MXPA97001115A - A method of configuration for a da handling system - Google Patents

Abstract

The present invention relates to a configuration method for a data access system to access the data elements stored in a distributed data structure, wherein the system comprises a hierarchy of nodes that have communication links between them , the hierarchy extends from a root node to a plurality of end nodes, this plurality of end nodes providing the distributed data structure and where the routes through the hierarchy to specify data elements stored in the nodes of endpoints are identifiable by pointers stored in nodes along each route and also where a request for access to a specific data item triggers a search message that passes through the hierarchy from one end node to the root node until it reaches a node that has a pointer relevant to the specific data element, after which the message The search path passes along the associated path to the end node that contains the data element that it retrieves, a method that includes the steps of: a) detecting a failure of part of the hierarchy that affects communication between at least one respective descendant node and a respective ascending node of the hierarchy; b) establishing a first additional communication link between the descendant node and an additional node, then the additional node becomes a second respective ascending node with the respective descendant node; establishing a second additional communication link between the second ascending node and the first ascending node, d) instructing the first ascending node to send messages destined to the descendant node towards the second ascending node from which node they pass to the descendant node; second ascending node periodically updates the other nodes of the hierarchy as the new place if the next node is the hierarchy who

Description

UH DB METHOD CONFIGURATION FOR PN SYSTEM DB DB DATA MANAGEMENT D E S C R I P C I O N The present invention relates to a configuration method for a data management system, to access data in a distributed data structure.

The structure of distributed data involved can be provided by very different arrangements. For example, it can be provided by different nodes in a parallel processing computer. In a communications network on the other hand, it can be provided at different switching points in the network, for example in local exchanges of , r ~. a public network. In a particular example, the invention can be used to provide information routes for tracking calls to users who can change their position with respect to logical or geographical places in the network.

The services available by means of communication networks are becoming increasingly complex. it has become important to introduce management systems for the associated service characteristics, which may be selectively available to a user, as long as no excessive computing or communication headers are introduced. An example is known as personalized telephone numbering schemes. In this, each user or group of users must assign a phone number that remains unchanged when the user or users change their place in the network.

For a service such as personal numbering, it is clearly important that the location data can be easily changed. From the point of view of the network, the data of the place is volatile but should always be quickly accessible in its current version.

Another important aspect of a network service such as personal numbering is that it is equally important as it is scalable. That is, it needs to be able to expand to a much larger number of users, perhaps indefinitely. For example, it is equally important that a network system can grow to support a number of millions of users. It is also desirable that these users generate variable load. For example, users who arrive at the office may wish to transfer their personal number from their place in the house to their place in the office. This will generate several million transactions in a space of a few hours. In a fixed network where a completely mobile service is supported, equivalent to that of a cellular radio network, then the heading or initiation carried by the fixed network in place information at that moment can carry many millions of transactions every hour.

It is an opinion to maintain a central database, which retains the relevant data to a service such as personal numbering. Any traffic that needs to be re-routed to a destination on the network can then access the database to obtain a current location for the destination. However, a system of this type generates a significant header in which the network should carry traffic simply by accessing the central database, either to the current site data or to download a current version. A provision based on central control is clearly vulnerable to overload.

Another solution, used in cellular (mobile) telephone networks is to provide each user with a home recorder, known as a "home location recorder". Each home location recorder retains important user data, such as the type sometimes known in smart grid technology as a user profile. That is, it must indicate, for example, which services of that user are subscribed and give details such as daily information time in relation to services of the type "call diversity" that can be applied indifferently at different times of the day. It will also contain a location data for the user. The home registrars together effectively provide a distributed data structure and are selected so that there is a mapping between the user's usual locations and the position of their home records in the network. However, because all calls automatically go to the user's home register, there is still a significant start in dealing with the user ~~ * when he waits for his or her usual place. Additionally, changes in the user's location mean that the location data in the home recorder has to be updated, which helps to signal the load. In addition, there are problems with the size of the user's base growths. The design response in a cellular network to an increased user base is to decrease the local cell sizes. However, not only is the dramatic load growth signage, according to a square law, but there is also the complexity of the management procedures. Providing data integrity and robustness is significant.

A slightly more sophisticated version of the "house place registration" (HLR) provision has also already been proposed, in which an additional record is provided which is known as the "visited place recorder", this repeats the character of the HLR, but it is also located in the network, in order to process requested calls when the called user is waiting for his / her usual place. The incoming calls will refer to either the registrar in those circumstances, depending on whether the incoming call appears on the network. However, if a call request is - "" "on track to the HLR, it simply passes to the visited place registrar.

THE INVENTION According to the present invention, there is provided a configuration method for a data access system to access the data elements stored in a distributed data structure, wherein the system comprises a hierarchy of nodes having communication links among them, the hierarchy extends from a root node to a plurality of end nodes, this plurality of end nodes providing the distributed data structure and where the routes through the hierarchy to specify data elements stored in the end nodes are identifiable by pointers stored in nodes along each route and also where a request for access where a specific data element triggers a search message that passes through the hierarchy from an end node towards the root node until it reaches a node that has a pointer relevant to the specific data element, after This is where the search message passes along the associated path to the end node that contains the data element that it retrieves, a method that includes the steps of: a) Detect a failure of part of the hierarchy that affects the communication between at least one respective descendant node and a respective ascending node of the hierarchy. b) Establish a first additional communication link between the descendant node and an additional node, then the additional node becomes a second respective ascending node with the respective descendant node. c) Establish a second additional communication link between the second ascending node and the first ascending node. d) Instruct the first ascending node to send messages destined to the descendant node towards the second ascending node from which node they pass to the descendant node; Y e) The second ascending node periodically updates the other nodes of the hierarchy as the new place if the descendant node is the hierarchical node.

By establishing a second additional communication link and then periodically updating the hierarchy as the new place in the hierarchy of the descendant node, the signaling required to update the hierarchy can be spread over time or confined to periods when the "load" of the hierarchy is light The additional communication link supplies the requests to the descendant node from nodes that have not been updated with the new place of the descendant node.

Personal numbering has already been mentioned before. It is expected to be a key ingredient in future communications services. That system assigns a number or address to an individual user (or potentially to a group of users). Then, the traffic destined for that user or those users will be tracked to the person or persons whenever it is located in the network. Although there is a need for a "hardware" current address associated with the user or users, so that the call can be routed to its destination, the "hardware" address is simply a translation of the destination personal number and will change if the important user moves his place in the network.

In an embodiment of the present invention, the "hardware" address is represented by the data element that an end node may contain. In a personal numbering system, a user may want to redirect their incoming calls to a "hardware" address on the network that is served by a different local exchange. Then the user will update the end node in the new exchange location so that the address of the important "hardware" is associated with the user's personal number. Therefore, that end node will preferably update the associated pointers in the nodes of the hierarchy to signal a route to the new local exchange for calls to that user's personal number and instruct the previous end node where the user was located to cancel your "hardware" address for that user's personal number. This means that the user's location data has been moved and changed. However, a data access system according to an embodiment of the present invention can track the user's movement, the entrapped being fired in the local exchange and provides the updated version of the "hardware" address.

A data access system according to an embodiment of the present invention may be capable of automatically reconfiguring and / or reconstructing its "space" of information either upon request or in response to a failure. For example, when the embodiments of the present invention have a hierarchical nature a "flood-fill" search procedure can be extended rapidly throughout the entire data access system so that the omitted data can potentially be collected and reconstructed. in a reconfigured system, if necessary.

Embodiments of the present invention can provide a relatively simple solution to the problem of entrainment, for a service such as personal numbering where the entrant has to change to a detailed level in the network, in real time, even in a switched telephone network. public (PSTN). In a PSTN the development and changes can on the other hand have very significant routes that are difficult for personal numbering.

BRIEF DESCRIPTION OF THE FIGURES Figure 1 shows a simple 4-level binary tree implementation of the network management system.

Figure 2 shows a flowchart for a simple information retrieval procedure executed by nodes in the implementation of Figure 1.

Figure 3 shows a bridging of two network management systems according to the implementation of Figure 1.

Figure 4 shows a network management system according to an embodiment of the present invention that has been expanded.

Figure 5 shows an example of response to a failure within a network management system according to an implementation as shown in Figure 1.

Figure 6 shows an information layer containing multiple network management systems according to the implementation shown in Figure 1.

Figure 7 is an illustrative diagram used in an explanation of a recovery mechanism according to the invention.

Figure 8 is a flow diagram; Y Figure 9 is a block diagram of an intelligent network architecture in which the network management system of Figure 1 can be implemented.

DETAILED DESCRIPTION OF THE INVENTION In a personal numbering system in a communications network, the service characteristics required by the user are: 1) The ability to program a local terminal equipment, such as a telephone, to be the destination device of his telephone number personalized. 2) the ability to revert to a predetermined location in the network, such as a home telephone, as the destination device; and 3) the ability to also select on the network the destination device for your personalized telephone number. r This means that the network and the associated management systems have to be able to cover the calls that arrive on the route to a "mobile target". The basis of the design of network management systems according to the embodiments of the present invention is a logical separation of the information management task from that of network infrastructure, implemented in a signaling network that can control the from a distributed database while it is simply expandable. Once the system, which comprises the signaling network with the distributed database, has localized information, usually in personal numbering, this is a "hardware" destination address, it effectively handles retro-control to the network infrastructure or supporting network.

J - An advantage of this division between the support and management network is that it facilitates the incorporation of end user control systems and networks, such as mobile communications, internal businesses, computer networks and the like. In general, because the management network is a separate conceptual structure, it can be physically treated as a separate structure and can be built directly on an existing system. Therefore, the functionality of existing systems can be argued instead of those that replace it. In addition, several disparate networks can be extended, unifying them.

With reference to figure 1, the simple base from which the network management system operates is node 1, 2, 3. Nodes 1, 2 and 3 are connected hierarchically as a tree, (it will be kept in mind that the network in the form of a tree is a logical structure that can be implemented within a real physical network, but the physical network will not necessarily be in the form of a tree). All nodes 1, 2 and 3 have at least one ancestor, except for the simple node 2 that forms the "root of the tree". Root node 2 is special only in the sense that it has no ancestors. This structure, consisting of nodes 1, with only one ancestor, has advantages for its simplicity, since multiple ascendants require more sophisticated control logic to deal with the extra decisions that have to be made. However, multi-ancestor structures offer more robustness when facing physical and logical failures and the embodiments of the invention are not limited to use with simple ascending structures.

In general, nodes 1, 2 and 3 will have more than one descendant except the end nodes 3, which deal directly with the support network, these usually being localized in local exchanges. These end nodes 3 need not have any descendants and for simplicity the following description is assumed to have no descendants.

With reference to Figure 1, the general principle of the route network comprising nodes 1, 2 and 3 and links, is that each end node 3 may potentially contain (or may provide) a current hardware address for one or more users If an incoming call request is received to place a call to a destination user via the network, the route network is used to find the current "hardware" address important to the destination user. It can be done that because the current hardware "address" for the destination user can be located through one or more of the end nodes 3, the correct end node 3 will point in the route network by middle of back pointers in nodes 1 and 2 above in hierarchy. It does not matter in which end node 3 between a network call request, the request will simply go through the hierarchy until it finds the back pointers, which will then proceed to the correct end node 3. That is, they will pass through the network until they find a node 1, 2 that contains a pointer in the direction of the correct end node 3. That pointer will go to the next node in the back node and so on it goes back through the hierarchy to the end node 3 that can provide the current user's hardware "address".

All nodes 1, 2 and 3 can have the same processing facility and data structure. The data will be present in the data structure only if that particular node has a role to play in directing a call request to a destination user. That role can be: i) simply pass a call request to your ascending node. ii) offer a pointer to the destination, if the node is not at the destination, but is on the path to the destination and pass the call request to the node indicated by the pointer; or iii) providing the current user "hardware" address of the destination user, if the referred node is the end node 3 for the destination user.

The information may be present in the data structure of a node if the pointer comprises a "key" and a link identifier and in the case of the end nodes 3 a "hardware" current address. (Additional information that may be present, except at the root node, is the list of ancestor nodes for use in case of failure.) This is discussed further in the following, Alternatively, instead of the node list, a new one can be searched. ascendant).

Although only the end nodes 3 will always contain a current "hardware" address, all nodes will generally be able to contain one. This is because, as will be seen later in this description, if the route network expands or contracts, the end nodes 3 become intermediate nodes 2 and vice versa. The "key" in general will identify the important user with the pointer. For example, the key can be that user's personal number. The link identifier identifies which of the links down to the next layer of the route network will pass a call request.

It will be noted that a route network as described above clearly does not provide a "home location recorder" for any user and therefore avoids any additional signal load involved in call requests that go to or through a data logger. home.

Using the signaling network shown in Figure 1, a user such as "P" will normally be served by an end node 3, such as the nearest local exchange. This means that the address of the P current network hardware is stored in that local exchange. Then the user P instigates a change in the decision to move from being served by a node A to be served by a different node B. The user P instigates the change through the access node B, which modifies its data to indicate that the user P can be found directly from node B with no additional route. Then node B passes a change command to its ascending node D and consequently D changes its data so that instead of route calls destined to user P to node A then en route to node B. node D, which it acts as an ancestor to node A, informs node A that there are no longer routes for user P and that therefore any call for user P will pass to the ascending node D.

It can be considered that, in this stage, user P has in fact to come from user P '. The structure of the data on the network is practically unchanged, only the nodes A and B having been affected together with the ascending node D.

There is an instant when only node B is able to find user P. This is at the stage that node B modifies its data to accommodate user P, but ascending node D has not the consequent changes. However, this moment has a very short life.

The changes if the user P moves to a much more distant node, for example the node C in figure 1, are necessarily much easier to reach. The initial node C will change its database to accommodate the user P and instruct its ascendant, node H in figure 1 that the change has been . The change will propagate through the tree, node C instructs node H, node H instructs node G and node G instructs node F, the root of the tree. The node F then "cleans" the previously existing route to the user P informing that his descendant node E no longer has call routes for the user P to one of his descendants. In turn, the node E will say to the descendant node D and the node D will say to the descendant node A.

Although these changes are more substantial, in the worst case scenario the number of changes involved will only be as large as 21og2 (N), where N is the total number of nodes in the tree. Very important, the changes are local, being triggered by communications between adjacent nodes. This updates the highly efficient distributed route table as well as only requires technology that already exists.

With reference to Figure 2, if a call request is made to establish a connection to a destination user in a support network, the route network is triggered in order to locate the destination user. The route process will be the same in each node that receives the call request, in which the node will make one or more of a series of decisions and acts on the output of these decisions.

The route process in any single node that receives a call request can be seen in the flow chart 20 of Figure 2. The initial response of a node is to examine its data structure for information in relation to the destination user, STAGE 22. Then, the three possible scenarios will be as follows: 1) The referred node is neither the target user nor located in its path.

In this case, in STAGE 22, the node will find that there is no information present. So the call request goes to the root of the network, then the node will verify if it is a root node itself, STAGE 23. The normal situation will be after the node is not a root node and the request for call simply passes to the ascending node, STEP 24. Consequently, the call request will make its way up through the network of nodes to the root node 2. 2) The referred node is in the path of the node in the destination user, but is not in itself the local exchange for the destination user.

In this case, the node will find in STAGE 22 that it does not retain information about the destination user. Therefore it proceeds to make a verification to know if it is in itself the node in the destination user. That is, verify if there is a local exchange, STAGE 25. In this scenario, the response will be negative and the node will pass the call request to a descendant node indicated by the present information, STAGE 26. Therefore, the call request at that stage it has passed to the root node of the network and is sent down again to the destination user. (Although it will be noted that it is not always necessary to track a call request through the root node). 3) The node is the node adjacent to the destination user.

In this case, by means of STAGES 22 and 25, the node will find that it has information about the destination user and this is also a local exchange. Therefore, it will inform of the download to the support network that will establish a connection to be set. That is, it will download the caller's location from which the call request originated and its own local information. The route network can now ethe operation, then the support network will start the call in itself, STAGES 27 and 28.

There is an alternative result, not discussed in the preceding. This is where a node establishes that it has no information about the destination user, but that it is a root node 02, and therefore has nowhere to address the call request. This result indicates that information has been omitted and, instead of sending the call request, node 2 will instigate a data search; STAGE 29.

Systems as described above, based on a tree network, use the concept of parallelism. Each node • "" "". 1, 2 and 3 handles an updated database in response to changes initiated by its neighbors. The hierarchical provision forces a limitation on the need to change information. Although it is possible for a requested data to change to propagate through the system, the structure is designed so that a change has limited consequences. This level of parallelism is evident from the geographical decomposition, although exactly the same decomposition can be applied with a node 1, 2 and 3.

The important feature is that changes are possible within the distributed data structure that are treated locally. Most nodes 1, 2 and 3 do not get involved when updating a piece of information, which has the effect of distributing the load in terms of updates through the network. The consequence of this is that the signal information in a network continues to be distributed even when the system is larger. In the case of a binary network, as described with reference to FIG. 1, the signal load per node 1, 2 and 3 is independent of the number of nodes in the network.

Each node 1, 2 and 3 can be provided by a parallel computer of distributed memory. Each processor inside the computer can be mapped directly to one of the ,? - communication nodes 1, 2 and 3 described in the previous example. Consequently, the computer control system can handle the database in the same way that nodes 1, 2 and 3 operate. This means that each "leaf" of the tree, ie the nodes they represent at the local exchange level, ie the lowest nodes 3, break the problem in the same way as the higher nodes of the network and therefore they use the same control processors.

It is a particular advantage that each node can be used to potentially control a network of lower nodes, always using the same control process. Therefore, to expand the scope of a route network as described above, it is possible to create a network based on the same architectural principles, create a new root node 1 and "degrade" the root nodes of the network original and the network to be placed under the new root node.

With reference to Figure 3, the original network may have had a root node "R" and the new network may have a root node R. "The new root node R is then placed as a new layer, which superimposes the nodes that were the previous root nodes R 'and R "and the new root R acts as a root for two networks. Each of the old root nodes R 'and R "needs to be updated to record the fact that they are no longer root nodes, after which the system will function properly.The same principle can be used to expand an existing route network to comply with the development and demands that are distributed geographically non-uniform Therefore, a node 3 that previously represented a local exchange can be converted to a node 2, higher in the route network, simply by adding an additional network below This is shown in Figure 4, where a node 40, previously at the lowest level of a route network, has been converted to intermediate node paper above 6 additional nodes 41 and 42, with the node lowest represented the local exchange that is now two layers lower.

In each case, as soon as each node whose state is changed has been informed, the route network will effectively select itself after a change in configuration. • The way in which a link failure is detected can be either because a connected node has tried to use the link and has failed or because a network monitoring mechanism has picked up the failure.

An example of the latter is the "heartbeat" signal provided in conventional C7 signaling systems.

In the simplest situation and having a single link failure, for example between the end node 3 'and an intermediate node 2", then the end node 3' will no longer be able to respond to call requests which will normally pass to its node 2"ascendant on the other side of the link that failed. In the normal routine network structure, there is no alternative link from that end node 3 'to either the original ascending node 2' or to any ascending node.

In order to reconfigure the route network when addressing a link failure, each descendant node 2, 3 has a list of ascending nodes by means of which it may try to establish reconnection in the route network. The descendant node 3 'affected by the decomposed link will therefore issue a reconnection request to each potential upstream node on its view, in turn, until an ascending node that works with ignition capacity is located. After this, the reconnection of a route through the network to the descendant node 3 'will be placed in the same manner when the reconnection is established and when a user moves from one end node 3 to another end node.

In practice, this reconnection exercise can be complex, depending on how high the important descendant node is in the hierarchy, since it will have to issue reconnection requests with respect to all pointers it contains.

In Figure 5, a new link 50 will now be established. However, since the route network has a logically defined structure in which the links are identified for example by addresses to each node, there are no difficulties in creating the new link 50. even when 2 layers of the hierarchy are expanded.

In the meantime, it may be the case that a call request attached to the end node 3 'is put in the route network in another way and therefore needs to be re-routed either to or through the downstream node 3' down the link what's wrong If the descendant node 3 'has already triggered a reconnection exercise, using a new ancestor, then the network will function in a normal manner. However, a reconnect exercise can generate delay, depending on the number of routes needed to be re-established. If the exercise is not completed, then the incoming call request can reach the intermediate node 2, ascending to the end node 3, but is unable to reach the end node 3 'The intermediate node 2' can not instigate a new link 51 to reach a descendant node as effectively as a descendant node can instigate a new link 50 to reach a new ancestor.This is because a new descendant node will only be important for a selection of the pointers carried by the intermediate node 2 ' In this case it is more effective for node 2"intermediate to issue a search request" flood / fill "in order to locate the new ascendant 2 'of the descendant nodes.

A filled flood search request is simply issued from one node to each other node to which it is connected. Each node when it receives a flood / fill search request will verify in order to know if it contains important data, or an important pointer. If so, then that node will respond to the request. If it is not, it will pass over the fill / flood search request to each node to which it is connected, except the node from which it was received. (Flood / fill search requests can be limited in practice, for example, by limiting the number of times the request can pass).

Therefore the intermediate node 2 will now locate the new ascending node 2 'for the end node 3' and be able to re-establish all the important routes.

Consequently, the mechanisms described above generate a new link 50 in the network that effectively "passes" the fault.

Of course you can use the same mechanisms if it is a link or a node that has failed.

There is a second mechanism to reestablish the network after it fails. A flood / fill search request for specific information can simply be issued through the network. Adopting this solution, an extensive reconfiguration of the network can be achieved.

The use of the flood / fill search request technique causes significant forward signaling and therefore is undesirable for less robust networks, ie those that are more likely to fail.

When a node has established a link to a new ascendant there is a need for the place of the "new" nodes to pass to the new ascending node. This has to be done as soon as possible if requests are passed to the node without delay. The need to update the network causes a break in signal activity after a change of direction that follows the establishment of a new connection between the node and the new ancestor. This can cause problems with the ability of the network to track requests between nodes especially when there are a number of reconnections during a given time interval as will be the case with any real network.

The present invention arose in an attempt to overcome this problem.

Figure 7 illustrates the recovery method. In route network 70 there is a link failure in link 71 between nodes 72 and 73.

A first step in the recovery mechanism is a detection step 80, as shown in FIG. 8. This step is carried out by the node 73 and detects that the link 71 between it and its ascending node 72 has failed. The node 72 carries out this stage as well.

A second stage 81 is performed by the node 73 and this is the establishment of a new connection 74 to a new ancestor 75. The new ancestor is selected from a list of possible ancestors retained by the node 73.

The new ascending node 75 then establishes a direct link 76 to the old ancestor 72, in a step represented by the box 82.

The old ascendant 72 is then requested by the new ascendant 75 to update its address for the node 73 and all the requests you receive go through the link 76 direct to the new link 75 and new connection 74 to the node 73.

Therefore, in this step the link failure has been overcome and again the node 73 is part of the network 70.

In the next step, represented by frame 84 is for the new node 75 ancestor to periodically update the other nodes in the network with the new node node place 73. This procedure can be performed "on demand", that is, when a node is received. requesting a node for the node 73, then the new ancestor 75 may send an update location message to the sending node. This may depend on whether the request is relocated for call connection requests. Alternatively, it may be carried out according to a predetermined pattern or carried out according to the activity of the network, i.e. the updating procedure may be carried out during a period when there is not much signal activity on the network .

The network management system, or data access system, described above provides a relatively reliable data structure because the data structure is distributed. The information that is being accessed, in this case hardware addresses for network users, is not held in a simple specific place. Although, in the first instance, there is only one route through the system to a data element of that information, route failures can be easily handled, as described above. An additional feature that will increase the security of connections to a user, is a backup data storage point or a "phantom" end node, will retain the current hardware address for a user. This can be located in a flood / fill search exercise so that reconnection with that user can still be established.

When a flood / fill search request is issued, a relatively high level of signal traffic can be generated. It may be that flood / fill search requests are used only in the case of data corruption, or on request, in order to minimize their use. It can be noted that, in a tree hierarchy as described above, the path through the route network to a local exchange or end node 3, which is in direct contact with a target user, is always less than 21og2. (N) '1, still making this search procedure relatively efficient.

In addition to the search, flood / fill requests can remove any reference to the specified data about the passing nodes, which cleans the network of spurious routes or data trajectories, the only exception of this will be in the local exchange , or end node 3, when the information is found. In this case, a location registration procedure can be initiated so that the data (or the data path) can be reconstructed. In addition, a call construction request can be passed on to the support network to ensure that the initial request is addressed. In the worst case, this total process to connect the calls takes 4log2 (N) 'stages, precisely double that taken in the case of an integral network.

In this way damage can be reconstructed on a large scale although it is more efficient to use route initiation requests directly for reconstruction. A combination of both processes leads to a rapid recovery process with the data being formed immediately when needed and the data not required being repaired slowly when the network capacity allows.

Management systems of the type described are easily expandable. Even in the simplified case that was presented, this is the tree structure, the generalization to a mechanism to retrieve distributed information is an obvious next step. Even in an environment that is dynamically repositioned information and the introduction of multiple copies, a hierarchical control system provides the fastest access to the closest source of required information. This is true even if an initial node of the data is ignored to the point that it does not recognize the database key or location address.

"" 'This can be better understood in terms of the operation of the customized telephone number system described above, from the point of view of the distributed database. The fact that the location of the users is constantly monitored and updated makes the internal structure of the management network information that of an automated database. If you add to that fact that there is a physical distribution of the data source, this is the movement of the users, who physically distributes the data, then this minimizes the number of transactions that each node will have to make. The net result is that of a distributed data structure that can be accessed from any remote terminal as if it is a coherent database, while the data transactions are updated in a parallel manner.

By direct analogy, information such as that stored in a memory is often located in many geographic locations. To have access to that information, you must know the place where it is stored. In a static environment, this is a trivial fact. However, there may be a problem. If there is only one copy of each piece of information, then simultaneous multiple access to a given piece of data is going to the present problems. The obvious solution to this is to migrate the data and produce multiple copies of them. That is, a data that is popular can be stored in several geographically separate sites, which reduces the containment of network access and load. Again, if this is done once and for all, there is no need to rush about finding the place of any data source from any given place since it will always be in the same place. However, in the real world it is likely that changes that are made continuously will be needed.

It is well known that the popularity of a piece of information is a transient thing. it is directly related to the fact that the information is volatile. It is often of value only when it is new and not when it begins to get old. Due to this change in relative value, instead of the data seen from the perspective of any given terminal, it can change over time. The information to all the users of a system about these changes of place incurs a high cost in the load of the network. Therefore, a system is necessary within which the location of information can be altered without telling the users, as long as the users retain their access, this is precisely what the hierarchical control structure of an embodiment of the present invention it can do and is consequently extremely versatile.

The "key" to a database can be through literally a key. It is simply a structure by which information is referenced. Similarly, references at the end of an article can be considered as keys. There is no necessary understanding in itself but it can be taken differently to be deciphered or at least partly deciphered. In the example of an article with references, a memory can be consulted and the memory will be directed to a place in the article in memory or directly to the correct type of memory.

There is an intrinsic hierarchy even at this level of key types, it is possible to superimpose different control hierarchies on a common source of distributed data.

With reference to Figure 6, at the end node level 60 of three different control hierarchies will be a common distributed data source. The three control hierarchies 61, 62 and 63 are of similar structure, each having a root node in the "top" and having access to the array of end nodes in the common source of distributed data, the different hierarchies of control can be easily connected to the degree that they can make comparisons and - 'share management information. In figure 6, some of the descendant nodes, in this case end nodes, only contain one type of data store where others contain 3, but this does not mean that this "simple descendant" can not access data from all of them. the other three types.

There is no reason why this scenario can not advance in that of general database structures.

For example, it is very possible to use this system to integrate different databases, the physical distribution presents the same data handling difficulties as that of algorithmic distribution, that is, where the algorithmic control of the data is separated into different structures. This will usually be a reference to two different database programs such as Sybase and Oracle. These programs already share a common interface language in that they respond to the same form of data requests from a remote terminal, the data request is first structured in its common language by means of an application program and the data received is transformed into the format relevant. Each database program can be considered as a physically separate or even physically separate data store. To unify the two systems in a hierarchical control structure as shown in Figure 6, all that is needed is a simple interface program that converts requests from the hierarchical control structure into relevant database requests. This allows hierarchical control structures to be applied to the database systems that exist in a very simple way.

The hierarchical route structures according to the embodiments of the present invention are capable of offering a general service in which they allow the recovery of data from any part of the hierarchical route network without necessarily knowing the place of the data source, this functionality it can be viewed from the perspective of a user, for example simply as a service structure that can be accessed at any point. It can be treated as a system to which the user can issue requests from any physical point and the system will solve the request for the user.

For a terminal user there will be a direct connection in the hierarchical control network. A user terminal can request requests in the system and those requests take the form of questions about the data that is contained in a data space of the route network. In the specific example of personal numbering, this information is the current hardware address of the user identified by a dialed telephone number. In practice, the route network will not normally return the current hardware address but instead proceeds to trigger the automatic connection of the call from the location of the current hardware address that has been found. This can be applied in other contexts. That is, any application in the route network can have a predetermined response once a piece of information is found. An answer may be to return information to the questioner although it may be simpler in relation to this return of the information as a separate task, current transport of the information between the two relevant addresses that are being handled by a conventional transport process, such as a support network.

- »The task of the route network is simply to find the location of the information. A different mechanism can be used to process the information once it has been found.

As is evident, this behavior is independent of the place of a user's terminal. As a result, the route network can be viewed as a local non-information database. More important, this database does not have to save the data in a specific place. If another site is found to be a more optimal storage location, the information can be moved without affecting any of the user's terminals or data access keys that users can use to retrieve the information.

With reference to Figure 9 and considering the personal numerical application of a route network according to an embodiment of the present invention, as described above, the route network will typically be provided at the service control point 90 of a intelligent network architecture. In an intelligent network architecture, a connection will be established again by a service switching point 91, using the route information obtained from an associated service control point 90. Alternatively, the intelligent network architectures will be highly flexible, the route network will replace a function of the service control point 90, and it will be located differently in the intelligent network architecture. The hardware that will provide the functions of a route network already known and used in intelligent network architectures. For example, the relevant hardware can be provided by a Unix workstation such as the Sun Sparc 5 that has a hard drive with a Gigabyte capacity.

Important features available for embodiments of the present invention may be listed as follows: 1) Optimization of geographic demand, for example by adding layers to give depth in highly populated regions in a personal numerical service network. 2) In memory problems, efficient location of demand can be provided, for example in services that are literally memories or services such as video on demand that is clearly a type of memory (library). 3) The signaling load is kept low and can truly be maintained in such a way that the signaling load at any point is constant even through the network when it has been fully expanded. This aspect is because when the demand increases, the network can generate local extra layers. 4) Embodiments of the invention can rest on conventional technology and software already in use and therefore does not require development of additional platform.

) Robustness. Although not described above, robustness can be improved by increasing the number of ascending nodes for a single ascending node. This is indicated in Figure 6. Another aspect of robustness is that the route network will be self-healing. 6) Because new layers can be introduced at the top of a route network according to an embodiment of the present invention, the existing route networks can be linked by a common root node superimposed in order to convert the original root nodes of the different networks to descendant nodes. This means that networks can be united which extend in different national territories. 7) Route networks of this type allow the introduction of new services since the nodes do not need to be updated with respect to a new service, the nodes simply provide a route and do not have knowledge of the data for which the route information is generated. In conventional networks the nodes include number translation resources that will be updated when new information is introduced into the network. These requirements are reduced or eliminated in networks operating according to the invention.

It will be noted that, although the hierarchical structures described above are generally binary tree structures, it is not necessary to be binary or tree-shaped. The structure should be considered, for example, as a three-dimensional structure instead of a two-dimensional structure and / or with multiple links instead of a binary solution. However, a significant aspect is the hierarchy in which the nodes of the "higher" layers are few and retain more information than the nodes of the "lower" layers.

The placement of a used structure can be used to fulfill the tasks you need to perform.

Although the embodiments referred to above with a personal numerical service have been described in connection with a fixed network, they can also be applied to a network that is an integration of a fixed cellular network and a mobile network.

Actually, realizations may be important for cellular networks alone.

Claims (10)

R E I V I N D I C A C I O N S

1. - A configuration method for a data access system to access the data elements stored in a distributed data structure, where the system comprises a hierarchy of nodes that have communication links between them, the hierarchy extends from a root node to a plurality of end nodes, this plurality of end nodes providing the distributed data structure and wherein the routes through the hierarchy to specify data elements stored at the end nodes are identifiable by pointers stored in nodes along each route and also where a request for access to a specific data item triggers a search message that passes through the hierarchy from an end node to the root node until it reaches a node that has a pointer relevant to the specific data element, after which the search message passes along the associated path to the end node containing the data element it retrieves, which method comprises the steps of: a) detecting a failure of part of the hierarchy affecting the communication between at least one respective descendant node and a respective ascending node of the hierarchy; b) establishing a first additional communication link between the descendant node and an additional node, then the additional node becomes a second respective ascending node with the respective descendant node; c) establishing a second additional communication link between the second ascending node and the first ascending node; d) instructing the first ascending node to send messages destined to the descendant node to the second ascending node from which node they pass to the descendant node; and e) the second ascending node periodically updates the other nodes of the hierarchy as the new place if the descendant node is the hierarchical node.

2. - Method according to clause l, wherein the data elements each comprise an address that identifies a data storage location in a data structure.

3. - Method according to clause 1, wherein the data elements each comprise an address that identifies a place in a communications network.

4. - Method according to clause 3, wherein receiving the search message in the end node containing the data element triggers the transmission of the data element to call control means in the communication network in such a way that it can be established a connection in the communications network in the identified place.

5. - Method according to any of the preceding clauses, wherein each pointer or pointer comprises an identifier associated with a data feed stored in the relevant specific end node, together with a link indicator to indicate the following communication link on the route to the end node.

6. - Method according to clause 1, wherein receiving the search message in the end node containing the data element triggers the download of the data element.

1. - Method according to any of the preceding clauses, wherein means are provided for loading a data element at any end node and pointer update means are provided which are responsible for loading a data element to trigger the consequent updating of the data elements. pointers stored in the nodes of the hierarchy.

8. - A management method for a communications network, which method comprises a data access system with a recovery method according to any of the preceding clauses.

9. - A method of handling according to the clause 8, for use in providing a personal numbering service in the communications network.

10. - A handling method according to clause 9, wherein each pointer along a route identifies a personal number for a specific network user and where the data element located at the end node reached by a search message that passed along that route comprises a hardware address for the user of the specific network in the communications network. SUMMARY Data elements stored in a distributed data structure are accessible by means of a hierarchical route network in which routes through the network are signaled to individual data elements. The network comprises communication links between nodes (1, 2, 3) that extend from a "root node" (2) to a plurality of end nodes (3). The end nodes (3) contain the data elements. To find a data element, a search message that enters the network through an end node (3) passes through the network to the root node (2) until it finds a signalized route to the important data element . After this it passes along the route to the end node (3) that contains the important data element. The invention is important for personal numbering services in a communications network. In this case, each of the data elements comprises hardware addresses for network users. If a user moves in relation to the network, it will change its hardware address and, in many cases, it will also change the important end node (3). However, the signaled route changes accordingly and the route network therefore provides an automatic entrainment of the user.