JPH07311704A - Directory system - Google Patents

Abstract

PURPOSE:To provide a directory system which can retrieve information regarding an object fast and obtain high reliability. CONSTITUTION:The directory system wherein a system control part 2 having responded to access to the information regarding the object actuates a directory access part to retrieve a directory to obtain the information is provided with a hash directory 6, a hash directory access part 8, and a hash function operation part 10; and when no information is obtained by the retrieval in the directory, the system control part 2 actuates a hash function operation part 10 to determine the identifier of the hash directory by making a hash function operate on the identifier of the object, and actuates the hash directory access part 8 identified with the identifier to retrieve the hash directory, thereby obtaining the information.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a directory system using a hash directory. In the information processing system, applications and data that are conventionally desired to be accessed (hereinafter referred to as objects).
The directory is used to read information about the machine (machine where the object exists, access method, etc.). The method of using the directory is such that when there is no information regarding the desired object in its own directory, other directory (s) are accessed. At that time, high speed and reliability are required.

[0002]

2. Description of the Related Art When accessing a directory as described above, when there is no desired information in its own directory, there are three known conventional access methods for accessing another directory. Has been.

The first access method is (1) in FIG.
As shown in, the access method is based on the multiple introduction mode. This access method obtains desired information by issuing a request to a plurality of other directories and selecting an appropriate response from the requests.

The second access method is an access method in a chain mode, as shown in (2) of FIG. According to this access method, when desired information is not obtained in its own directory, a request is issued to one directory other than its own, the directory is circulated until the request is satisfied, and the obtained information is The desired information is obtained by returning to the request source via the route opposite to the request.

In the third access method, first, when desired information cannot be obtained in its own directory, a request is issued to another directory other than its own, and when the request is not satisfied, another directory is introduced. The desired information is obtained by sequentially performing the procedure of receiving the information.

[0006]

In all of the above three access methods, desired information is obtained by repeating the sequential processing.

Therefore, it takes a considerable amount of time to obtain the desired information. In particular, this tendency becomes remarkable when the number of cells increases, which is not suitable for applications requiring real-time performance. A cell means a management unit of an object group, and each cell has a directory. Also, even if you have a replica in the directory,
It may be difficult to determine the extent of replicas to have. The reason is that the wider the range of replicas, the shorter the search time, but if the contents of the directory (information about objects) are updated accordingly, the consistency between the original and the replicas is maintained. Because it will be difficult.

The present invention was created in view of the above technical problems, and an object of the present invention is to provide a directory system capable of high-speed retrieval of information on objects and high reliability.

[0009]

FIG. 1 is a block diagram showing the principle of the invention according to claims 1 to 5. According to the first aspect of the present invention, as shown in FIG. 1, the system control unit 2 which responds to the occurrence of the access to the information about the object activates the directory access unit 3 to search the directory 4 to retrieve the information about the object. The subdirectory system to be obtained is provided with a hash directory 6, a hash directory access section 8, a directory system having a hash function operating section 10, and a communication control section 12 connecting the directory systems, and the directory search is performed. When there is no information about the object in, the system control unit 2 activates the hash function acting unit 8, acts a hash function on the identifier of the object, determines the hash directory identifier of the hash directory, and determines the hash directory. Knowledge When the hash directory 6 identified by the child is the own directory, the hash directory access unit 8 is activated to search the hash directory 6 to obtain information about the object, and the determined hash directory identifier is the own directory. When the hash directory identifier is not for the system, the hash directory identifier is transferred to the directory system having the hash directory identified by the hash directory identifier via the communication control unit (12) to request the retrieval of the information on the object, The hash directory access unit (8) of the directory system having the hash directory identified by the directory identifier is activated to search the hash directory (6) of the directory system to search for the object. Wherein the return information about-object to the requested directory system.

According to a second aspect of the present invention, as shown in FIG. 1, in the directory system according to the first aspect, when there is no actual hash directory, a sub-hash function is applied to the hash directory identifier to cause an actual hash directory. Is determined, and the existing hash directory (6) is searched with the hash directory identifier.

According to a third aspect of the present invention, as shown in FIG. 1, in the directory system according to any of the first and second aspects, an actual hash directory 6 exists.
When a new hash directory is added to, a hash function is applied to the identifier of the object for the new existing hash directory to determine the hash directory identifier of the hash directory, and the existing hash directory is determined by the hash directory identifier. When the hash directory corresponding to the hash directory identifier of the hash directory determined by searching the directory and applying the hash function to the object identifier does not actually exist, the subhash function is applied to the determined hash directory identifier to realize the existence. The hash directory identifier of the hash directory to be determined is determined, and the existing hash directory 6 is searched with the hash directory identifier.

According to a fourth aspect of the present invention, as shown in FIG. 1, in the directory system according to any one of the first to third aspects, when the directory becomes an obstacle, a hash function is directly added to the identifier of the object. It is characterized by making it act.

According to a fifth aspect of the present invention, as shown in FIG. 1, in the directory system according to any one of the first to third aspects, the hash directory 6 is constructed by adding the directory management information to the information about the object. However, when the directory 4 fails, the hash directory 6 is searched with the directory management information to restore the information of the failed directory.

[0014]

In the invention described in claim 1, the system control unit 2 is
In response to the occurrence of the access to the information about the object, the directory access unit 3 is activated to start the directory 4
When there is no information related to the object, the system control unit 2 activates the hash function operating unit 8 to apply a hash function to the identifier of the object to determine the hash directory identifier of the hash directory 6. When the hash directory is in the local system, the hash directory access unit 8 is activated with the hash directory identifier to search the hash directory 6 to obtain information about the object.

When the hash directory identifier is not in the own system, the hash directory identifier is requested via the communication control unit 12 to the directory system identified by the hash directory identifier to retrieve the information about the object, and the information about the object is searched from there. Is to get.

In this way, the hash directory having the information about the object can be found immediately, so that the overhead due to the interaction between the directories can be greatly reduced.

According to the second aspect of the present invention, when there is no existing hash directory, the sub-hash function is further applied to the hash directory identifier to obtain the hash directory identifier of the existing hash directory, and the hash directory identifier is identified. It gets information about an object from a hash directory.

Therefore, even if the hash directory identifier of the existing hash directory cannot be determined by merely operating the hash function, the hash directory identifier of the existing hash directory can be determined to obtain information about the object. it can.

According to the third aspect of the present invention, by using the hash function and the sub-hash function, even when the hash directory is added, the information about the desired object can be obtained from the existing hash directory or the additional hash directory. Obtainable.

According to the fourth aspect of the present invention, when the directory becomes an obstacle, the hash function can be directly applied to the identifier of the object without accessing the directory. Therefore, the directory becomes an obstacle. Also, it is possible to obtain information about a desired object.

According to the fifth aspect of the present invention, by adding the directory management information to the information about the object, it is possible to search the hash directory other than the failure occurrence by the directory management information and restore the failure occurrence directory.

[0022]

FIG. 2 shows an embodiment of the invention described in claims 1 to 5. In FIG. 2, 30 is a system control unit,
32 is a directory access unit, 34 is a directory,
36 is a hash function acting unit, 38 is a hash directory access unit, 40 is a hash directory, and 42 is a communication control unit. 44 is a directory system,
System control unit 30, directory access unit 32, directory 34, hash function acting unit 36, hash directory access unit 38, hash directory 4
0 and the communication control unit 42 form a directory system. The directory system is provided for each cell. A cell is a management unit of an object group as will be apparent from the later description. The cell here is the OSF / DCE (Open Systems Foundation / Distr).
ibuted Computing Environment) A generic term for cells and the management area of each directory in the cell.

The system control unit 30 controls registration and retrieval of information regarding objects. Directory access unit 32 activated by system control unit 30
Accesses the directory 34. Directory 3
When there is no information about the object in 4, the hash function acting unit 36 applies the hash function to the identifier of the object to determine the hash directory identifier of the hash directory. The hash directory access unit 38 searches the hash directory 40 in response to the hash directory identifier. When the hash directory identifier determined by the hash function acting unit 36 is not the hash directory in its own directory system, the system control unit 30 accesses another hash directory via the communication control unit 42.

In FIG. 2, the system control unit 30 corresponds to the system control unit 2 in FIG. 1, and the directory access unit 32 corresponds to the directory access unit 3 in FIG. The directory 34 corresponds to the directory 4 in FIG. 1, and the hash directory 40 corresponds to the hash directory 6 shown in FIG. The hash directory access unit 38 corresponds to the hash directory access unit 8 in FIG. 1, and the hash function acting unit 36 corresponds to the hash function acting unit 10 shown in FIG.

The operation according to the invention of claim 1 in the directory system configured as described above will be described with reference to FIGS.
Will be described below. First, registration of information about an object will be described.

When an object generated by an information processing system (not shown) (this object is called object A) transfers a registration request for information about the object to the directory system, the system controller 30 _{0 of the} directory system. Activates the directory access unit 32 ₀ of its own directory system and registers the information about the object in the directory 34 ₀ (see (1) and (2) in FIG. 3).

Next, the system control unit 30 ₀ activates the hash function acting unit 36 ₀ to act a predetermined hash function on the identifier of the registered object, and the resulting hash directory identifier of the hash directory is obtained. decide.

[0028] If the hash directory 44 ₀ identified by the determined hashed directory identifier, which is to determine whether or not within the own directory system is a hash directory in the local directory system, The system control unit 30 ₀ uses the hash directory access unit 38 in its own directory system.
₀ is activated and registered in the hash directory 40 ₀ specified by the hash directory identifier. Wherein when the determination indicates that this is the other directory system, the system control unit 30 _0, directory system 44 ₁ having a hash directory specified by the hash directory identifier the determined start the communication control unit 42 The information about the object is transmitted to (see (3) in FIG. 3).

[0029] receives information about the object, and the system control unit 30 ₁ of the directory system 44 ₁ that is identified by the hash directory identifier, information about the object to launch a hash directory access unit 38 ₁ of the own directory system 44 ₁ Is registered in the hash directory 40 ₁ (see FIG. 3).
(See (4)).

Next, a search for information about the object A registered as described above will be described. When an object belonging to a certain cell (this object is referred to as an object B. The object B is not necessarily in the same cell as the object A) wants to obtain information about the object A, the object B first Directory system in own cell 44 ₂ (subscript 2 is cell #
It shows that it is 2. The same applies hereinafter. ) Is requested to retrieve information about object A ((1) in FIG. 4).
See). In response to the request, the system control unit 30 ₂ activates the directory access unit 32 to activate the directory 34
₂ is searched for information on object A (see (2) in FIG. 4). When the information is obtained by searching, the system control unit 30 ₂ passes the information to the object B.

When the information about the object A cannot be obtained by the above-mentioned search, the system control unit 30 ₂ applies a predetermined hash function to the identifier of the object A to determine the hash directory identifier of the hash directory. Hash directory identified by the hash directory identifier, if those of its own directory system 44 _2, the system control unit 30 ₂ searches the hash directory 40 ₂ Start hash Directory Access unit 38 ₂ The obtained information is passed to the object B.

If the hash directory identified by the determined identifier is not in the own directory system, the system control unit 30 ₂ causes the communication control unit 42
₂ is activated and the directory system having the corresponding hash directory is requested to retrieve the information on the object A (see (3) in FIG. 4).

In the requested directory system, for example, the directory system 44 ₁ , the system control unit 30 ₁ activates the hash directory access unit 38 ₁ to search the hash directory 40 ₁ and obtain information about the object A ( See (4) in Fig. 4). The system control unit 30 ₁ of the directory system 44 ₁ that is requested returns information about the object A obtained by starting the communication control unit 42 ₁ of the requesting directory system 44 ₂ (in FIG. 4 (5) see) . The request-source directory system 44 ₂ passes the received information about the object to the object B via the system control unit 30 ₂ of the directory system 44 ₂ (see (6) in FIG. 5). When there is no information about the desired object in the directory 34 ₂ , the information about the object A can be immediately obtained from the hash directory, so that the overhead due to the interaction between the directories is significantly reduced and the information about the object can be obtained at high speed. I can do it.

Next, embodiments of the inventions described in claims 2 and 3 will be described below. This embodiment is an example of using a hash function and a subhash function. It is assumed that the hash directory identifier obtained by applying the hash function to the identifier of the object is both represented by an integer, and the hash function is an operation for obtaining the remainder divided by 100. Therefore, 100 hash directories allocated from 0 to 99 can be provided as the hash function identifiers. However, it is assumed that the hash directories that actually exist are 10 hash directories from 0 to 9, that is, the number of cells is 10. Then, from 0 to 99 obtained by operating the hash function in order to distribute the hash directory identifiers to the 10 hash directories.
In order to further distribute the integers up to 0 to 9 as a subhash function, it is assumed that an operation for obtaining a remainder divided by 10 is used.

Under such a condition, registration of the object A whose object identifier is 3210 (see FIG. 9) will be described. 32 as the object identifier
When the hash function of 100 is applied to the identifier of object A to which 10 is assigned, number 10 can be obtained as the hash directory identifier of the hash directory.

However, under the above example, since there are only 10 hash directories, when the 10 subhash functions are applied to the 10 hash directory identifiers, the hash directory identifier of the 0th hash directory is obtained. Comes.

Since the hash directory identifier of the hash directory thus obtained is number 0, the information about the object A having the identifier 3210 is registered in the 0th hash directory.
Similarly, among the hash directory identifiers obtained by applying the hash function, the hash directory identifiers other than the 10th hash directory identifier 20, 30, 40, 50, 60, 7
When 0, 80 and 90 act the 10 sub-hash functions, they all become 0-th hash directory identifiers. Therefore, the 0-th hash directory identifier obtained by sequentially operating the hash function and the sub-hash function is Any object having an identifiable object identifier is registered in the 0th hash directory (see hash directory # 0 in FIG. 5).

Similarly, the hash directory identifier 1,
The information about the object of the object that has determined 11, 21, ..., 91 is registered in the first hash directory. Similarly, the information regarding the object of the object for which the hash directory identifiers 2, 22, ..., 92 are determined is registered in the second hash directory. The same applies hereinafter.

Further, the retrieval of the information about the object in a certain directory system is performed in the same manner as described above. Next, assume that the 10th hash directory is newly installed. This is equivalent to the 10th cell being newly established.

In this case, from the information about the object registered in the 0th hash directory, the information that should have been originally registered in the 10th hash directory is extracted, and those pieces of information are extracted. Move from the directory to the 10th hash directory.

When this processing is performed, the information on the object A having the identifier 3210 is transferred to the 10th hash directory (see FIG. 6). And
For objects having identifiers whose hash directory identifiers can be determined to be 11th to 99th by the action of the hash function, 10 subhash functions are applied to those hash directory identifiers in the same manner as in the conventional case. The hash directory identifiers up to the second are determined, and information about the object is registered in the hash directory identified by these hash directory identifiers. For example, when the identifier of the object is 5411, 11 hash directory identifiers are obtained as the hash directory identifier by operating the hash function, and by further operating the sub-hash function, the first hash directory identifier becomes Since it is obtained, the information about the object of the object having the identifier 5411 is registered in the first hash directory.

The same applies to the identifiers of other objects. The operation of retrieving the information about the object from the directory system and passing the information to the requested object is also performed in the same manner as described above.

By using the sub-hash function, when only the hash function is used as the number of hash directories, even when the number of hash directories becomes insufficient, the hash directories can be distributed and registered in that number. Moreover, it is possible to deal with the addition or deletion of the hash directory.

Next, an embodiment of the invention described in claim 4 will be described below. This embodiment is an embodiment when a failure occurs in a directory of a certain directory system.

The operation of this embodiment is such that, when the object B in the embodiment described in claim 1 wants to obtain information about the object A, when the directory of the cell (directory system) to which the object B belongs fails. It is an example of how to obtain information about.

Since the information about the object cannot be obtained from the directory of the directory system to which the object B belongs, the process for searching the hash directory of the directory system to which the object B belongs is started. Since this search processing has been described in detail in the invention described in claim 1, the description thereof will be omitted. Therefore, even if a directory fails, one object can obtain information about another object.

Finally, an embodiment of the invention described in claim 5 will be described below. This embodiment is an invention relating to recovery of a directory when a failure occurs in the directory.

In this case, a management cell identifier is added to each hash directory as shown in FIG. 7 (see FIG. 6). When a failure occurs in the directory 34, an object belonging to a cell in which the directory temporarily becomes a failure using the management cell identifier in the 0th (# 0) hash directory to the 9th (# 9) hash directory as a key. By retrieving and collecting only the information regarding the object and temporarily re-registering the information regarding those objects in the directory in which the failure has occurred, it is possible to recover the directory in which the failure occurred. For example, when the directory 34 of the 0th cell becomes an obstacle, information regarding the objects of the hash directories whose management cell identifiers of the hash directories 0th to 9th hash directories is 0 is temporarily collected and retrieved. Re-register in the directory that has no major failures.

By doing so, the directory in which the temporary failure has occurred can be restored.

[0050]

As described above, according to the present invention, even if there is no information about an object in the directory, the information about the object is registered by applying a predetermined hash function to the identifier of the object. Since the hash directory can be found immediately, the overhead due to the interaction between the directories can be greatly reduced, and a countermeasure can be taken when a failure occurs. Therefore, a fast and reliable directory service can be provided. By using the sub-hash function, it is possible to deal with addition and deletion of hash directories.

[Brief description of drawings]

FIG. 1 is a principle block diagram of the invention according to claims 1 to 5.

FIG. 2 is a diagram showing an embodiment of the invention described in claims 1 to 5;

FIG. 3 is a diagram showing an example of registration of information about objects.

FIG. 4 is a diagram illustrating a search example of information regarding an object.

FIG. 5 is a diagram showing initial setting to each hash directory.

FIG. 6 is a diagram illustrating a processing example when a hash directory is added.

FIG. 7 is a diagram showing restoration of a directory that has become a temporary failure by using a hash directory.

FIG. 8 is a diagram showing a data configuration example in a hash directory.

FIG. 9 is a diagram showing a conventional directory search example.

[Explanation of symbols]

 2 system control unit 3 directory access unit 4 directory 6 hash directory 8 hash directory access unit 10 hash function acting unit 12 communication control unit 30 system control unit 32 directory access unit 34 directory 36 hash function acting unit 38 hash directory access unit 40 hash directory 42 Communication control unit

Claims (5)

[Claims] 1. A directory system in which a system control unit (2) which responds to an access to information about an object activates a directory access unit (3) to search a directory (4) to obtain information about the object. , A hash directory (6), a hash directory access unit (8), a hash function operating unit (10), and a communication control unit (12) for connecting the directory system.
And when there is no information about the object in the search of the directory, the system control unit (2) activates a hash function acting unit (8) to apply a hash function to the identifier of the object to perform the hash operation. The hash directory identifier of the directory is determined, and when the hash directory (6) identified by the hash directory identifier belongs to the own directory system, the hash directory access unit (8) is activated to open the hash directory (6). When information about the object is retrieved and the determined hash directory identifier does not belong to the own directory system, the hash directory identifier is identified by the hash directory identifier via the communication control unit (12). Transfers to a directory system having a directory to request retrieval of information about an object, and activates a hash directory access unit (8) of the directory system having the hash directory identified by the hash directory identifier to activate the hash directory of the directory system. A directory system, wherein (6) is searched and information about the object is returned to the requested directory system. 2. The directory system according to claim 1, wherein when there is no existing hash directory, a subhash function is applied to the hash directory identifier to determine the hash directory identifier of the existing hash directory, and the hash directory identifier is used. A directory system characterized by searching an existing hash directory (6). 3. The directory system according to any one of claims 1 to 4, wherein when a new hash directory is added to the existing hash directory (6), an object is added to the new existing hash directory. Of the hash directory of the hash directory determined by applying a hash function to the identifier of the hash directory identifier, searching the existing hash directory with the hash directory identifier, and applying the hash function to the identifier of the object to determine the hash of the hash directory When the hash directory corresponding to the directory identifier does not actually exist, the sub-hash function is applied to the determined hash directory identifier to determine the hash directory identifier of the existing hash directory. , Directory system characterized by searching the hash directory (6) for real in the hash directory identifier. 4. The directory system according to any one of claims 1 to 5, wherein when the directory becomes an obstacle, information about the object is obtained by directly acting a hash function on the identifier of the object. And the directory system. 5. The directory system according to claim 1, wherein the hash directory (6) is configured by adding directory management information to the information about the object, and when the directory (4) temporarily fails. A directory system characterized by searching a hash directory (6) with directory management information and recovering information of a temporarily failed directory.