WO2009012810A1 - Database system with multiple layer distribution - Google Patents
Database system with multiple layer distribution Download PDFInfo
- Publication number
- WO2009012810A1 WO2009012810A1 PCT/EP2007/057566 EP2007057566W WO2009012810A1 WO 2009012810 A1 WO2009012810 A1 WO 2009012810A1 EP 2007057566 W EP2007057566 W EP 2007057566W WO 2009012810 A1 WO2009012810 A1 WO 2009012810A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- database
- data cluster
- slave
- data
- slave database
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/27—Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
- G06F16/275—Synchronous replication
Definitions
- the present invention generally relates to database systems and, in particular, to database systems with a plurality of databases distributed in different locations and accessible by different applications.
- a distributed database may be regarded as a plurality of databases physically or logically distributed, likely under control of a central database management system, and wherein storage devices are not all necessarily attached to a common CPU.
- the distributed database might be built up with multiple computers located in the same physical location, or may be dispersed over a network of interconnected computers.
- the distribution of databases instances is not necessarily a consequence of data distribution itself but also for the purpose of data replication in order to obtain high available systems.
- the distributed database system is generally required to accomplish: a transparent distribution, so that the applications interact with the distributed database system as if it were one compact logical system; and transparent transactions, so that each transaction maintains a database integrity across the plurality of distributed databases.
- the transparent distribution where the plurality of databases is distributed in different locations, requires a similar performance for local applications requesting data from a closely located database and for global applications requesting data from a far away located database. This is achieved in a traditional distributed database system by the usage of memory caches in areas closely located with the requester applications. Each memory cache temporary saving data usable by the closely located applications.
- a database system with a two-layer distribution that is, with a master database, which may be distributed in a number of database instances or just being a centralized instance, and with a plurality of slave databases acting as memory caches and provided in areas closely located with the requester applications
- cache management logic that takes care about managing data in the slave databases, managing consistency between cached data in the slave databases and master data in the master database, and managing consistency between different caches in different instances of the slave database.
- WAN Wide Area Network
- HSS Home Subscriber Server
- HLR Home Location Register
- the present invention is aimed to obviate at least some of the above disadvantages and provides for a database system with a master database and a plurality of slave databases to act as memory caches, each slave database connected with a number of applications allowed to request execution of an operation to the database system, and for a method of handling subscriber data in said database system.
- the database system and the method incorporating features to allow effective replication and updating mechanisms where applied in a WAN and shared in a telecommunication system such as a cellular network or an IP Multimedia Subsystem (hereinafter IMS) network.
- IMS IP Multimedia Subsystem
- This method comprises the conventional steps of receiving an operation request for a given subscriber from an application, and returning the execution result towards the requester application.
- the method also includes, in accordance with this first aspect of the present invention a step of configuring so-called data clusters at the master database, wherein each data cluster identifies a number of subscriber data associated with an operation and each data cluster is assigned a priority per slave database, per application, or per combinations thereof; a step of determining at the master database the priority of the data cluster associated with the received operation; and either the steps of providing the data cluster associated with the operation from the master database towards the slave database, replicating the received data cluster at the slave database, and executing the operation with the received data cluster at the slave database, where the priority of the data cluster is higher than for previous operations; or the step of executing the operation with the data cluster at the master database, where the priority of the data cluster is lower than for a previous operation.
- the method may comprise a further step of dynamically updating the priority assigned to the data cluster, the update based on accountability of successful or unsuccessful movements of the data cluster from one slave database to one another and reaching a predefined threshold value for the successful or unsuccessful movement. This step is useful where a fine tune is fetched to accommodate each replicated data cluster in the most suitable slave database.
- the method may include a step of instructing the previous slave database the removal of a data cluster where the priority of the data cluster at said slave database is lower than for a new operation.
- the step of executing the operation with the received data cluster may include a step of updating the master database with the modified data cluster.
- the step of executing the operation with the data cluster at the master database may include a step of updating another slave database with the modified data cluster.
- This method is particularly advantageous where the step of receiving the operation request for a given subscriber is carried out at a slave database connected with the application, since the slave databases are preferably located in a location close to the requester applications.
- the method may further include a step of sending the operation request for the given subscriber from the slave database towards the master database.
- the method may include alternative steps to the step of sending the operation request towards the master database.
- the method may include an additional step of configuring data clusters at the slave database, wherein each data cluster identifies a number of subscriber data associated with the operation; and, additionally, the method may further include a step of requesting a data cluster associated with the operation request for the given subscriber from the slave database towards the master database.
- the step of returning the execution result towards the requester application may be carried out at the slave database connected with the application.
- this step of returning the execution result towards the requester application may include a step of returning the result from the master database to the slave database and a step of returning the result from the slave database towards the requester application.
- the method may be applicable in a database system having a unique centralized master database as well as in a database system wherein the master database comprises a plurality of interconnected cooperating databases.
- the method may further comprise a step of determining a first cooperating database in charge of the data cluster, and a step of providing either the data cluster or the result of executing the operation towards a slave database through a second cooperating database.
- a database system storing subscriber data for subscribers of a telecommunication network, and comprising: a master database; a plurality of slave databases to act as memory caches, each slave database connected with a number of applications allowed to request execution of an operation to the database system; a receiver for receiving an operation request for a given subscriber from an application; and a sender for returning the execution result towards the requester application.
- the master database and the plurality of slave databases are improved in accordance with this second aspect of the present invention to accomplish the above method.
- the master database includes a configuration unit for configuring data clusters, each data cluster identifying a number of subscriber data associated with an operation and assigned a priority per slave database, application, or combinations thereof; a memory module for storing data cluster values per subscriber basis; a processing unit for determining the priority of the data cluster associated with the operation, and for executing the operation with the data cluster where the priority of the data cluster is lower than for a previous operation; and a sender for providing the data cluster associated with the operation towards the slave database where the priority of the data cluster is higher than for previous operations.
- each slave database includes a memory module for replicating a number of data clusters from the master database; a receiver for receiving the data cluster associated with the operation from the master database; and a processing unit for executing the operation with the received data cluster.
- the processing unit of the master database in cooperation with the configuration unit may be arranged for dynamically updating the priority assigned to the data cluster. This update may be based on accountability of successful, or unsuccessful, movements of the data cluster from one slave database to one another and reaching a predefined threshold value for the successful, or unsuccessful, movement.
- the processing unit of the master database in cooperation with the sender may be arranged for instructing a slave database the removal of a data cluster where the priority of the data cluster at the slave database is lower than for a new operation, and the processing unit of the slave database in cooperation with the receiver may be arranged for removing said data cluster from the memory module at the slave database.
- the processing unit in cooperation with a sender of the slave database may be arranged for submitting an update towards the master database with the modified data cluster, and the processing unit of the master database in cooperation with the receiver is arranged for updating said data cluster in the memory module.
- the processing unit of the master database in cooperation with the sender is arranged for submitting an update towards a slave database
- the processing unit of each slave database in cooperation with the receiver is arranged for updating said data cluster in the memory module.
- each slave database may be arranged for receiving the operation request for a given subscriber from the application, since the slave databases are preferably located in a location close to the requester applications.
- each slave database may include a sender for sending the operation request for the given subscriber towards the master database, and the master database may further include a receiver for receiving said operation request.
- each slave database may include a configuration unit for configuring data clusters at the slave database, each data cluster identifying a number of subscriber data associated with the operation.
- each slave database may include a sender for requesting a data cluster associated with the operation request for the given subscriber towards the master database.
- each slave database may include the sender for returning the execution result towards the requester application.
- the sender of the master database may be arranged for returning the execution result towards the slave database
- the receiver of the slave database may be arranged for receiving said result
- the sender of the slave database may be arranged for returning said result towards the requester application.
- This database system may be provided with an arrangement having a unique centralized master database as well as with an arrangement having a plurality of interconnected cooperating databases where data clusters may be distributed.
- the sender of each cooperating database may be arranged for providing either the data cluster or the result of executing the operation towards the slave database directly of through another cooperating database.
- the database system described above is usable in a sort of data layered architecture and can be shared amongst a number of front-end servers to provide external communications towards other entities in a telecommunication network.
- a Home Location Register (hereinafter HLR) holding subscriber data for subscribers of a GSM network, may comprise a front-end server for external communications and the above database system, in accordance with the second aspect of the invention, for storing the subscriber data for said subscribers.
- HLR Home Location Register
- a Home Subscriber Server holding subscriber data for subscribers of an IP Multimedia Subsystem (hereinafter IMS) network
- IMS IP Multimedia Subsystem
- a front-end server for external communications and the above database system, in accordance with the second aspect of the invention, for storing the subscriber data for said subscribers.
- the invention may be practised by one or more computer programs, which may be loadable into an internal memory of a computer with input and output units and with a processing unit, the one or more computer programs comprising executable software, which may be separable in different portions, adapted to carry out the above method steps in the above different entities, server or devices, when running in a number of computers.
- the executable software, or portions thereof may be recorded in a carrier readable medium.
- FIG. 1 is a basic block diagram illustrating a database system having a centralized master database and a plurality of slave databases closely located to a number of applications connected to said slave databases.
- FIG. 2 is a basic block diagram illustrating a database system having a number of cooperating databases acting as a master database, and a plurality of slave databases closely located to a number of applications connected to said slave databases.
- FIG. 3a, 3b and 3c illustrate alternative embodiments for assigning a priority to data clusters per slave database, application requesting an operation, and combinations thereof.
- FIG. 4 illustrates an exemplary table held at the master database of the database system, and dynamically updated during operation, to indicate those particular data clusters replicated in given slave databases and having a given priority.
- FIG. 5 is a basic sequence diagram illustrating the retrieval of a data cluster for executing an operation request at a slave database where the priority of the operation is higher than for previous operations.
- FIG. 6 is a basic sequence diagram illustrating the execution of an operation request at the master database where the priority of a new operation is lower than for a current operation.
- FIG. 7 is a basic block structure presenting the structural elements that a master database and a slave database may comprise in accordance with several embodiments of the invention.
- An important aspect behind this invention is the configuration of subscriber data as data clusters, wherein each data cluster represents an optimal amount of subscriber data associated with each particular operation to be requested from any application.
- the level of granularity selected for each data cluster is another important aspect for the purpose of the present invention and can, nevertheless, be adjusted during operation based on different performance measures obtained with previous values.
- this method includes an initial step of configuring data clusters at the master database 1, wherein each data cluster identifies a number of subscriber data associated with an operation and wherein each data cluster is assigned a priority per slave database, per application, or per combinations thereof.
- Fig. 3a shows a first exemplary configuration wherein each exemplary data cluster 103a is assigned a priority 102a per each application 101a allowed to request execution of an operation to the database system; whereas Fig. 3b shows a second exemplary configuration wherein each exemplary data cluster 103b is assigned a priority 102b per each slave database 104b to act as memory cache; whereas Fig. 3c shows a third exemplary configuration wherein each exemplary data cluster 103c is assigned a priority 102c per each slave database 104c to act as memory cache and per each application 101c allowed to request execution of an operation to the database system.
- the master database 1 includes a configuration unit 10 for configuring data clusters 103a, 103b, 103c, wherein each data cluster identifies a number of subscriber data associated with an operation and is assigned a priority 102a,
- the Fig. 5 illustrates a basic sequence of actions carried out in accordance with a first embodiment of the method wherein an operation associated with a high-priority data cluster is requested to the database system during a step S-IOO.
- Such operation is referred to, for the sake of simplicity, as a "Read / Write Operation" to indicate that the data cluster values might be modified as a result of executing a 'write operation', and are not modified as a result of executing a 'read operation'. Nevertheless, these operations may always require a process rather than a simple read or write instruction in order to produce a result.
- the operation for a given subscriber is submitted from a first application 3a, namely Application- 1, during the step S-IOO and is received at first slave database 2m, namely Slave DB-I.
- first slave database 2m namely Slave DB-I.
- the database system may have a common receiver for receiving operations from the applications.
- this step S-100 of receiving the operation request for a given subscriber is carried out at a slave database 2m, which is connected with the first application 3 a.
- the slave database 2m may include a receiver 21 arranged for receiving the operation request for a given subscriber from the application 3 a.
- the slave database 2m receiving the operation searches during a step S- 105 whether a data cluster associated with the received operation exists at the slave database, and determines during a step S-110 that no such data cluster exists therein.
- Such data cluster is not replicated at the slave database 2m and actions are required from the master database 1 to obtain an execution result to be returned towards the requester application.
- the slave database 2m includes a memory module 20 for replicating a number of data clusters from the master database 1, and a processing unit 22 arranged for determining whether a data cluster associated with the operation exists in the memory module at the slave database.
- the method may further include a step S-115 of sending the operation request for the given subscriber from the slave database 2m towards the master database 1, as Fig. 5 illustrates.
- the method may include a step of configuring data clusters at the slave database
- the slave database 2m may include a sender 23 for sending the operation request for the given subscriber towards the master database 1, and the master database 1 may further include a receiver 14 for receiving said operation request.
- the slave database 2m may include a configuration unit 24 for configuring data clusters at the slave database 2m, wherein each data cluster identifies a number of subscriber data associated with the operation and, additionally, the slave database 2m may include a sender 23 for requesting a data cluster associated with the operation request for the given subscriber towards the master database 1.
- the method illustrated in Fig. 5 continues with a step S- 120 of determining at the master database 1 the priority of the data cluster associated with the operation.
- the master database 1 includes a processing unit 12 for determining the priority of the data cluster associated with the operation. This may be achieved by consulting any table built up as the exemplary tables in Fig. 3a, 3b, or 3c, or the like.
- the processing unit 12 determines during a step S- 125 that the priority of the data cluster associated with this new operation is higher than for previous operations, if any, and triggers the replication of the data cluster towards the slave database currently in charge of the operation.
- the method includes a step S- 130 of providing the data cluster associated with the operation from the master database 1 towards the slave database 2m, a step of replicating the received data cluster at the slave database 2m, and a step S- 140 of executing the operation with the received data cluster at the slave database 2m.
- the master database 1 includes a sender 13 for providing the data cluster associated with the operation towards the slave database 2m where the priority of the data cluster is higher than for previous operation; whereas the receiver 21 of the slave database 2m is arranged for receiving the data cluster associated with the operation from the master database 1, and the processing unit 22 of the slave database 2m is arranged for executing the operation with the received data cluster.
- the method may include a step of updating the master database 1 with modified data cluster values.
- the processing unit 22 in cooperation with the sender 23 of the slave database 2m is arranged for submitting an update towards the master database 1 with modified data cluster values, where the received data cluster is modified as a result of executing the operation, and the processing unit 12 of the master database 1 in cooperation with the receiver 14 is arranged for updating said data cluster in the memory module 11.
- the master database may update during a step S-135 a dynamic table, such as the exemplary one shown in Fig. 4, to track in what particular slave database 104 a particular data cluster 103 is replicated and with what priority was submitted there.
- This update may be done by the processing unit 12 of the master database 1, likely acting on the memory module 11 therein.
- the method may include a step S- 128 of instructing the another slave database 2p the removal of a data cluster.
- the processing unit 12 of the master database 1 in cooperation with the sender 13 is arranged for instructing a slave database 2p the removal of a data cluster where the priority of the data cluster at the slave database 2p is lower than for a new operation
- the processing unit 22 of the slave database 2p in cooperation with the receiver 21 is arranged for removing said data cluster from the memory module 20.
- the method may be enhanced with an additional mechanism to dynamically update the priority assigned to data clusters during the step of configuration. This update may be based on accountability of successful, or unsuccessful, movements of the data cluster from one slave database to one another, and reaching a predefined threshold value for the successful or unsuccessful movement.
- the method may include a step of increasing a so-called 'successful counter' for the priority assigned to the data cluster submitted to the slave database 2m, or a step of decreasing the 'unsuccessful counter' for the priority assigned to the data cluster withdrawn from the slave database 2p, or both, until reaching a predefined threshold high- value or low-value, where the corresponding assigned priority is increased or decreased.
- the processing unit 12 of the master database 1 in cooperation with the configuration unit 10, the memory module 11 or both, is arranged for dynamically updating the priority assigned to the data cluster, the update based on accountability of successful or unsuccessful movements of the data cluster from one slave database to one another and reaching a predefined threshold value for the successful or unsuccessful movement.
- the method ends with a step S-145 of returning the execution result towards the requester application 3a.
- the step S-145 of returning the execution result towards the requester application 3 a may be carried out at the slave database 2m, which is connected with the application 3 a.
- the database system may have a common sender for returning execution results towards the applications.
- the slave database 2m includes a sender 23 for returning the execution result towards the requester application 3 a.
- FIG. 6 illustrates a basic sequence of actions carried out in accordance with this second embodiment of the method wherein an operation associated with a low-priority data cluster is requested to the database system during a step S- 150.
- an operation associated with a low-priority data cluster is requested to the database system during a step S- 150.
- such operation is referred to as a "Read / Write Operation", and may always require a process rather than a simple read or write instruction in order to produce a result.
- This second embodiment has quite a few actions in common with the above first embodiment and, for the sake of simplicity, those particular or alternative features for the common actions, which are enough detailed for the first embodiment, may be partially or totally omitted for this second embodiment without any intention of departing from the above description.
- the operation for a given subscriber is submitted from a first application 3b, namely Application-2, during the step S- 150 and is received at a second slave database 2n, namely Slave DB-2.
- the database system may have a common receiver for receiving operations from the applications.
- this step S- 150 of receiving the operation request for a given subscriber is carried out at the slave database 2n, which is connected with the second application 3b.
- the slave database 2n may include a receiver 21 arranged for receiving the operation request for a given subscriber from the application 3b.
- the slave database 2n receiving the operation searches during a step S- 155 whether a data cluster associated with the received operation exists at the slave database, and determines during a step S- 160 that no such data cluster exists therein.
- Such data cluster is not replicated at the slave database 2n and actions are required from the master database 1 to obtain an execution result to be returned towards the requester application.
- the slave database 2n includes a memory module 20 for replicating a number of data clusters from the master database 1 , and a processing unit 22 arranged for determining whether a data cluster associated with the operation exists in the memory module at the slave database.
- the method may further include a step S- 165 of sending the operation request for the given subscriber from the slave database
- the method may include a step of configuring data clusters at the slave database 2n, each data cluster identifying a number of subscriber data associated with the operation and, additionally, a step of requesting the data cluster associated with the operation request for the given subscriber from the slave database 2n towards the master database 1.
- the slave database 2n may include a sender 23 for sending the operation request for the given subscriber towards the master database 1, and the master database 1 may further include a receiver 14 for receiving said operation request.
- the slave database 2n may include a configuration unit 24 for configuring data clusters at the slave database 2n, wherein each data cluster identifies a number of subscriber data associated with the operation and, additionally, the slave database 2n may include a sender 23 for requesting a data cluster associated with the operation for the given subscriber towards the master database 1.
- the method illustrated in Fig. 6 continues with a step S- 170 of determining at the master database 1 the priority of the data cluster associated with the operation.
- the master database 1 includes a processing unit 12 for determining the priority of the data cluster associated with the operation. This may be achieved by consulting any table built up as the exemplary tables in Fig. 3a, 3b, or 3c, or the like.
- the processing unit 12 determines during a step S- 175 that the priority of the data cluster associated with this new operation is lower than for a previous operation, and triggers the execution of the operation at the master database 1, and with the data cluster therein, during a step S- 180.
- the processing unit 12 of the master database 1 is arranged for executing the operation with the data cluster where the priority of the data cluster is lower than for a previous operation.
- the method may include a step of locally updating the master database 1 with modified data cluster values.
- the modified data cluster associated with the new operation had been previously replicated in another slave database 2q, namely Slave DB-4, and the priority of the data cluster at the another slave database 2q is higher than for the new operation, the method may include a step S- 182 of updating the another slave database 2q with modified data cluster values.
- the processing unit 12 of the master database 1 in cooperation with the sender 13 is arranged for submitting an update towards the another slave database 2q with modified data cluster values, where the data clusters are modified at the master database 1, and the processing unit 22 of the slave database 2q in cooperation with the receiver 21 is arranged for updating said data cluster in the memory module 20.
- the method in accordance with the second embodiment illustrated in Fig. 6, ends with a step of returning the execution result towards the requester application 3b.
- the method may include a step S- 190 of returning the execution result towards the requester application 3b from the slave database 2n, which is connected with the application 3b.
- the database system may have a common sender for returning execution results towards the applications.
- this step of returning the execution result towards the requester application may include a step S- 185 of returning the result from the master database 1 to the slave database 2n and a step S- 190 of returning the result from the slave database 2n towards the requester application 3b.
- the slave database 2n may include a sender 23 arranged for returning the execution result towards the requester application 3b.
- the sender 13 of the master database 1 may be arranged for returning the execution result towards the slave database 2n
- the receiver 21 of the slave database 2n may be arranged for receiving said result
- the sender 23 of the slave database 2n may be arranged for returning said result towards the requester application 3b.
- the method may be applicable in a database system having a unique centralized master database, as shown in Fig. 1, as well as in a database system wherein the master database comprises a plurality of interconnected cooperating databases, as shown in Fig. 2.
- the method may further comprise a step of determining a first cooperating database Ie in charge of the data cluster, and a step of providing either the data cluster or the result of executing the operation towards a slave database 2m, 2n through a second cooperating database Ia, Ic.
- this database system may be provided with an arrangement 1 having a unique centralized master database as well as with an arrangement 1 having a plurality of interconnected cooperating databases Ia, Ib, Ic, Id,
- each cooperating database may be arranged for providing either the data cluster or the result of executing the operation towards the slave database 2m, 2n, 2p, 2q directly of through another cooperating database.
- the database system and the method are usable in a so-called data layered architecture wherein a number of front-end servers, which are arranged to provide external communications towards other entities in a telecommunication network, share the database system.
- a first instance of an entity in a data layer architecture may be a HLR holding subscriber data for subscribers of a GSM network, wherein the HLR comprises a front-end server for external communications and the above database system for storing the subscriber data for said subscribers.
- a second instance of an entity in a data layer architecture may be a HSS holding subscriber data for subscribers of an IMS network, wherein the HSS comprises a front-end server for external communications and the above database system for storing the subscriber data for said subscribers.
- the invention also provides for a computer program, loadable into an internal memory of a computer with input and output units as well as with a processing unit, the computer program comprising executable software adapted to carry out method steps as described above when running in the computer, and wherein the executable software may be recorded in a carrier readable in a computer.
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention addresses the access to subscriber data in a telecommunication system, and provides for a database system with a master database and with a plurality of slave databases acting as memory caches closely located with the requester applications and for a method of handling such database system. The method includes the steps of: configuring data clusters at the master database, each data cluster identifying subscriber data associated with an operation and assigned a priority; determining at the master database the priority of the data cluster associated with a received operation; and either providing the data cluster associated with the operation from the master database towards the slave database, replicating the received data cluster at the slave database, and executing the operation with the received data cluster at the slave database, where the priority of the data cluster is higher than for previous operations; or executing the operation with the data cluster at the master database, where the priority of the data cluster is lower than for a previous operation.
Description
DATABASE SYSTEM WITH MULTIPLE LAYER DISTRIBUTION
TECHNICAL FIELD
[0001] The present invention generally relates to database systems and, in particular, to database systems with a plurality of databases distributed in different locations and accessible by different applications.
BACKGROUND
[0002] Regarding database technologies, the concept of distributed databases has been widely spread for some time in order to address scalability issues. A distributed database may be regarded as a plurality of databases physically or logically distributed, likely under control of a central database management system, and wherein storage devices are not all necessarily attached to a common CPU. Thus, the distributed database might be built up with multiple computers located in the same physical location, or may be dispersed over a network of interconnected computers. Generally speaking, the distribution of databases instances is not necessarily a consequence of data distribution itself but also for the purpose of data replication in order to obtain high available systems.
[0003] Where considering a database system distributed in different physical locations, one has to take into account the different nature of the applications allowed to access such database system in terms of its connections to particular database instances and respective distances, as well as in terms of the data distribution amongst said particular database instances. In this respect, and depending on particular models of data distribution to apply, one may distinguish between local applications, which are connected to a specific database instance having all required data and which do not required data from remote database instances, and global applications, which is connected to any database instances and which requires data from remote database instances.
[0004 ] Even where local and global applications concurrently coexist to access the distributed database system, and particularly where the local and global applications carry out communication functions between network nodes of a telecommunication network, the distributed database system is generally required to accomplish: a transparent distribution,
so that the applications interact with the distributed database system as if it were one compact logical system; and transparent transactions, so that each transaction maintains a database integrity across the plurality of distributed databases.
[0005] The transparent distribution, where the plurality of databases is distributed in different locations, requires a similar performance for local applications requesting data from a closely located database and for global applications requesting data from a far away located database. This is achieved in a traditional distributed database system by the usage of memory caches in areas closely located with the requester applications. Each memory cache temporary saving data usable by the closely located applications.
[0006] On the other hand, where memory caches are provided in areas closely located with the requester applications, the integrity of the database system to be maintained by each transparent transaction, as one compact logical system, requires an updating of all the memory caches each time a transaction modifies data in any particular memory cache.
[0007 ] In other words, where a database system with a two-layer distribution is provided, that is, with a master database, which may be distributed in a number of database instances or just being a centralized instance, and with a plurality of slave databases acting as memory caches and provided in areas closely located with the requester applications, there is a need for a sort of cache management logic that takes care about managing data in the slave databases, managing consistency between cached data in the slave databases and master data in the master database, and managing consistency between different caches in different instances of the slave database.
[0008] Nowadays, different mechanisms are known to address these three previous issues. For instance, the issue of managing data to be cached in the slave database may be addressed by cache algorithms like a so-called "Least Recently Used", a so-called "Least Frequently Used", and the like; whereas managing consistency between cached data in the slave databases and master data in the master database, as well as between different caches in different instances of the slave database, may be addressed by cache coherence models. Regarding the cache coherence models, the most widely known are the so-called "directory- based", "snooping" and "snarfmg". These tree models, where applied to the two-layer DB architecture presented above, have different consequences.
[0009] Regarding the directory-based cache coherence model, there is a directory entry for each data block to be cached which contains information about the caching state of the data block in the system, and the locations of the slave database caching said data block. By checking the state and the locations, one can determine which instances of the slave database need to be updated for an operation in order to maintain coherence.
[0010] Regarding the snooping cache coherence model, at each slave database location, there is a monitor that is aware about changes in data cached in other locations of the slave database. Where these changes take place, the monitor removes the cache data.
[0011] Regarding the snarfmg cache coherence model, at each slave database location, there is a monitor that is aware of changes in the master database and the monitor updates the cached data where there is a change in the master database.
[0012] These three models, and corresponding implementation mechanisms, are very inefficient where the distributed database system is used in a Wide Area Network (hereinafter WAN) and is shared in a telecommunication system by a number of possibly different subscriber register front-ends, such as Home Subscriber Server (HSS) front-ends and Home Location Register (HLR) front-ends may be. In such scenario, the distributed database system is expected to provide almost real time responses as well as real time coherence whilst the slave database locations are geographically separated by long distances, however, the WAN delays adversely affect continuous replications and updates.
SUMMARY
[0013] The present invention is aimed to obviate at least some of the above disadvantages and provides for a database system with a master database and a plurality of slave databases to act as memory caches, each slave database connected with a number of applications allowed to request execution of an operation to the database system, and for a method of handling subscriber data in said database system. The database system and the method incorporating features to allow effective replication and updating mechanisms where applied in a WAN and shared in a telecommunication system such as a cellular network or an IP Multimedia Subsystem (hereinafter IMS) network.
[ 0014 ] Thus, in accordance with a first aspect of the present invention, there is provided a method of handling subscriber data in a database system with a master database and with a plurality of slave databases intended to act as memory caches, wherein each slave database is connected with a number of applications allowed to request execution of an operation to the database system. This method comprises the conventional steps of receiving an operation request for a given subscriber from an application, and returning the execution result towards the requester application.
[0015] The method also includes, in accordance with this first aspect of the present invention a step of configuring so-called data clusters at the master database, wherein each data cluster identifies a number of subscriber data associated with an operation and each data cluster is assigned a priority per slave database, per application, or per combinations thereof; a step of determining at the master database the priority of the data cluster associated with the received operation; and either the steps of providing the data cluster associated with the operation from the master database towards the slave database, replicating the received data cluster at the slave database, and executing the operation with the received data cluster at the slave database, where the priority of the data cluster is higher than for previous operations; or the step of executing the operation with the data cluster at the master database, where the priority of the data cluster is lower than for a previous operation.
[0016] Regarding the priority assigned to each data cluster, the method may comprise a further step of dynamically updating the priority assigned to the data cluster, the update based on accountability of successful or unsuccessful movements of the data cluster from one slave database to one another and reaching a predefined threshold value for the successful or unsuccessful movement. This step is useful where a fine tune is fetched to accommodate each replicated data cluster in the most suitable slave database.
[0017 ] In particular, where the data cluster had been replicated in other slave database for a previous operation, the method may include a step of instructing the previous slave database the removal of a data cluster where the priority of the data cluster at said slave database is lower than for a new operation. Moreover, where the data cluster is modified as a result of executing the operation at the slave database, the step of executing the operation with the received data cluster may include a step of updating the master database with the
modified data cluster. Furthermore, where the data cluster is modified as a result of executing the operation at the master database, the step of executing the operation with the data cluster at the master database may include a step of updating another slave database with the modified data cluster.
[0018] This method is particularly advantageous where the step of receiving the operation request for a given subscriber is carried out at a slave database connected with the application, since the slave databases are preferably located in a location close to the requester applications. In such a case, and even more advantageously where the method further includes a step of determining that no data cluster associated with the operation exists at the slave database, the method may further include a step of sending the operation request for the given subscriber from the slave database towards the master database.
[0019] The method may include alternative steps to the step of sending the operation request towards the master database. In this respect, the method may include an additional step of configuring data clusters at the slave database, wherein each data cluster identifies a number of subscriber data associated with the operation; and, additionally, the method may further include a step of requesting a data cluster associated with the operation request for the given subscriber from the slave database towards the master database.
[ 0020 ] On the one hand, the step of returning the execution result towards the requester application, in this method, may be carried out at the slave database connected with the application. On the other hand, this step of returning the execution result towards the requester application may include a step of returning the result from the master database to the slave database and a step of returning the result from the slave database towards the requester application.
[0021] The method may be applicable in a database system having a unique centralized master database as well as in a database system wherein the master database comprises a plurality of interconnected cooperating databases. For the latter, the method may further comprise a step of determining a first cooperating database in charge of the data cluster, and a step of providing either the data cluster or the result of executing the operation towards a slave database through a second cooperating database.
[0022] In accordance with a second aspect of the present invention, there is provided a database system storing subscriber data for subscribers of a telecommunication network, and comprising: a master database; a plurality of slave databases to act as memory caches, each slave database connected with a number of applications allowed to request execution of an operation to the database system; a receiver for receiving an operation request for a given subscriber from an application; and a sender for returning the execution result towards the requester application.
[0023] In this database system, the master database and the plurality of slave databases are improved in accordance with this second aspect of the present invention to accomplish the above method.
[0024 ] Thus, the master database includes a configuration unit for configuring data clusters, each data cluster identifying a number of subscriber data associated with an operation and assigned a priority per slave database, application, or combinations thereof; a memory module for storing data cluster values per subscriber basis; a processing unit for determining the priority of the data cluster associated with the operation, and for executing the operation with the data cluster where the priority of the data cluster is lower than for a previous operation; and a sender for providing the data cluster associated with the operation towards the slave database where the priority of the data cluster is higher than for previous operations. Apart from that, each slave database includes a memory module for replicating a number of data clusters from the master database; a receiver for receiving the data cluster associated with the operation from the master database; and a processing unit for executing the operation with the received data cluster.
[0025] Regarding the priority assigned to each data cluster, the processing unit of the master database in cooperation with the configuration unit may be arranged for dynamically updating the priority assigned to the data cluster. This update may be based on accountability of successful, or unsuccessful, movements of the data cluster from one slave database to one another and reaching a predefined threshold value for the successful, or unsuccessful, movement.
[0026] In particular, where the data cluster had been replicated in other slave database for a previous operation, the processing unit of the master database in cooperation with the
sender may be arranged for instructing a slave database the removal of a data cluster where the priority of the data cluster at the slave database is lower than for a new operation, and the processing unit of the slave database in cooperation with the receiver may be arranged for removing said data cluster from the memory module at the slave database. Moreover, where the data cluster is modified as a result of executing the operation at the slave database, the processing unit in cooperation with a sender of the slave database may be arranged for submitting an update towards the master database with the modified data cluster, and the processing unit of the master database in cooperation with the receiver is arranged for updating said data cluster in the memory module. Furthermore, where the data cluster is modified as a result of executing the operation at the master database, the processing unit of the master database in cooperation with the sender is arranged for submitting an update towards a slave database, and the processing unit of each slave database in cooperation with the receiver is arranged for updating said data cluster in the memory module.
[0027 ] As advantageous as for the above method, the receiver of each slave database may be arranged for receiving the operation request for a given subscriber from the application, since the slave databases are preferably located in a location close to the requester applications. In such a case, and even more advantageously where the processing unit of each slave database is arranged for determining that no data cluster associated with the operation exists in the memory module at the slave database, each slave database may include a sender for sending the operation request for the given subscriber towards the master database, and the master database may further include a receiver for receiving said operation request.
[0028] Alternatively to the sending of the operation request towards the master database, each slave database may include a configuration unit for configuring data clusters at the slave database, each data cluster identifying a number of subscriber data associated with the operation. In this case, each slave database may include a sender for requesting a data cluster associated with the operation request for the given subscriber towards the master database.
[0029] Regarding the return of the execution result towards the requester application, each slave database may include the sender for returning the execution result towards the
requester application. On the other hand, the sender of the master database may be arranged for returning the execution result towards the slave database, the receiver of the slave database may be arranged for receiving said result, and the sender of the slave database may be arranged for returning said result towards the requester application.
[0030] This database system may be provided with an arrangement having a unique centralized master database as well as with an arrangement having a plurality of interconnected cooperating databases where data clusters may be distributed. In the latter arrangement, the sender of each cooperating database may be arranged for providing either the data cluster or the result of executing the operation towards the slave database directly of through another cooperating database.
[ 0031 ] In accordance with a third aspect of the invention, the database system described above is usable in a sort of data layered architecture and can be shared amongst a number of front-end servers to provide external communications towards other entities in a telecommunication network.
[0032] For instance, a Home Location Register (hereinafter HLR) holding subscriber data for subscribers of a GSM network, may comprise a front-end server for external communications and the above database system, in accordance with the second aspect of the invention, for storing the subscriber data for said subscribers.
[0033] Also for instance, a Home Subscriber Server (hereinafter HSS) holding subscriber data for subscribers of an IP Multimedia Subsystem (hereinafter IMS) network, may comprise a front-end server for external communications and the above database system, in accordance with the second aspect of the invention, for storing the subscriber data for said subscribers.
[ 0034 ] The invention may be practised by one or more computer programs, which may be loadable into an internal memory of a computer with input and output units and with a processing unit, the one or more computer programs comprising executable software, which may be separable in different portions, adapted to carry out the above method steps in the above different entities, server or devices, when running in a number of computers. In particular, the executable software, or portions thereof, may be recorded in a carrier readable medium.
BRIEF DESCRIPTION OF THE DRAWINGS
[0035] The features, objects and advantages of the invention will become apparent by reading this description in conjunction with the accompanying drawings, in which:
[0036] FIG. 1 is a basic block diagram illustrating a database system having a centralized master database and a plurality of slave databases closely located to a number of applications connected to said slave databases.
[0037 ] FIG. 2 is a basic block diagram illustrating a database system having a number of cooperating databases acting as a master database, and a plurality of slave databases closely located to a number of applications connected to said slave databases.
[0038] FIG. 3a, 3b and 3c illustrate alternative embodiments for assigning a priority to data clusters per slave database, application requesting an operation, and combinations thereof.
[0039] FIG. 4 illustrates an exemplary table held at the master database of the database system, and dynamically updated during operation, to indicate those particular data clusters replicated in given slave databases and having a given priority.
[0040] FIG. 5 is a basic sequence diagram illustrating the retrieval of a data cluster for executing an operation request at a slave database where the priority of the operation is higher than for previous operations.
[0041] FIG. 6 is a basic sequence diagram illustrating the execution of an operation request at the master database where the priority of a new operation is lower than for a current operation.
[0042] FIG. 7 is a basic block structure presenting the structural elements that a master database and a slave database may comprise in accordance with several embodiments of the invention.
DETAILED DESCRIPTION
[0043] The following describes some preferred embodiments for a method of handling subscriber data in a database system with a master database 1 and with a plurality of slave
databases 2m, 2n, 2p, 2q to act as memory caches, as illustrated in Fig. 1 and Fig. 2, each slave database connected with a number of applications 3a, 3b, 3c, 3d, 3f allowed to request execution of an operation to the database system; as well as embodiments of the database system illustrated in Fig. 7.
[0044 ] An important aspect behind this invention is the configuration of subscriber data as data clusters, wherein each data cluster represents an optimal amount of subscriber data associated with each particular operation to be requested from any application. The level of granularity selected for each data cluster is another important aspect for the purpose of the present invention and can, nevertheless, be adjusted during operation based on different performance measures obtained with previous values.
[0045] Therefore, this method includes an initial step of configuring data clusters at the master database 1, wherein each data cluster identifies a number of subscriber data associated with an operation and wherein each data cluster is assigned a priority per slave database, per application, or per combinations thereof. In this respect, Fig. 3a shows a first exemplary configuration wherein each exemplary data cluster 103a is assigned a priority 102a per each application 101a allowed to request execution of an operation to the database system; whereas Fig. 3b shows a second exemplary configuration wherein each exemplary data cluster 103b is assigned a priority 102b per each slave database 104b to act as memory cache; whereas Fig. 3c shows a third exemplary configuration wherein each exemplary data cluster 103c is assigned a priority 102c per each slave database 104c to act as memory cache and per each application 101c allowed to request execution of an operation to the database system.
[0046] To this end, as Fig. 7 illustrates, the master database 1 includes a configuration unit 10 for configuring data clusters 103a, 103b, 103c, wherein each data cluster identifies a number of subscriber data associated with an operation and is assigned a priority 102a,
102b, 102c per slave database 104b, application 101a, or combinations thereof 101c, 104c; and also a memory module 11 for storing data cluster values per subscriber basis.
[0047 ] Once data clusters are configured in the master database 1, the Fig. 5 illustrates a basic sequence of actions carried out in accordance with a first embodiment of the method wherein an operation associated with a high-priority data cluster is requested to the
database system during a step S-IOO. Such operation is referred to, for the sake of simplicity, as a "Read / Write Operation" to indicate that the data cluster values might be modified as a result of executing a 'write operation', and are not modified as a result of executing a 'read operation'. Nevertheless, these operations may always require a process rather than a simple read or write instruction in order to produce a result.
[0048] In this first embodiment, the operation for a given subscriber is submitted from a first application 3a, namely Application- 1, during the step S-IOO and is received at first slave database 2m, namely Slave DB-I. Generally speaking, the database system may have a common receiver for receiving operations from the applications. However, in accordance with an advantageous embodiment, this step S-100 of receiving the operation request for a given subscriber is carried out at a slave database 2m, which is connected with the first application 3 a. To this end, as Fig. 7 illustrates, the slave database 2m may include a receiver 21 arranged for receiving the operation request for a given subscriber from the application 3 a.
[0049] In this first embodiment illustrated in Fig. 5, the slave database 2m receiving the operation searches during a step S- 105 whether a data cluster associated with the received operation exists at the slave database, and determines during a step S-110 that no such data cluster exists therein. Such data cluster is not replicated at the slave database 2m and actions are required from the master database 1 to obtain an execution result to be returned towards the requester application. To this end, as Fig. 7 illustrates, the slave database 2m includes a memory module 20 for replicating a number of data clusters from the master database 1, and a processing unit 22 arranged for determining whether a data cluster associated with the operation exists in the memory module at the slave database.
[0050] In this respect, the method may further include a step S-115 of sending the operation request for the given subscriber from the slave database 2m towards the master database 1, as Fig. 5 illustrates. Alternatively or complementary, and not illustrated in any drawing, the method may include a step of configuring data clusters at the slave database
2m, each data cluster identifying a number of subscriber data associated with the operation and, additionally, a step of requesting the data cluster associated with the operation request for the given subscriber from the slave database 2m towards the master database 1. To this end, as Fig. 7 illustrates, the slave database 2m may include a sender 23 for sending the
operation request for the given subscriber towards the master database 1, and the master database 1 may further include a receiver 14 for receiving said operation request. Alternatively or complementary, the slave database 2m may include a configuration unit 24 for configuring data clusters at the slave database 2m, wherein each data cluster identifies a number of subscriber data associated with the operation and, additionally, the slave database 2m may include a sender 23 for requesting a data cluster associated with the operation request for the given subscriber towards the master database 1.
[0051] Whatever previous alternative is followed, the method illustrated in Fig. 5 continues with a step S- 120 of determining at the master database 1 the priority of the data cluster associated with the operation. To this end, as Fig. 7 illustrates, the master database 1 includes a processing unit 12 for determining the priority of the data cluster associated with the operation. This may be achieved by consulting any table built up as the exemplary tables in Fig. 3a, 3b, or 3c, or the like.
[0052] In this case illustrated in Fig. 5, the processing unit 12 determines during a step S- 125 that the priority of the data cluster associated with this new operation is higher than for previous operations, if any, and triggers the replication of the data cluster towards the slave database currently in charge of the operation. Thus, the method includes a step S- 130 of providing the data cluster associated with the operation from the master database 1 towards the slave database 2m, a step of replicating the received data cluster at the slave database 2m, and a step S- 140 of executing the operation with the received data cluster at the slave database 2m. To this end, as Fig. 7 illustrates, the master database 1 includes a sender 13 for providing the data cluster associated with the operation towards the slave database 2m where the priority of the data cluster is higher than for previous operation; whereas the receiver 21 of the slave database 2m is arranged for receiving the data cluster associated with the operation from the master database 1, and the processing unit 22 of the slave database 2m is arranged for executing the operation with the received data cluster.
[0053] Particular applicable in this case, where the data cluster is modified at the slave database 2m as a result of executing the operation during the step S- 140, the method may include a step of updating the master database 1 with modified data cluster values. To this end, as Fig. 7 illustrates, the processing unit 22 in cooperation with the sender 23 of the slave database 2m is arranged for submitting an update towards the master database 1 with
modified data cluster values, where the received data cluster is modified as a result of executing the operation, and the processing unit 12 of the master database 1 in cooperation with the receiver 14 is arranged for updating said data cluster in the memory module 11.
[0054 ] Once the data cluster has been submitted for replication at the slave database, the master database may update during a step S-135 a dynamic table, such as the exemplary one shown in Fig. 4, to track in what particular slave database 104 a particular data cluster 103 is replicated and with what priority was submitted there. This update may be done by the processing unit 12 of the master database 1, likely acting on the memory module 11 therein.
[0055] In particular, where the data cluster associated with the new operation had been previously replicated in another slave database 2p, namely Slave DB-3, and the priority of the data cluster at the another slave database 2p is lower than for the new operation, the method may include a step S- 128 of instructing the another slave database 2p the removal of a data cluster. To this end, as Fig. 7 illustrates, the processing unit 12 of the master database 1 in cooperation with the sender 13 is arranged for instructing a slave database 2p the removal of a data cluster where the priority of the data cluster at the slave database 2p is lower than for a new operation, and the processing unit 22 of the slave database 2p in cooperation with the receiver 21 is arranged for removing said data cluster from the memory module 20.
[0056] As already commented above, the method may be enhanced with an additional mechanism to dynamically update the priority assigned to data clusters during the step of configuration. This update may be based on accountability of successful, or unsuccessful, movements of the data cluster from one slave database to one another, and reaching a predefined threshold value for the successful or unsuccessful movement. For instance, in the present embodiment, the method may include a step of increasing a so-called 'successful counter' for the priority assigned to the data cluster submitted to the slave database 2m, or a step of decreasing the 'unsuccessful counter' for the priority assigned to the data cluster withdrawn from the slave database 2p, or both, until reaching a predefined threshold high- value or low-value, where the corresponding assigned priority is increased or decreased. To this end, as Fig. 7 illustrates, the processing unit 12 of the master database 1 in cooperation with the configuration unit 10, the memory module 11 or both, is arranged for dynamically
updating the priority assigned to the data cluster, the update based on accountability of successful or unsuccessful movements of the data cluster from one slave database to one another and reaching a predefined threshold value for the successful or unsuccessful movement.
[ 0057 ] The method, in accordance with the embodiment illustrated in Fig. 5, ends with a step S-145 of returning the execution result towards the requester application 3a. In this first embodiment, aligned with the reception at the slave database 2m of the operation for a given subscriber from the application 3a, the step S-145 of returning the execution result towards the requester application 3 a may be carried out at the slave database 2m, which is connected with the application 3 a. Generally speaking, the database system may have a common sender for returning execution results towards the applications. However, in accordance with this advantageous embodiment, as Fig. 7 illustrates, the slave database 2m includes a sender 23 for returning the execution result towards the requester application 3 a.
[0058] A second embodiment of this method is illustrated in Fig. 6. Once data clusters have been configured in the master database 1 , as for the above first embodiment, the Fig. 6 illustrates a basic sequence of actions carried out in accordance with this second embodiment of the method wherein an operation associated with a low-priority data cluster is requested to the database system during a step S- 150. As before and for the same purpose, such operation is referred to as a "Read / Write Operation", and may always require a process rather than a simple read or write instruction in order to produce a result.
[0059] This second embodiment has quite a few actions in common with the above first embodiment and, for the sake of simplicity, those particular or alternative features for the common actions, which are enough detailed for the first embodiment, may be partially or totally omitted for this second embodiment without any intention of departing from the above description.
[0060] In this second embodiment, the operation for a given subscriber is submitted from a first application 3b, namely Application-2, during the step S- 150 and is received at a second slave database 2n, namely Slave DB-2. As for the previous embodiment, the database system may have a common receiver for receiving operations from the applications. However, in accordance with an advantageous embodiment, this step S- 150 of
receiving the operation request for a given subscriber is carried out at the slave database 2n, which is connected with the second application 3b. To this end, as Fig. 7 illustrates, the slave database 2n may include a receiver 21 arranged for receiving the operation request for a given subscriber from the application 3b.
[0061] In this second embodiment illustrated in Fig. 6, the slave database 2n receiving the operation searches during a step S- 155 whether a data cluster associated with the received operation exists at the slave database, and determines during a step S- 160 that no such data cluster exists therein. Such data cluster is not replicated at the slave database 2n and actions are required from the master database 1 to obtain an execution result to be returned towards the requester application. To this end, as Fig. 7 illustrates, the slave database 2n includes a memory module 20 for replicating a number of data clusters from the master database 1 , and a processing unit 22 arranged for determining whether a data cluster associated with the operation exists in the memory module at the slave database.
[0062] In this respect and as for the first embodiment, the method may further include a step S- 165 of sending the operation request for the given subscriber from the slave database
2n towards the master database 1, as Fig. 6 illustrates. Alternatively or complementary, and not illustrated in any drawing, the method may include a step of configuring data clusters at the slave database 2n, each data cluster identifying a number of subscriber data associated with the operation and, additionally, a step of requesting the data cluster associated with the operation request for the given subscriber from the slave database 2n towards the master database 1. To this end, as Fig. 7 illustrates, the slave database 2n may include a sender 23 for sending the operation request for the given subscriber towards the master database 1, and the master database 1 may further include a receiver 14 for receiving said operation request. Alternatively or complementary, the slave database 2n may include a configuration unit 24 for configuring data clusters at the slave database 2n, wherein each data cluster identifies a number of subscriber data associated with the operation and, additionally, the slave database 2n may include a sender 23 for requesting a data cluster associated with the operation for the given subscriber towards the master database 1.
[0063] Whatever previous alternative is followed, the method illustrated in Fig. 6 continues with a step S- 170 of determining at the master database 1 the priority of the data cluster associated with the operation. To this end, as Fig. 7 illustrates, the master database 1
includes a processing unit 12 for determining the priority of the data cluster associated with the operation. This may be achieved by consulting any table built up as the exemplary tables in Fig. 3a, 3b, or 3c, or the like.
[0064 ] In this case illustrated in Fig. 6, the processing unit 12 determines during a step S- 175 that the priority of the data cluster associated with this new operation is lower than for a previous operation, and triggers the execution of the operation at the master database 1, and with the data cluster therein, during a step S- 180. To this end, as Fig. 7 illustrates, the processing unit 12 of the master database 1 is arranged for executing the operation with the data cluster where the priority of the data cluster is lower than for a previous operation.
[0065] In particular, where the data cluster is modified as a result of executing the operation during the step S- 180, the method may include a step of locally updating the master database 1 with modified data cluster values. Moreover, where the modified data cluster associated with the new operation had been previously replicated in another slave database 2q, namely Slave DB-4, and the priority of the data cluster at the another slave database 2q is higher than for the new operation, the method may include a step S- 182 of updating the another slave database 2q with modified data cluster values. To this end, as Fig. 7 illustrates, the processing unit 12 of the master database 1 in cooperation with the sender 13 is arranged for submitting an update towards the another slave database 2q with modified data cluster values, where the data clusters are modified at the master database 1, and the processing unit 22 of the slave database 2q in cooperation with the receiver 21 is arranged for updating said data cluster in the memory module 20.
[0066] The method, in accordance with the second embodiment illustrated in Fig. 6, ends with a step of returning the execution result towards the requester application 3b. In particular, the method may include a step S- 190 of returning the execution result towards the requester application 3b from the slave database 2n, which is connected with the application 3b.
[0067 ] Generally speaking, the database system may have a common sender for returning execution results towards the applications. However, in this second embodiment illustrated in Fig. 6, this step of returning the execution result towards the requester application may include a step S- 185 of returning the result from the master database 1 to
the slave database 2n and a step S- 190 of returning the result from the slave database 2n towards the requester application 3b.
[0068] To this end, as Fig. 7 illustrates, the slave database 2n may include a sender 23 arranged for returning the execution result towards the requester application 3b. In particular, the sender 13 of the master database 1 may be arranged for returning the execution result towards the slave database 2n, the receiver 21 of the slave database 2n may be arranged for receiving said result, and the sender 23 of the slave database 2n may be arranged for returning said result towards the requester application 3b.
[0069] The method, under the first or second embodiments respectively illustrated in Fig. 5 and Fig. 6 or combinations thereof, may be applicable in a database system having a unique centralized master database, as shown in Fig. 1, as well as in a database system wherein the master database comprises a plurality of interconnected cooperating databases, as shown in Fig. 2. For the latter, the method may further comprise a step of determining a first cooperating database Ie in charge of the data cluster, and a step of providing either the data cluster or the result of executing the operation towards a slave database 2m, 2n through a second cooperating database Ia, Ic. To this end, this database system may be provided with an arrangement 1 having a unique centralized master database as well as with an arrangement 1 having a plurality of interconnected cooperating databases Ia, Ib, Ic, Id,
Ie where data clusters may be distributed. In the latter arrangement, the sender 13 of each cooperating database may be arranged for providing either the data cluster or the result of executing the operation towards the slave database 2m, 2n, 2p, 2q directly of through another cooperating database.
[0070] As already disclosed above, the database system and the method, under the first or second embodiments or combinations thereof, are usable in a so-called data layered architecture wherein a number of front-end servers, which are arranged to provide external communications towards other entities in a telecommunication network, share the database system.
[0071] A first instance of an entity in a data layer architecture may be a HLR holding subscriber data for subscribers of a GSM network, wherein the HLR comprises a front-end
server for external communications and the above database system for storing the subscriber data for said subscribers.
[ 0072 ] A second instance of an entity in a data layer architecture may be a HSS holding subscriber data for subscribers of an IMS network, wherein the HSS comprises a front-end server for external communications and the above database system for storing the subscriber data for said subscribers.
[0073] The invention also provides for a computer program, loadable into an internal memory of a computer with input and output units as well as with a processing unit, the computer program comprising executable software adapted to carry out method steps as described above when running in the computer, and wherein the executable software may be recorded in a carrier readable in a computer.
[0074 ] The invention is described above in respect of several embodiments in an illustrative and non-restrictive manner. Obviously, variations, and combinations of these embodiments are possible in light of the above teachings, and any modification of the embodiments that fall within the scope of the claims is intended to be included therein.
Claims
1. A method of handling subscriber data in a database system with a master database (1) and with a plurality of slave databases (2m, 2n, 2p, 2q) to act as memory caches, each slave database connected with a number of applications (3a, 3b, 3c, 3d, 3f) allowed to request execution of an operation to the database system, the method comprising the steps of:
- receiving (S-IOO, S- 150) an operation request for a given subscriber from an application (3 a, 3b); and
- returning (S- 145, S- 190) the execution result towards the requester application (3a, 3b);
and characterized by the steps of:
- configuring data clusters (103a, 103b, 103c) at the master database (1), each data cluster identifying a number of subscriber data associated with an operation and assigned a priority (102a, 102b, 102c) per slave database (104b), application ( 101 a), or combinations thereof (101c, 104c);
- determining (S-120, S-170) at the master database (1) the priority of the data cluster associated with the operation; and
- either providing (S- 130) the data cluster associated with the operation from the master database (1) towards the slave database (2m), replicating the received data cluster at the slave database (2m), and executing (S- 140) the operation with the received data cluster at the slave database (2m), where the priority of the data cluster is higher than for previous operations;
- or executing (S- 180) the operation with the data cluster at the master database (1), where the priority of the data cluster is lower than for a previous operation.
2. The method of claim 1, wherein the step of receiving (S-100, S- 150) the operation request for a given subscriber is carried out at a slave database (2m, 2n) connected with the application (3a, 3b).
3. The method of claim 1, wherein the step of returning (S- 145, S- 190) the execution result towards the requester application (3a, 3b) is carried out at a slave database (2m, 2n) connected with the application (3a, 3b).
4. The method of claim 1, further comprising a step of determining (S-I lO, S- 160) that no data cluster associated with the operation exists at the slave database (2m, 2n).
5. The method of claim 2, further comprising a step of sending (S-115, S- 165) the operation request for the given subscriber from the slave database (2m, 2n) towards the master database (1).
6. The method of claim 1, further comprising a step of configuring data clusters at the slave database (2m, 2n), each data cluster identifying a number of subscriber data associated with the operation.
7. The method of claim 6, further comprising a step of requesting a data cluster associated with the operation request for the given subscriber from the slave database (2m, 2n) towards the master database (1).
8. The method of claim 1, wherein the step of returning (S- 185, S- 190) the result towards the requester application (3b) includes a step of returning (S- 185) the result from the master database (1) to the slave database (2n) and a step of returning (S- 190) the result from the slave database (2n) towards the requester application (3b).
9. The method of claim 1, further comprising a step of instructing (S-128) a slave database (2p) the removal of a data cluster where the priority of the data cluster at the slave database (2p) is lower than for a new operation (S-100).
10. The method of claim 1, wherein the step of executing (S-140) the operation with the received data cluster at the slave database (2m) includes a step of updating the master database (1) where the data cluster is modified as a result of executing the operation.
11. The method of claim 1, wherein the step of executing (S- 180) the operation with the data cluster at the master database (1) includes a step of updating (S-182) another slave database (2q) with modified data cluster values where the data cluster is modified as a result of executing the operation, and where he priority of the data cluster at the another slave database (2q) is higher than for the new operation.
12. The method of claim 1, wherein the master database (1) includes a plurality of interconnected cooperating databases (Ia, Ib, Ic, Id, Ie), and further comprising a step of determining a first cooperating database (Ie) in charge of the data cluster, and a step of providing either the data cluster or the result of executing the operation towards a slave database (2m, 2n) through a second cooperating database (Ia, Ic).
13. The method of claim 1, further comprising a step of dynamically updating the priority assigned to the data cluster, the update based on accountability of successful or unsuccessful movements of the data cluster from one slave database to one another and reaching a predefined threshold value for the successful or unsuccessful movement.
14. A database system storing subscriber data for subscribers of a telecommunication network, the database system comprising: a master database (1); a plurality of slave databases (2m, 2n, 2p, 2q) to act as memory caches, each slave database connected with a number of applications (3a, 3b, 3c, 3d, 3f) allowed to request execution of an operation to the database system; a receiver (14, 21) for receiving (S-IOO, S-150) an operation request for a given subscriber from an application (3a, 3b); and a sender (13, 23) for returning (S- 145, S- 190) the execution result towards the requester application (3a, 3b);
and the database system characterized in that:
- the master database (1) includes a configuration unit (10) for configuring data clusters (103a, 103b, 103c), each data cluster identifying a number of subscriber data associated with an operation and assigned a priority (102a, 102b, 102c) per slave database (104b), application (101a), or combinations thereof (101c, 104c); a memory module (11) for storing data cluster values per subscriber basis; a processing unit (12) for determining (S-120, S-170) the priority of the data cluster associated with the operation, and for executing (S- 180) the operation with the data cluster where the priority of the data cluster is lower than for a previous operation; and a sender (13) for providing (S- 130) the data cluster associated with the operation towards the slave database (2m) where the priority of the data cluster is higher than for previous operations; and
- each slave database (2m, 2n, 2p, 2q) includes a memory module (20) for replicating a number of data clusters from the master database (1); a receiver (21) for receiving the data cluster associated with the operation from the master database (1); and a processing unit (22) for executing (S-140) the operation with the received data cluster.
15. The database system of claim 14, wherein the processing unit (22) of each slave database (2m, 2n, 2p, 2q) is arranged for determining (S-I lO, S- 160) whether a data cluster associated with the operation exists in the memory module at the slave database.
16. The database system of claim 14, wherein the receiver (21) of each slave database (2m, 2n, 2p, 2q) is arranged for receiving (S-100, S- 150) the operation request for a given subscriber from the application (3a, 3b).
17. The database system of claim 14, wherein each slave database (2m, 2n, 2p, 2q) includes the sender (23) for returning (S- 145, S- 190) the execution result towards the requester application (3a, 3b).
18. The database system of claim 16, wherein each slave database (2m, 2n, 2p, 2q) includes a sender (23) for sending (S-115, S-165) the operation request for the given subscriber towards the master database (1), and the master database (1) further includes a receiver (14) for receiving said operation request.
19. The database system of claim 16, wherein each slave database (2m, 2n, 2p, 2q) includes a configuration unit (24) for configuring data clusters at the slave database (2m, 2n), each data cluster identifying a number of subscriber data associated with the operation.
20. The database system of claim 19, wherein each slave database (2m, 2n, 2p, 2q) includes a sender (23) for requesting a data cluster associated with the operation request for the given subscriber towards the master database (1).
21. The database system of claim 17, wherein the sender (13) of the master database (1) is arranged for returning (S- 185) the execution result towards the requester application (3b) through the slave database (2n), the receiver (21) of the slave database (2n) is arranged for receiving said result, and the sender (23) of the slave database (2n) is arranged for returning (S- 190) said result towards the requester application (3b).
22. The database system of claim 14, wherein the processing unit (12) of the master database (1) in cooperation with the sender (13) is arranged for instructing a slave database (2p) the removal of a data cluster where the priority of the data cluster at the slave database (2p) is lower than for a new operation, and wherein the processing unit (22) of the slave database (2p) in cooperation with the receiver (21) is arranged for removing said data cluster from the memory module (20).
23. The database system of claim 14, wherein the processing unit (22) in cooperation with a sender (23) of each slave database (2m, 2n, 2p, 2q) is arranged for submitting an update towards the master database (1) where the received data cluster is modified as a result of executing the operation, and wherein the processing unit (12) of the master database (1) in cooperation with the receiver (14) is arranged for updating said data cluster in the memory module (11).
24. The database system of claim 14, wherein the processing unit (12) of the master database (1) in cooperation with the sender (13) is arranged for submitting an update towards a slave database (2q) where the data clusters are modified at the master database (1) as a result of executing the operation, and wherein the processing unit (22) of each slave database (2m, 2n, 2p, 2q) in cooperation with the receiver (21) is arranged for updating said data cluster in the memory module (20).
25. The database system of claim 14, wherein the master database (1) includes a plurality of interconnected cooperating databases (Ia, Ib, Ic, Id, Ie) where data clusters are distributed, the sender (13) of each cooperating database being arranged for providing either the data cluster or the result of executing the operation towards the slave database (2m, 2n) directly of through another cooperating database (Ia, Ic).
26. The database system of claim 14, wherein the processing unit (12) of the master database (1) in cooperation with the configuration unit (10) is arranged for dynamically updating the priority assigned to the data cluster, the update based on accountability of successful or unsuccessful movements of the data cluster from one slave database to one another and reaching a predefined threshold value for the successful or unsuccessful movement.
27. A Home Location Register "HLR" holding subscriber data for subscribers of a 2nd generation cellular network, and comprising: a front-end server (3a, 3b, 3e) for external communications and the database system of claim 14 for storing the subscriber data for said subscribers.
28. A Home Subscriber Server "HSS" holding subscriber data for subscribers of an IP Multimedia Subsystem "IMS" network, and comprising: a front-end server (3c, 3d,
3f) for external communications and the database system of claim 14 for storing the subscriber data for said subscribers.
29. A computer program, loadable into an internal memory of a computer with input and output units as well as with a processing unit, the computer program comprising executable software adapted to carry out the method steps according to any of claims
1 to 13 when running in the computer.
30. The computer program of claim 29, wherein the executable software is recorded in a carrier readable in a computer.
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/EP2007/057566 WO2009012810A1 (en) | 2007-07-23 | 2007-07-23 | Database system with multiple layer distribution |
EP07787809A EP2171619A1 (en) | 2007-07-23 | 2007-07-23 | Database system with multiple layer distribution |
US12/670,198 US8818942B2 (en) | 2007-07-23 | 2007-07-23 | Database system with multiple layer distribution |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/EP2007/057566 WO2009012810A1 (en) | 2007-07-23 | 2007-07-23 | Database system with multiple layer distribution |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2009012810A1 true WO2009012810A1 (en) | 2009-01-29 |
Family
ID=38659379
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/EP2007/057566 WO2009012810A1 (en) | 2007-07-23 | 2007-07-23 | Database system with multiple layer distribution |
Country Status (3)
Country | Link |
---|---|
US (1) | US8818942B2 (en) |
EP (1) | EP2171619A1 (en) |
WO (1) | WO2009012810A1 (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2011042064A1 (en) * | 2009-10-09 | 2011-04-14 | Telefonaktiebolaget Lm Ericsson (Publ) | Method for handling data stored by a communication system |
CN110413692A (en) * | 2019-07-30 | 2019-11-05 | 中国工商银行股份有限公司 | Data processing method and server for data-base cluster |
Families Citing this family (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7974982B2 (en) * | 2008-02-04 | 2011-07-05 | Disney Enterprises, Inc. | System and method for device profiling using cascaded databases |
US8612383B2 (en) * | 2008-11-05 | 2013-12-17 | Mastercard International Incorporated | Method and systems for caching objects in a computer system |
CN102142008B (en) * | 2010-12-02 | 2013-04-17 | 华为技术有限公司 | Method and system for implementing distributed memory database, token controller and memory database |
WO2012076976A1 (en) * | 2010-12-08 | 2012-06-14 | N&N Chopra Consultants Pvt. Ltd. | System and method for integrating software functionalities on n-layer architecture platform |
US10317858B2 (en) * | 2011-04-07 | 2019-06-11 | Infosys Technologies, Ltd. | Architecture and method for centrally controlling a plurality of building automation systems |
US9043573B2 (en) * | 2012-06-07 | 2015-05-26 | Netapp, Inc. | System and method for determining a level of success of operations on an abstraction of multiple logical data storage containers |
US8996810B2 (en) | 2012-12-10 | 2015-03-31 | Facebook, Inc. | System and method of detecting cache inconsistencies |
US9571603B2 (en) * | 2013-09-17 | 2017-02-14 | Cisco Technology, Inc. | Redundancy network protocol system |
CN104462225B (en) * | 2014-11-12 | 2018-01-12 | 华为技术有限公司 | The method, apparatus and system of a kind of digital independent |
EP3076308B1 (en) * | 2015-03-31 | 2017-11-29 | Sap Se | Telecommunication method for handling a database query in a telecommunication system |
US10210027B1 (en) * | 2015-09-17 | 2019-02-19 | EMC IP Holding Company LLC | Cluster management |
CN110069365B (en) * | 2019-04-26 | 2023-07-04 | 腾讯科技(深圳)有限公司 | Method for managing database and corresponding device, computer readable storage medium |
US10802973B1 (en) | 2019-07-01 | 2020-10-13 | Bank Of America Corporation | Data access tool |
CN111737173B (en) * | 2020-06-24 | 2022-03-22 | 浪潮(北京)电子信息产业有限公司 | I2C bus communication control method, device, system and readable storage medium |
US11405451B2 (en) * | 2020-09-30 | 2022-08-02 | Jpmorgan Chase Bank, N.A. | Data pipeline architecture |
Family Cites Families (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6523032B1 (en) * | 2000-05-12 | 2003-02-18 | Oracle Corporation | Servicing database requests using read-only database servers coupled to a master database server |
JP3772690B2 (en) * | 2001-04-25 | 2006-05-10 | 日本電気株式会社 | Service system and service method used therefor |
EP1470684B1 (en) * | 2002-01-21 | 2007-11-07 | Nokia Corporation | Method and system for changing a subscription |
JP4080366B2 (en) * | 2003-04-01 | 2008-04-23 | シャープ株式会社 | Network terminal, network system, and network terminal control method |
US6954821B2 (en) * | 2003-07-31 | 2005-10-11 | Freescale Semiconductor, Inc. | Crossbar switch that supports a multi-port slave device and method of operation |
US7917495B1 (en) * | 2004-06-11 | 2011-03-29 | Seisint, Inc. | System and method for processing query requests in a database system |
US7523110B2 (en) * | 2005-03-03 | 2009-04-21 | Gravic, Inc. | High availability designated winner data replication |
US7353311B2 (en) * | 2005-06-01 | 2008-04-01 | Freescale Semiconductor, Inc. | Method of accessing information and system therefor |
US9058372B2 (en) * | 2006-08-23 | 2015-06-16 | Kyocera Corporation | Database management in a wireless communication system |
US20080096550A1 (en) * | 2006-10-24 | 2008-04-24 | Nokia Corporation | Automatic replacement of a mobile sub-scriber identity code |
-
2007
- 2007-07-23 US US12/670,198 patent/US8818942B2/en active Active
- 2007-07-23 WO PCT/EP2007/057566 patent/WO2009012810A1/en active Application Filing
- 2007-07-23 EP EP07787809A patent/EP2171619A1/en not_active Withdrawn
Non-Patent Citations (3)
Title |
---|
HUICAN ZHU, BEN SMITH, TAO YANG: "A Scheduling Framework for Web Server Clusters with Intensive Dynamic Content Processing", TECHNICAL REPORT TRCS-98-29, CS DEPT, UCSB, November 1998 (1998-11-01), pages 1 - 23, XP002458591, Retrieved from the Internet <URL:http://citeseer.ist.psu.edu/zhu98scheduling.html> [retrieved on 20071108] * |
MIGUEL BARREIRO, JOSE LUIS FREIRE, VICTOR M. GULIAS, JUAN J. SANCHEZ: "Exploiting Sequential Libraries on a Cluster of Computers", PROC. OF THE ERLANG WORKSHOP 2001. FLORENCE, ITALY., September 2001 (2001-09-01), pages 1 - 8, XP002458592, Retrieved from the Internet <URL:http://www.erlang.se/workshop/juanjo.ps> [retrieved on 20071113] * |
PLATTNER C ET AL: "Ganymed: scalable replication for transactional Web applications", LECTURE NOTES IN COMPUTER SCIENCE, SPRINGER VERLAG, BERLIN, DE, vol. 3231, 2004, pages 155 - 174, XP002370197, ISSN: 0302-9743 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2011042064A1 (en) * | 2009-10-09 | 2011-04-14 | Telefonaktiebolaget Lm Ericsson (Publ) | Method for handling data stored by a communication system |
US9282143B2 (en) | 2009-10-09 | 2016-03-08 | Telefonaktiebolaget L M Ericsson (Publ) | Method for handling data stored by a communication system |
CN110413692A (en) * | 2019-07-30 | 2019-11-05 | 中国工商银行股份有限公司 | Data processing method and server for data-base cluster |
Also Published As
Publication number | Publication date |
---|---|
US20100257137A1 (en) | 2010-10-07 |
US8818942B2 (en) | 2014-08-26 |
EP2171619A1 (en) | 2010-04-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2009012810A1 (en) | Database system with multiple layer distribution | |
US8176256B2 (en) | Cache regions | |
US8261020B2 (en) | Cache enumeration and indexing | |
US7941401B2 (en) | Distributed data management system | |
US8484417B2 (en) | Location updates for a distributed data store | |
Bronson et al. | {TAO}:{Facebook’s} distributed data store for the social graph | |
CN111858097A (en) | Distributed database system and database access method | |
US8700855B2 (en) | System and method for supporting a tiered cache | |
US20140195566A1 (en) | Distributed data management system | |
EP2288997B1 (en) | Distributed cache arrangement | |
EP2002343B1 (en) | Multi-cache cooperation for response output caching | |
US20020198883A1 (en) | Web system having clustered application servers and clustered databases | |
US10055349B2 (en) | Cache coherence protocol | |
CN100437590C (en) | Method for prefetching object | |
US20070118632A1 (en) | System and method for providing a directory service network | |
US20110225120A1 (en) | System for maintaining a distributed database using leases | |
CN103365987B (en) | Clustered database system and data processing method based on shared-disk framework | |
US11366598B1 (en) | Dynamic lease assignments in a time-series database | |
US9922031B2 (en) | System and method for efficient directory performance using non-persistent storage | |
Gu et al. | Practically realizable efficient data allocation and replication strategies for distributed databases with buffer constraints | |
Patgiri | MDS: In-depth insight | |
CN117539915B (en) | Data processing method and related device | |
KR101419379B1 (en) | Method of reducing network load and structure of node for multi processor system with distributed memory | |
Hao | Large-Scale Key-Value Stores | |
Srinivasa et al. | Comparative Study on Mostly Used NoSQL Databases |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 07787809 Country of ref document: EP Kind code of ref document: A1 |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
REEP | Request for entry into the european phase |
Ref document number: 2007787809 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2007787809 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 12670198 Country of ref document: US |