installed bases. The route information table of area 2 in this embodiment is as shown in Table 11.

[0102]

[Table 11]

The primary selection means executes the estimation process by using all the sites registered in the route storage table as the candidate sites for primary selection. As a result, the primary selection means outputs the base closest to the input user terminal. The evaluation value table for secondary selection is a table in which the evaluation values based on the transfer speed and the load are written for the communication paths between the bases registered in the path information table as shown in Table 12. In this embodiment, the larger the evaluation value for secondary selection is, the better information transfer can be expected.

[0104]

[Table 12]

The secondary selection evaluation value acquisition means is a subroutine that is activated by the secondary selection means, and uses the two pieces of identification information of the bases as arguments and sets the secondary selection evaluation value between the two bases to the second. It is acquired from the evaluation value table for next selection, and this evaluation value for secondary selection is returned as a return value.

The secondary selection means is activated by using the identification information (denoted as z1) of the candidate bases selected as a result of the primary selection and the candidate base list for secondary selection (assumed as Y2) as an argument, and performs the secondary selection. , A subroutine for returning the site selection information (denoted as Z2) as a return value. The secondary selection means activates the secondary selection evaluation value acquisition means using z1 and y2 as arguments for each of the candidate bases y2 included in Y2, and acquires the secondary selection evaluation value regarding y2. Then, the candidate bases are selected in order from the candidate base having the best evaluation value for secondary selection, and Z is selected.
Generates 2.

The main routine program is the same as that of the third embodiment, except for the following points. The list of mirror server installation bases indicated by S (V) acquired by the server information acquisition means is used as a secondary selection candidate base list (Y2). The list of candidate bases when starting the base selection processing corresponding to the main routine of the second embodiment is a predetermined primary selection candidate base list. The output means in this site selection process outputs site selection information (denoted as Z1) that includes only the candidate site with the lowest evaluation value (denoted as z1) from the primary selection candidate site list. Further, after acquiring this site selection information, the main routine program activates the secondary selection means with z1 and Y2 as arguments, and acquires the site selection information Z2 as a return value. Then, the response transmission means is activated using the site selection information Z2, the URL of the mirror server, and the list S (V) of installation sites as arguments.

The site selection operation will be described for the same case as in the first to third embodiments. The primary selection means performs estimation with all the locations in this route information table as candidates, and outputs location K as the estimation result of the location closest to the user terminal. The evaluation value acquisition means for secondary selection refers to the evaluation value table for secondary selection, and the evaluation value of each candidate base is H = 3, L =
Determined as 1. Then, H is set as a candidate base to which the user should output H as a server base near the user.

In this embodiment, the base H is set as the optimum server.
Server is selected. This is a result of using a table that reflects the execution speed of communication between the user terminal site and each site, and solves the fourth problem of the problems of the conventional technology.

(Fifth Embodiment) In a fifth embodiment of the present invention, a mode in which the evaluation value information holds the maximum value and the minimum value of the common path length n will be described. That is, in the second embodiment,
An example in which the evaluation value information holds the maximum value and the minimum value of the common path length n instead of the maximum value and the minimum value of the evaluation value will be described. The maximum value of the common route of a certain candidate base y stored in the evaluation value table is nH (U, y), and the minimum value is nL.
Expressed as (U, y). Let L be the initial value of nH (U, y),
The initial value of nL (U, y) is set to 0. The evaluation value updating means performs the updating process shown in the following formulas (5) and (6) instead of the updating process shown in the formulas (3) and (4). If r (U, m) = r (y, m) is not satisfied nH (U, y)
Update: nH (U, y) ← Min (L-2 * (m−1), nH (U, y)) (5) r (U, m) = r (y, m) Minimum value nL
Update (U, y): nL (U, y) ← Max (L-2 * m, nL (U, y)) (6) Table 13 shows an example of the process of updating the evaluation value information in this embodiment.
Shown in.

[0111]

[Table 13]

When the network interface of the relay device is redundantly N-folded, the route information table stores a plurality of network addresses for the relay device. In this case, in determining whether the first relay device and the second relay device are the same, if the network address of the first relay device matches any of the plurality of network addresses, the first relay device The device and the second relay device shall be considered to be the same.

[0113]

As described above, according to the present invention,
It is possible to significantly reduce the number of packet transfers that occur in the site selection process, reduce the time required for the estimation process, and reduce the network load due to the estimation process. In addition, it is possible to evaluate the distance between the user terminal and the site on a scale other than the number of relay devices, and it is possible to more appropriately estimate the nearby site. In addition, it is possible to execute estimation processing by multiple site selection devices while continuing the estimation process to the site selection device that is closer to the user terminal, and occurs when the user terminal is far from the site where estimation is activated. The estimation error of the relationship between the length and the length of the path, which becomes easier, is reduced. Further, the size of the route information table and the server information table of the estimation device at each base is reduced, so that the required memory resources are reduced.

That is, the present invention realizes a faster and more accurate site selection device with a smaller memory resource consumption as compared with the prior art, and realizes a high speed information distribution service and a large capacity information distribution service. Realize efficiently.

[Brief description of drawings]

FIG. 1 is a diagram showing a basic configuration of a site selection device according to the present invention.

FIG. 2 is a diagram showing a configuration of the present invention for improving calculation of an evaluation value.

FIG. 3 is a diagram showing a site selection system according to the present invention.

FIG. 4 is a diagram showing a site selection system according to the present invention.

FIG. 5 is a diagram showing an arrangement example of a site selection system.

FIG. 6 is a diagram showing a format of a request message.

FIG. 7 is a diagram showing a format of a response message.

FIG. 8 is a flowchart showing the flow of site selection processing.

FIG. 9 is a diagram showing an example of a network used for explanation.

FIG. 10 is a diagram showing a configuration of a site selection device according to a conventional technique.

FIG. 11 is a flowchart showing a flow of processing for selecting a site according to a conventional technique.

FIG. 12 is a diagram for explaining the concept of evaluation values.

FIG. 13 is a diagram showing a configuration of a traceroute executing unit.

FIG. 14 is a flowchart showing a flow of execution of traceroute.

[Explanation of symbols]

1, 46 Relay device information acquisition means 2,42 Route information table 3 Evaluation value updating means 4,47 Continuation judgment means 5,44,48 Output means 6 Evaluation value table 7 Primary selection means 8 Secondary selection means 9 Secondary selection evaluation value acquisition means 10, 17 Request receiving means 11, 19 Response transmission means 12, 20 Request transmission means 13, 21 Response receiving means 14, 22 Re-requesting means 15 Starter 16 site selection device table 18 Server information acquisition means 30, 31, 40 site selection device 41 traceroute execution means 43 Means for obtaining the evaluation value of each candidate site Ra-Ro relay device D1-D3 site selection device U user terminal Sa-Sl server

Continuation of the front page (56) Reference JP-A-11-232201 (JP, A) JP-A-11-261580 (JP, A) JP-A-2001-273223 (JP, A) JP-A-2001-290787 (JP, A) ) JP 2001-202318 (JP, A) JP 2001-109694 (JP, A) JP 2001-69169 (JP, A) JP 2000-172618 (JP, A) JP 2000-138920 (JP, A) ) JP-A-9-305517 (JP, A) Special Table 2002-507092 (JP, A) 59th National Convention of Information, 4S-9 ACM SIGCOMM'95, p288-298 (58) Fields investigated (Int. Cl. ⁷ , DB name) H04L 12/56 G06F 13/00

Claims (23)

(57) [Claims] 1. An observation base where an observation point is placed and one or more candidate bases are defined, and information for identifying a terminal base accommodating a user terminal is used as input information for the observation base. A means for selecting a base from the candidate bases is provided, and the means for selecting stores a list of identification information of relay devices on the route from the observation base to the candidate base for each of the candidate bases. A route information table that stores the distance from the observation point of the relay device for each of the identification information of the relay device, and a route from the observation point to the terminal point and a route from the observation point to the candidate point The evaluation value defined by using the common path length n which is a common part, or the evaluation value in the calculation process includes a value to be held in the calculation process instead of the evaluation value. An evaluation value table that holds evaluation value information for each candidate site, a relay device information acquisition unit that sequentially acquires, one by one, the identification information of the relay devices on the route from the observation site to the user site, and the relay device. Each time the information acquisition means acquires one piece of identification information of a relay device, the relay device identified by this identification information is set as the first relay device, and each of the candidate bases is on the route from the observation base to the candidate base. And, the identification information of the second relay device that is as far as the first relay device from the observation point is acquired from the route information table, and the identification information of the first relay device and the second relay device are acquired. It is determined whether the first relay device and the second relay device are the same based on the identification information of the determination result, and the determination result and the distance m from the observation point to the first or second relay device. By using the evaluation value information of the evaluation value table, it is determined whether the value of the length n of the common path or the range of possible values of n can be specified, and if the value can be specified, the candidate base Evaluation value updating means for updating evaluation value information, and one of the relay apparatus information acquisition means acquired by the relay device information acquisition means whether or not the end condition defined so that it can be determined using the contents of the evaluation value table is satisfied Determined each time the update of the evaluation value table relating to the identification information of the relay device is completed, and if this end condition is not satisfied, the identification information of the unacquired relay device on the route from the observation base to the terminal base is determined. Further, a continuation judging means for instructing to sequentially acquire one by one, and if the continuation judging means judges that the end condition is satisfied, the magnitude relation between the candidate bases of the evaluation value of the evaluation value table According to the above, the output terminal is configured to determine the length relationship between the candidate bases of the distance to the terminal base, preferentially select the candidate bases from the candidate bases having this short distance, and output the base selection information indicating the selected base. A site selection device characterized by being provided. 2. The site selection device according to claim 1, wherein the distance between the sites is defined as the number of relay devices on the route between the sites. 3. The relay device information acquisition means includes means for acquiring the identification information of the relay device from the one closest to the observation base along a route from the observation base to the terminal base, and the evaluation value table. Includes means for setting the initial value of the evaluation value information, which is a value before the acquisition of the identification information of the relay device by the relay device information acquisition means, to a value that the evaluation value cannot take, the evaluation value updating means If the first relay device and the second relay device are the same or the value of the evaluation value information is other than the initial value, the evaluation value information is not updated and the first relay When the device and the second relay device are different and the value of the evaluation value information is an initial value, the distance from the observation base of the relay device identified by the identification information of the relay device acquired last time is set. As the length n of the common path Serial calculating an evaluation value based selection system of claim 1 further comprising means for changing place the value of the evaluation value information by this evaluation value. 4. The termination condition is that the distance of the common part between the route from the observation base to the candidate base and the route from the observation base to the user base has been obtained for all the candidate bases. The site selection device according to claim 1. Wherein said termination condition, the evaluation value based selection device to that billed to claim 1 wherein the the magnitude relation can be identified in the candidate site. 6. The site selection device according to claim 1, wherein the termination condition is that a predetermined number of the candidate sites having the evaluation value equal to or less than a predetermined value are found. 7. The site selection device according to claim 1, wherein a length of a route from the observation site to the candidate site is defined as L, and the evaluation value is defined as L-2n. 8. An arbitrary monotonically increasing or monotonically decreasing function is defined as f, and an output value of f with respect to an input X is defined as f (X),
The site selection device according to claim 1, wherein a length of a route from the observation site to the candidate site is L, and the evaluation value is defined as f (L-2n). 9. The evaluation value table comprises means for storing, as the evaluation value information, evaluation value information including the common route length n or a value to be held in the calculation process in the calculation process of the common route length, The output unit that outputs the site selection information calculates the evaluation value of each of the candidate sites with reference to the common path length n related to the candidate site stored in the evaluation value table, and the candidate site of the evaluation value. The site selection device according to claim 1, further comprising means for determining a magnitude relationship between the sites. 10. The evaluation value table is such that the maximum value of the range of possible values of the evaluation value is the maximum value, the minimum value of the range is the minimum value, and the maximum value and the minimum value are the evaluation values. Means to store as information, the maximum value which is a value at the time before acquisition of the identification information of the relay device by the relay device information, the initial value of the minimum value, the range that the evaluation value can be specified at this time And a means for setting the evaluation value updating means, wherein the evaluation value updating means determines whether the first relay device and the second relay device are the same among the maximum value or the minimum value and the second value. When a identifiable item is calculated by using the distance m of the one relay device from the observation point and the maximum value is identifiable, and the calculated maximum value is smaller than the maximum value stored in the evaluation value table. Is the maximum stored in the table. The minimum value stored in the evaluation value table is calculated when the replacement minimum value can be specified by the calculated maximum value and the calculated minimum value is larger than the minimum value stored in the evaluation value table. The site selection device according to claim 1, further comprising means for replacing the site by the minimum value. 11. The minimum value of the distance between arbitrary relay devices is d.
Since the evaluation value table sets the initial value of the maximum value of the evaluation values in the evaluation value table to a value of L and refers to the above to determine the transfer destination, a high-speed relay device can be realized. Includes a unit that sets an initial value of the minimum value of the fixed evaluation value having an address length to a value of −L, and the evaluation value updating unit is the same as the first relay device and the second relay device. When it is determined that the maximum value can be specified, the maximum value is calculated as L-2m, and when it is determined that the first relay device and the second relay device are different, the minimum value is determined. The site selection device according to claim 7 or 10, further comprising means for determining that the minimum value is identifiable and calculating the minimum value as L-2 (m-d). 12. f is an arbitrary monotonically increasing function, d is a minimum value of a distance between arbitrary relay devices, and the evaluation value table is an initial value of the maximum value of the evaluation values in the evaluation value table. Is set to a value of f (L) and an initial value of the minimum value of the evaluation value is set to a value of f (-L), the evaluation value updating means includes the first relay device and the first relay device. When it is determined that the second relay device is the same, it is determined that the maximum value can be specified, the maximum value is calculated as f (L-2m), and the first relay device and the second relay device are calculated. When it is determined that the minimum values are different, it is determined that the minimum value can be specified, and this minimum value is set to f (L-2 (m
-D)) as a means for calculating
Site selection device described. 13. f is an arbitrary monotonically decreasing function, d is a minimum value of a distance between arbitrary relay devices, and the evaluation value table is an initial value of the maximum value of the evaluation values in the evaluation value table. To a value of f (-L) and an initial value of the minimum value of the evaluation value to a value of f (L), the evaluation value updating means includes the first relay device and the first relay device. When it is determined that the second relay device is the same, it is determined that the minimum value can be specified, the minimum value is calculated as f (L-2m), and the first relay device and the second relay device are calculated. If it is determined that the maximum values are different, it is determined that the maximum value can be specified, and the maximum value is set to f (L-2 (m
-D)) as a means for calculating
Site selection device described. 14. A minimum value of a distance between arbitrary relay devices is d.
In the evaluation value table, instead of the range of possible values of the evaluation value, the maximum value of this range is the maximum value of the range of possible values of the common route length n, and the minimum value of this range is A means for storing this maximum value and this minimum value as the minimum value as the evaluation value information, and an initial value of this maximum value as the value of the length of the route from the observation point to the candidate point, and the initial value of this minimum value. The evaluation value updating means determines that the minimum value can be specified when the first relay device and the second relay device are the same, A means for calculating the minimum value as m and determining the maximum value as identifiable when the first relay device and the second relay device are different from each other and calculating the maximum value as m-d. 11. The site selection device according to claim 9 or 10, including: . 15. The evaluation value updating unit searches the route information table for a route including the identification information of the first relay device; and when a route is found as a result of the search, the route is searched. Means for updating the evaluation value information, assuming that the relay device identification information is acquired by the relay device information acquisition means for each of the relay devices on the section from the observation base to the first relay device above. The site selection device according to claim 1, including the site selection device. 16. The relay device information acquisition means receives a distance k as an input, and a relay device located at a point distant from the observation base by k on a route from the observation base to the terminal base or in the vicinity of the point. Means for acquiring the identification information of the relay device, a means for setting the k to the average of the lengths of the routes in the route information table in the first acquisition of the identification information of the relay device, and the identification information of the relay device for the second time and thereafter. 2. The base selection according to claim 1, further comprising means for setting k to the average of the lengths of the routes when the route including the identification information of the relay device acquired last time in the acquisition of the route exists in the route information table. apparatus. 17. A list of candidate bases different from the list of candidate bases is defined as a list of candidate bases for secondary selection, and the closest candidate base indicated by the means for outputting the base selection information is first selected. The evaluation value for the secondary selection based on a scale different from the scale used for the primary selection is the evaluation value for each communication path from the primary site selected as a candidate site to each of the secondary selection candidate sites. The secondary selection evaluation value acquisition means for acquiring, and the secondary selection means for preferentially selecting the candidate site for secondary selection from the one having a favorable evaluation value for secondary selection. Site selection device. 18. The site selection device according to claim 17, wherein the evaluation value for secondary selection is an evaluation value based on a transfer speed of a communication path between the candidate site and the candidate site for secondary selection. 19. The site selection device according to claim 17, wherein the evaluation value for secondary selection is an evaluation value based on a load on a communication path between the candidate site and the candidate site for secondary selection. 20. A secondary selection distance table storing the secondary selection evaluation values between the secondary selection candidate bases and the candidate bases, wherein the secondary selection evaluation value acquisition means is provided. 18. The site selection device according to claim 17, further comprising: means for acquiring the secondary selection evaluation value by referring to the secondary selection evaluation value table. 21. Request transmission means for transmitting a site selection request message including information for identifying a terminal site to the site selection device according to claim 1, and a selected site or another site selection. Response receiving means for receiving a response message including device identification information from this site selection device, and transmitting a site selection request message to this site selection device when this response message includes the identifier of another site selection device A re-request unit and the site selection device according to any one of claims 1 to 20 are provided, wherein the site selection device receives the site selection request message and sets the terminal site indicated by the message as the input information. And a response message including identification information of a candidate base included in the base selection information output with respect to the input information. Response generating means for generating a message and transmitting the response message to the originator of this site selection request message, and a site selecting device for storing the identification information of the other site selecting device in association with the identification information of the candidate site. A table, the response transmission means refers to the site selection device table, and if the identification information of the site selection device corresponding to the candidate site indicated by the site selection information exists in the table, the identification information And a means for including the response message in the response message. 22. A list of identifiers of servers that provide a service with respect to identification information of a service, and server information acquisition means for acquiring installation site information of each of these servers, wherein the request transmission means is the site The request receiving means includes means for including service identification information in the selection request message, and the request receiving means acquires a list of identifiers of servers providing the service indicated by the site selection request message by using the server information acquiring means. The response transmitting means includes means for setting each installation site of each server as the candidate site, wherein the response transmission means is an identifier of a server installed at the candidate site indicated by the site selection information from the list of identifiers of the server. 22. The site selection system according to claim 21, further comprising means included in. 23. A computer apparatus having predetermined hardware and predetermined basic software installed in the hardware, and further installing the computer apparatus, the computer apparatus according to claim 1. A recording medium having software recorded therein, which is a site selection device or a device corresponding to the site selection system according to claim 21 or 22.