FR2925722A1 - Computer network, has archiving server whose file collection agent does not transfers data of source server in collected file, if search result represents presence of data corresponding to calculated code in mass memory of archiving server - Google Patents

Abstract

The network has a source server (2) e.g. computer system, with a file collection agent (8), and an archiving server (5) comprising an archiving management module (9) for permitting archiving of files on a mass memory (3) e.g. magnetic hard disc. A communication connection system (1) exchanges information between the servers. The agent does not transfers data of the source server contained in a collected file and corresponds to a calculated digital code, if search result carried out by the module represents presence of data corresponding to the calculated digital code in the mass memory. The communication connection system is constituted by a wired connection system or radiofrequency connection system.

Description

The invention relates to a computer network with advanced functions of automatic file archiving.

Throughout the text, the term files refers to files as stored in mass memory and likely to be managed as such by at least one operating system and file management (OS) device. A file is designated by a unique identifier (file name) allowing access to the data of this file stored in mass memory. A file includes a set of useful data for at least one computing device and / or at least one user, constituting a content of this file. Unless otherwise specified, the term data used in reference to a file designates content data of a file. The archiving of files essentially aims at storing the file content data in mass memory. Moreover, in a manner known per se, certain data, called metadata, are descriptive files, some of its properties and / or all or part of its content. The file metadata can be part of the contents of the file, all or part of the contents of another file, or dynamically generated by a software process-especially an operating system-from the file. In general, the invention relates to a computer network comprising: at least one file server, said source server, equipped with at least one file collection agent to be archived, at least one archiving server adapted to enable the archiving of files on at least one mass memory associated with this archiving server, and at least one network digital link adapted to allow the exchange of digital information between the servers. It is already known (see for example US 2004/0236801) such computer networks in which: the archive server is provided with at least one archive management agent adapted to be able to issue a collection trigger command to destination of a collection agent of a source server, this trigger command incorporating data representative of predetermined collection criteria, - said collection agent is adapted to execute on the source server a file search corresponding to the collection criteria of the collection server. a command received from the archive management agent. These computer networks, and known archival systems they include, have several series of disadvantages. First, they require the establishment of a high-performance digital link, that is, high bandwidth, between each source server and the archive server for file transfer. However, the development of modern computer networks involves optimizing the bandwidth of the digital links as much as possible, by allocating this bandwidth as much as possible to the transmission of useful data to the different users of the network. In particular, it should be noted that these known systems require many data transfers which, in reality, are not strictly necessary for archiving files. Indeed, for example, the same set of data representing a file content can be duplicated on different source servers, or even on the same source server. In this case, the corresponding files are systematically archived by known systems regardless of their context. It should also be noted that file archiving is essentially considered, from the point of view of a user, as a secondary administration task of the computer system. Moreover, in the same way, it is important to ensure that the collection agent occupies only a minimum of the computing resources of the source server, so as not to disrupt the operation, and to leave the maximum of these resources available to the user of this source server. Also, modern computer networks are increasingly heterogeneous, and allow the communication of a variety of workstations as well as the format of the files contained, the operating systems and file management. The file archiving system on the network must therefore be compatible with these different formats and operating systems. However, most known archiving systems are closely dependent on the operating system and file management, and are not compatible with other file management systems. For example, known archiving systems that are compatible with the Windows operating system are generally not compatible with the Linux or Unix operating system. This problem arises even more with the development of the files associated with metadata, which are data describing the contents of the file and / or computer properties of the file and / or other information associated with the files and useful to the user and / or software applications that may work on this file. As such, the inventors have determined that a successful network archiving system should be compatible with: the different categories of metadata that may be encountered, and the different formats in which this metadata may be recorded, any system of information provided with at least one file server, independently of the software applications that manage and process the information and content of the data (e.g., electronic document management applications or documents), unstructured data in association with metadata collected and / or enriched by the archiving server, in particular to subsequently provide the possibility of performing contextual searches on the archived information and / or presenting the restored information in context function. The invention aims to propose a solution to these problems. Accordingly, the invention aims in particular to provide a computer network with an archiving system which, on the one hand, is compatible with a low bandwidth in the digital link between each archiving server and each source server , and / or which, in any case, occupies only a small part of this bandwidth; avoids the multiplication of identical duplicate data set transfers on different source files, but allows the archiving of metadata associated with each of these source files; on the other hand, uses only a very small portion of the computing resources of each source server and is extremely undisturbed by the operation of each source server; moreover, it is compatible with computer architectures and very diverse software applications; allows the acquisition and / or creation and / or association, with archived data and files, of various metadata, including metadata for contextual management of archived information; enables dynamic and automatic, scalable management of each archive server. To do this, the invention relates to a computer network 20 comprising: at least one file server, said source server, with a file collection agent to archive, at least one archive server adapted to allow storing files on at least one mass memory associated with this archive server, and having an archive management module, and at least one network digital link adapted to allow the exchange of information between the servers, wherein: said archive management module is adapted to be able to issue a collection command to at least one collection agent of at least one source server, this collection command incorporating data representative of predetermined collection criteria -including in particular data identifying at least one start time of the collection and / or at least one time window of realization of the collection by said collection agent, said collection agent being adapted to execute on the source server a file search corresponding to the collection criteria of a command received by said collection agent and sent by the archive management agent, characterized in for each file collected by said collection agent, the latter is adapted to: calculate, from data contained in the collected file and stored on the source server (that is to say on at least one memory mass of the latter), and according to a predetermined single calculation method, a numerical code, said calculated code, comprising at least a number, said identifier code, whose value depends solely on said data and said calculation method and is clean to said data, to transmit the calculated code to the archiving management module, in that the archiving server is adapted to be able, on receipt of a calculated code, to find out if it exists, in memory of mass of the archive server, data corresponding to this calculated code-in particular at least one record containing this calculated code and / or the identifier code-, in that it is adapted to perform a transfer to the archive server said data from the source server contained in the collected file and corresponding to the calculated code, if the result of the search performed by the archive management module is representative of the fact that data corresponding to the calculated code are not present in the memory of mass of the archive server, said archive management module being adapted to store in mass memory the data contained in the collected file and the calculated code associated with these data, and in that it is adapted not to transfer to the archiving server said data of the source server contained in the collected file and corresponding to the calculated code, if the result of the search carried out by the The archive management odule is representative of the fact that data corresponding to the calculated code are present in mass memory of the archive server. Advantageously and according to the invention, said archiving management module is adapted to be able to transmit to the collection agent digital data representative of the result of said search of data stored in mass memory of the archive server and corresponding to the code calculated, and in that said collection agent is adapted to: where the result of the search performed by the archive management module is representative of the fact that data corresponding to the calculated code are not present in the mass memory of the archiving server, generating a transfer authorization command to the archiving server of the data of the source server contained in the file collected, said archiving management module being adapted to memorize in mass memory that data contained in the file collected and this calculated code associated with these data, if the result of the search carried out by the archiving management module is representative of the fact that data corresponding to the calculated code are present in mass memory of the archive server, do not generate an authorization command transfer of data contained in the file collected. Furthermore, advantageously and according to the invention, the archiving management module is adapted to carry out said search by searching whether there exists, in mass memory of the archive server, at least one record containing the identifier code. More particularly, advantageously, a network according to the invention is suitable for: if the result of the search carried out by the archiving management module is representative of the fact that the identifier code is not present in the mass memory of the server d archiving, generating a transfer command to the archiving server of the data of the source server contained in the file collected, said archiving management module being adapted to memorize in mass memory this data contained in the file collected and this identifier code associated with these data, if the result of the search performed by the archive management module is representative of the fact that the identifier code is present in the mass memory of the archive server, do not generate a data transfer command contained in the collected file. Advantageously and according to the invention, the calculated code, and in particular the identifier code, is calculated from the collected file data other than the metadata of this collected file. In particular, the value of the identifier code depends only on the data of the collected file other than the metadata of this collected file, and is representative only of the content data of the collected file, but not the metadata of this collected file. In addition, advantageously and according to the invention, for each file collected, the file collection agent is adapted to collect metadata associated with this file and, if the result of the search performed by the archiving management module is representative of the fact that data corresponding to the calculated code are present in the mass memory of the archiving server, generating a partial transfer authorization command adapted to allow only the transfer of all or part of these metadata to the management module of archiving, said archiving management module being adapted to memorize in mass memory these metadata in association with the data of the file collected. Thus, in a network according to the invention, the metadata associated with a collected file are transmitted irrespective of the result of said search carried out by the archiving management module with regard to the calculated code. It should also be noted that the value of the calculated code, and in particular the value of the identifier code, does not depend on the metadata associated with the file collected, these metadata, whether or not contained in the file collected, being not used in the calculation method of the calculated code, in particular in the method for calculating the identifier code. Furthermore, advantageously and according to the invention, for each file collected, the file collection agent is adapted to collect and associate with this collected file, metadata resulting from data and / or metadata recorded on the source server concerned. In particular, it is metadata collected on the source server other than technical metadata. Advantageously and according to the invention, for each file collected, the file collection agent is adapted to collect and associate with this collected file, metadata chosen from: - MDT technical metadata describing all or part of the technical content of the file itself and associated with this file; metadata developed by a separate software application and recorded according to a predetermined structure, for example header metadata and / or metadata stored in a database and / or metadata recorded in a separate file (eg metadata about a style or format ...); metadata generated according to at least one predetermined computer processing by the collection agent, for example: contextual metadata generated from data representing the position of the file collected in the source server file management system, and / or the computer environment of this file collected in the source server file management system, and / or attributes of the collected file and / or, • static metadata corresponding to one or more values entered by the user, for example during of the configuration of the collection, • computed metadata obtained by a simple calculation from the data of the collected file, for example by a simple indexing calculation. The use of such metadata in a network according to the invention, recorded in the archive server in association with the data of a file collected provides many advantages. In particular, these metadata can be the subject of subsequent analyzes, cross-checks and logical deductions. More generally, the archiving thus achieved becomes dynamic in the sense that it is likely to evolve over time. The data and metadata stored in the archive server can be indexed, immediately after each record, or offline. Indexing methods may be added at any time, as needed, in an archiving server according to the invention. In addition, it is possible to use tools for optimizing the recording of data, and in particular for optimizing the resources of the archive server mass memories, according to data lifecycles and / or or associated metadata, data storage plans (based on, for example, data retention times, migrations and format conversions, in particular according to software evolutions, data destruction strategies, etc.). In a network according to the invention, it is even possible to fully control the archive server using the different metadata associated with the archived files. It should be noted in this respect that this control of the archiving server can be performed automatically by a software application according to the metadata, without any intervention of an operator. In addition, it is possible to integrate advanced search engines using different metadata and / or indexing results to facilitate subsequent search and / or contextual presentation of archived data.

Moreover, advantageously and according to the invention, for each file collected and transferred to the archive server with at least one calculated code, said archive management module is adapted: to calculate a verification code, from the data contained in the collected file as received by the archive server 9 and according to said predetermined single calculation method used by the collection agent to calculate said corresponding identifier code, - to compare the value of this verification code calculated from received data, with the value of this identifier code transmitted with said data of the collected file, and for memorizing in mass memory the data contained in the collected file (preferably with the calculated code, or at least at least the code identifier) if and only if both values of the verification code and the identifier code are equal.

Thus, said calculated code, and more particularly the identifier code, is used to check the integrity of the data of each file collected after each transfer to the archiving server. Also, this same calculated code, and more particularly the identifier code, can be used to check the integrity of the data transferred in the event of a file being restored, that is to say from the archiving server to the archive. a source server. Furthermore, according to an advantageous embodiment, the archiving management module includes a collection management module adapted to centralize the communications between the archiving management module and each collection agent.

Preferably, the collection management module comprises a file transfer module adapted to control the transfer of data from each file collected between a source server and the archive server. It should be noted that in a network according to the invention, the data of the same file can be stored in mass memory of a single source server and then transferred entirely in traditional point-to-point mode, that is, for example, by FTP transfer, or on the contrary duplicated on various source servers (for example formed of user workstations). In the latter case, the calculated code makes it possible to avoid transferring the same file several times to the archiving server from several source servers.

In addition, advantageously and according to the invention, the file transfer module is adapted to limit the number of simultaneous transfers to the archive server and / or the data transfer rate to a predetermined maximum threshold value.

The invention also relates to a network characterized in combination by all or some of the characteristics mentioned above or below. Other objects, features and advantages of the invention will appear on reading the following description of its preferred embodiments, given solely by way of non-illustrative example and which refers to the appended figures in which: FIG. FIG. 2 is a schematic flow chart showing an embodiment of the main steps performed by a collection agent in a network according to the invention for initializing a collection. FIG. 3 is a block diagram illustrating an example of a dialogue between a collection agent and an archiving management module of a network according to the invention for the initialization of a collection, FIG. 4 is a flowchart. schematic representation of an embodiment 20 of the main steps performed by a collection agent in a network according to the invention during a collection step, the f Fig. 5 is a block diagram illustrating an example of a dialogue between a collection agent and an archive management module of a network according to the invention in a collection step, Fig. 6 is a diagram of a example of a file illustrating its constitution (data and metadata) and the metadata that can be associated with the data of this file in the archive server of a network according to the invention, FIG. 7 is a schematic flowchart representing an exemplary embodiment of FIG. steps performed by a code verification module calculated in a network according to the invention. The general organization of a computer network according to the invention is shown schematically in FIG. 1. The network firstly comprises a system 1 of communication links adapted to allow the exchange of data and digital information, that is to say say, forming a digital network. This link system 1 may have any architecture, incorporate or not specific point-to-point links, access to one or more public network (s) such as the Internet, access to one or more networks (x) private computer (s) ... Moreover, the invention is compatible with any technology (s) for the realization of such a system 1 of links (wired, radiofrequency, satellite, via a mobile phone network ...). At least one file server, said source server 2 is connected to this system 1 of links. Such a source server 2 is a computer system comprising at least one central unit equipped with a microprocessor-based digital data processor (s) and associated random memory (s), and at least one data processing system. OS operation, that is to say a set of software adapted to perform administration and management functions of the computer system, and in particular file management functions recorded on at least one mass memory 3, 4 associated (by the operating system OS) to this source server 2. Such mass memory may be a mass memory 3 connected locally to the source server 2, (for example one or more hard disk (s) ) magnetic (s) and / or electronic (s) and / or optical (s), internal (s) or external (s), or other). The mass memory associated with the source server 2 may comprise a mass memory 4 of network type, that is to say directly connected to the link system 1, the source server 2 accessing this mass memory 4 via the system 1 of links. Such a network memory 4 may also include one or more magnetic hard disk (s) and / or electronic (s) and / or optical (s), and / or one or more unit (s) combining several hard disks (for example RAID type) or any other mass memory 4 capable of being connected autonomously to a network. Of course, several separate mass memories 3, 4 may be associated with the same source server 2. A source server 2 within the meaning of the present invention may consist of any known computer system, not only of the personal microcomputer type, fixed, portable or portable, but also for example of the type personal digital assistant, portable digital phone, or otherwise type of computing center or other. Furthermore, the network according to the invention also comprises at least one archive server 5 connected to said link system 1.

Such an archiving server 5 is a computer system comprising at least one central unit equipped with digital processing capacity with microprocessor (s) and associated random memory (s), and at least one operating system (not shown in FIG. 1), that is to say a set of software adapted to perform administration and management functions of the computer system, and in particular file management functions recorded on the minus a mass memory 6, 7 associated (by the operating system) with this archiving server 5. Such a mass memory may be a mass memory 6 connected locally to the archive server 5, (for example one or more hard disk (s) magnetic (s) and / or electronic (s) and / or optical (s), internal (s) or external (s), or other). The mass memory associated with the archive server 5 can comprise a mass memory 7 of network type, that is to say directly connected to the link system 1, the archive server 5 accessing this mass memory 7 via the system 1 of links. Such a network mass memory 7 may also include one or more magnetic hard disk (s) and / or electronic (s) and / or optical (s), and / or one or more unit (s) combining several hard disks (for example RAID type) or any other mass memory 7 may be connected autonomously to a network. Of course, several separate mass memories 6, 7 can be associated with the same archive server 5.

Each source server 2 comprises a file collection agent 8 to be archived, which is adapted to execute on the source server 2 a search of files corresponding to collection criteria of a collection command received by this collection agent 8 via the System 1 links from the archive server 5. The archive server 5 is meanwhile has an archive management module 9 which is adapted to allow the definition and registration of a collection command incorporating data representative of predetermined collection criteria, and for issuing such a collection command on the system 1 of links to at least one collection agent 8 of at least one source server 2. The management module Archiving 9 of the archiving server 5 includes a collection management module 10 which is the only point of access to the archiving server 5 by the collection agents 8 of the different source servers 2. In addition, Collection management dule 10 is adapted to allow a user to define a collection command, i.e. to enter predetermined collection criteria. For example, the collection criteria of a collection order can be chosen from: - the source server (s) 2 concerned by the collection, - the date and time of the first launch of the collection command. collection on each source server 2, at each time period (time window defined by a start date and time and an end date and time, these dates and times being definitively definable or, on the contrary, relatively report to predetermined events such as the ignition of a server, the shutdown of a server, a periodicity ...) during which the collection is authorized to run, or criteria relating to all or part of the path of the files involved in the collection, - criteria for the inclusion and / or exclusion of the files in the collection, in particular according to technical metadata (file type, file size, author, owner, date of creation, date of modification ion, date of last access, ...) and / or file header metadata and / or indexing data ...

The collection management module 10 of the archive management module 9 is also adapted to send to each source server 2 concerned on the system 1 of links, a triggering command of the corresponding collection. The triggering of a collection by a collection agent 8 10 can be controlled by the collection management module 9 according to various criteria which are for example chosen from: - temporal properties (fixed date, daily, weekly, monthly, annual periodicity , etc.), external trigger commands such as manual operations by an administrator, or a command from a specific software application of the archive server 5 or external to this archive server 5 ( via the system 1 of links) ... Each collection agent 8 is previously configured so that the address of the archiving server 5 associated with it is known to it (this address being recorded on the source server 2 ). Thus, in a network according to the invention, the collection agents 8 of the different source servers 2 are not autonomous, but on the contrary entirely under the control of the archiving management module 9 of an archive server. In other words, the operation of the archive management module 9 and 25 of the collection agents 8 is of the master / slave type, the archive management module 9 constituting a master able to control the operation of the collection agents 8, and the collection agents 8 having a slave operation remotely controlled via the link system 1 by the archiving management module 9. This type of operation appears in particular in FIGS. 2 and 3. After a start-up step 20 followed of a collection initialization step 21, the collection agent 8 is placed on a stop test step 22 in which it is examined whether the collection agent 8 must be stopped or not, whether or not it has received a stop command signal from the archive management module 9. If yes, a stopping step 23 of the collection agent 8 is executed. If not, two processes are run in parallel. In a first process 24, 25 of active collections configuration, the first step 24 examines whether a message 33 is received from the archive management module 9. This message 33 can be a stop signal of a collection in progress and / or deactivation of an already active collection, an activation signal of a new collection or a modification signal of an active collection. If such a message 33 is received, then a collection configuration step 25 is performed, in which data representative of the collection definition 34 is received from the archive server 5 if this step corresponds to a new collection to be configured or an active collection change. Then, the parameters of the corresponding active collection are created or modified. Then, collection agent 8 returns to step 22. If no message 33 is received in step 24, step 25 is not performed. The second parallel collection execution process 26, 27 makes it possible to execute the various active collections configured by the first process 24, 25. This process 26, 27 for executing the collections comprises a first step in which it is examined whether an active collection 25 should be started or not. For example, in step 26, a predetermined collection start time recorded in the collection definition data 34 can be compared to the date and time of the current system. If the collection start time is reached, this collection is started in step 27. Otherwise, the collection agent 8 returns to the stop test step 22. Thus, in a network according to the invention, the collection configuration and collection execution processes execute concurrently. In this way, several collections can be activated in parallel, as well as the configuration of new collections. Once activated, a collection remains active unless this collection has been scheduled to be executed only once, or if the archive management module 9 emits a signal of interruption of this collection. As seen in FIG. 3, when the collection management module 10 of the archive management module 9 determines that a collection command must be sent to a collection agent 8 on a source server 2, it issues a connection request 31 on the network, that is to say to the source server 2 via the system 1 links. In the event that a connection can be established, the collection agent 8 transmits an acknowledgment signal 32 validating the establishment of the connection, to the archiving server 5.

Simultaneously, the step of starting the collection agent 8 is performed as mentioned above. The collection management module 10 of the archiving management module 9 then sends a command 33 for triggering and identification of collection, making it possible to trigger a new collection by the collection agent 8 as mentioned above. , this command 33 being processed by the test 24. The collection management module 10 of the archiving management module 9 of the archive server 5 then sends the source server 2 all of the data representative of the definition of the collection. 34 to perform, which allow the collection 25 to be configured by the collection agent 8 as described above. When the collection is complete, the collection agent 8 sends an end of collection signal 35 to the collection management module 10 of the archive server 5. FIG. 4 illustrates the main steps performed by the collection agent 8 at the end of the collection period. During a collection step 27. FIG. 5 illustrates the exchanges occurring between the collection agent 8 and the archiving management module 9 of the archive server 5 during this collection step 27. The step 40 corresponds to the start of the collection. During the subsequent test 41, the collection agent 8 executes a collection module 13 which traverses the file system of the source server 2 as the background task of this source server 2. If no new file to be archived is detected, a stop signal of the collection is issued in step 42, which allows to control the test 29 described above triggering the end of the collection. If a new file to be archived is detected, the test 43 examines whether this file is eligible for collection considering the collection parameters received in the collection definition 34. If the file is not eligible for this collection, the process returns in test 41 to detect a new file. If the file is eligible, a step 44 is executed in which a numerical code, said calculated code, is calculated from the data contained in the file as stored on the corresponding mass memory 3 and / or 4 of the source server 2 This calculated code comprises at least one number, said identifier code, whose value depends solely on these data and a predetermined calculation method executed by a calculation module 12 of the collection agent 8. The value of the identifier code does not depend in particular on the metadata that may be associated with the file collected, the data of the file corresponding to its content, that is to say the data other than metadata, being extracted from the file to proceed with the calculation of the identifier code, or more generally to the calculation of each number corresponding to the calculated code. The calculation method used in this step 44 may be arbitrary, since it makes it possible to identify in a one-to-one way the content of the data, called content data, of the file that are not metadata.

In other words, if the value of the content data varies, the calculated code, and in particular the identifier code, must vary. On the contrary, if all the content data remain the same, two successive calculations performed with the same calculation method must provide the same value for the calculated code, and in particular for the identifier code.

In an exemplary embodiment, the calculated code comprises: a numerical value indicating the total size of the content data stored in mass memory; an intermediate code which is the result of a first verification calculation carried out on the content data, for example of the simple sum calculation type or CRC (cyclic redundancy verification calculation); an identification code that is the result of a complete one-to-one calculation of the content data, for example a more elaborate cyclic redundancy check calculation (CRC-8, ..., CRC-16, CRC-32 , CRC64-ISO, CRC64-ECMA-182) and / or a calculation based on at least one cryptographic hashing function (SHA-0, SHA-1, SHA-2, SHA-256, SHA-384, MD2, MD4 , MD5, RIPEND, Tiger, Whirlpool) applied to the content data. Preferably, all the source servers 2 whose collection agent 8 performs the archiving of the files in the same archive server 5 use the same method of calculating the calculated code.

The code calculated in step 44 is sent to the archiving server 5 at 53. Prior to this, the collection agent 8 sends a connection request signal 51 to the archive management module 9, and the latter issues a signal 52 of acknowledgment of connection validation if this connection can be established. The archiving management module 9 then triggers, upon receipt of this calculated code, a module 11 for checking the presence of this calculated code stored on one or the other of its associated mass memories 6, 7. FIG. 7 gives an exemplary embodiment of the verification module I1. In step 71, it is examined whether the size of the data (and which is the first number contained in the computed code in the example mentioned above) corresponds to a content data size already archived in mass memory. 6, 7 of the archive server 5. If the size of the content data of the collected file is new, it means that this content data has never been archived. In this case, a signal 54 is recorded in step 77 according to which the content data of the file collected has not already been archived, which then makes it possible to proceed to step 48 of transferring these content data. 9. On the other hand, if there already exists content data corresponding to this size stored in mass memory 6, 7, it is examined, during a test step 73, among the various data items. of content corresponding to this size, if there exists archived content data stored in mass memory 6, 7 of the archive server 5 for which the value of the intermediate code obtained from the first verification calculation (which is a simple calculation , for example of the CRC type), is the same as that contained in the calculated code. This intermediate code value can be stored in mass memory 6, 7 of the archive server 5 in association with the corresponding content data. In this case, it suffices in a step 72 immediately preceding the test step 73, to read this intermediate code value as it is recorded and to compare it with the value received in the calculated code received corresponding to the content data. of the collected file.

If this intermediate code value is not recorded with the content data, it is sufficient, in step 72, to calculate it using, of course, the same calculation method that was used to calculate the value of the content. intermediate code contained in the received calculated code corresponding to the content data of the calculated file. If the value of the intermediate code contained in the calculated code received does not correspond to any of the content data stored in mass memory 6, 7 of the archiving server 5, this means that the content data of the file collected have never been archived, and recorded in step 77 a signal 54 according to which the content data of the file collected have not already been archived, which then proceeds to step 48 of transferring these content data. If, on the other hand, there are archived content data for which the value of the intermediate code corresponds to that of the received computation code, the value of the identification code for these various archived content data is determined during step 74. This determination can be made either by simple reading when the value of the identification code is recorded in association with the content data in the mass memory 6, 7 of the archiving server 5, or by a new calculation performed on the data. content from the same calculation method as that used to compute the identification code of the content data of the collected file. Then, in step 75, the value of the identification code contained in the received calculated code corresponding to the file collected is compared with the value of the identification code thus determined for the various archived content data. If there is archived content data that has the same identification code value, a signal is recorded in step 76 that the file content data is already present in the server's mass memory 6, 7. archiving. Otherwise, the transfer step 48 is performed. The verification module 11 therefore performs a search on the mass memories 6, 7 to determine if there exists an archived file for which such calculated code having this value has already been recorded. The result of this verification is then communicated in 54 or 55 to the collection agent 8 of the corresponding source server 2 via the link system 1. The collection agent 8 then performs a test 47 from the result it receives. If this result 54 indicates that the calculated code is not already present on one of the mass memories 6, 7 of the archive server 5, the collection agent 8 authorizes the transfer of the content data of this collected file. to the archiving server 5, during the transfer step 48. On the other hand, if the result 55 indicates that the calculated code is already present on one of the mass memories 6, 7 of the archiving server 5, step 48 is not performed and the process is put back to the test 41 for detecting a new file. In parallel, after the step 44 of calculating the calculated code, the collection agent 8 performs a step 45 of creation of metadata associated with the file collected, then a step 46 of systematic transfer of these metadata to the archive server 5. At the end of step 46, the process is returned to test 41 for the detection of a new file to be collected. Thus, regardless of the value of the calculated code, the metadata of a collected file is always transferred to the archiving server 5. If the module 11 for verifying the computed code determines that the content data of the file collected have already been archived on the archive server 5, only these metadata are transferred. They are then added to all the information concerning the file collected and already previously recorded on one or other of the mass memories 6, 7 of the archiving server 5. On the other hand, if the module 11 of verification of the calculated code indicates that the content data of the collected file have not been previously archived or have been modified since the last archiving (computation code not present in the mass memories 6, 7 associated with the archiving server 5), the data content of the file collected and stored on the mass memories 3, 4 associated with the source server 2, are transferred with the metadata to be archived. In step 45 of creating the metadata associated with the collected file, the collection agent 8 executes a module 14 for extracting and calculating metadata. The metadata collected by this module 14 includes MDT technical metadata that are stored in association with the file content data on the mass memory 3, 4 associated with the source server 2. These MDT metadata are generated by the operating system. OS of the source server 2, for example metadata describing the technical content of the file itself, its date of creation, its volume (number of bits), ... Moreover, the file can also be associated with one or more applications (s) specific software (s) 19 likely (s) to generate MDE header metadata in association with this file, these MDE metadata being recorded in this file, with the file content data, on the mass memory 3, 4 of the source server 2. For example, a word processor file, which can be identified as such by its extension, can be the subject of generated MDE metadata and recorded by a word processing software that allows you to create or modify it (for example a title, a summary, keywords, the identification of an author, ...). It can also be for example data representative of the state of a document, access rights to a file, a digital signature, authentication certificates ... This header metadata MDE are accessed by opening the file itself, or through a header extractor API software module. The metadata extraction and calculation module 14 of the collection agent 8 is adapted to be able to identify these MDE header metadata, to collect them (by sending targeted queries, for example on databases, or a direct request, for example via a header extractor API, concerning the collected file) and save them with all the metadata associated with the file that will be transferred to the archiving server 5. In addition, the module 14 of extraction and calculation of metadata is adapted to be able to generate MDG metadata itself according to at least predetermined computer processing, and to associate them with the file collected, these metadata MDG being for example metadata representative of the context, on the source server 2, the collected file. For example, the MDG metadata generated by the metadata extraction and calculation module 14 are chosen from: metadata, called MDGS static metadata, which can be defined by a user input during the configuration of the collection for example via the archive management module 9; for example, it is possible to add metadata to qualify all the files corresponding to a predetermined collection or to a predetermined collection part (for example one may want to associate the commercial qualification to all the files of all the corresponding directories to a predetermined collection); metadata, called MDGC calculated metadata, obtained by a calculation, preferably a simple calculation, performed on the content data of the file concerned: for example, in the case of a text file, it is possible to perform a light text indexing providing statistical results on the words of the text; in the case of an image file, it is possible to indicate whether or not the content data of the file corresponds to a black image, or to a blank image, or others ... FIG. 6 illustrates an example of a file, and different categories of metadata that can be extracted and calculated from this file. In the example, the file 61 is for example a file in the MXF format. This file includes content data D1, D2, D3. It also contains metadata that requires the opening of the file 61 or a metadata extractor to access it, namely MDE header metadata: MDE1, MDE2, MDE3, intermediate metadata MDI4, MDI5 separating the groups of data content, and end-of-file MDF metadata. Moreover, the operating system of the source server 2 on which the file 61 is initially recorded has associated technical metadata MDT with this file 61. The module 14 for extracting and calculating metadata first gathers the different metadata. MDE header, MDT techniques, and the intermediate MDI4, MDI5, and MDF end metadata. Moreover, a metadata creation module 62 makes it possible, by one or more simple calculations, to generate computed MDGC metadata that are also associated with the metadata of the file 61 by the module 14 for extracting and calculating metadata. Also, the metadata extraction and calculation module 14 associates static type MDGS metadata entered by a user during collection configuration to qualify the properties of the collected files. Also, the metadata extraction and calculation module 14 can associate with the file 61 metadata, called MDC contextual metadata. These contextual MDC metadata are generated by a module 63 for calculating contextual metadata from the technical metadata MDT and / or for example all or part of the header metadata MDE1, MDE2. These contextual metadata MDC are for example metadata representing the position of the file collected in the file management system of the source server 2 and / or the computer environment of this file collected in the file management system of the source server 2 and / or attributes of the collected file ...

Furthermore, the metadata extraction and calculation module 14 can extract metadata, called MDEX extracted metadata, from another file 64, separate from the file 61 but associated with the latter as a metadata file. Indeed, some file formats provide for specific metadata files, the contents of this specific metadata file constituting metadata for the content of a second file. This is the case, for example, of XML files, or of text editor files associated with style sheets. As can be seen, the collection agent 8 makes it possible to group all these metadata in a set of metadata. associated with the collected file. Thus, the collection agent 8 makes it possible to associate a large number of different metadata with each file collected, and then to transfer these metadata 65 to the archiving server 5, so that these metadata 65 are archived in association with the calculated code. and so with the archived file. This then makes it possible to obtain an archive server managed and controlled from metadata in a flexible, dynamic and fast manner. It should be noted that all or some of the processes mentioned in FIG. 6 can be carried out at the level of the archiving management module 9, in real time during the recording of the archived file or, on the contrary, at a later time, since the The corresponding metadata will have been previously transferred to the archiving server 5. In FIG. 1, there is shown schematically MD 1 metadata of a collected file 61 contained in this same file 61 in association with the corresponding DATA data (ie that may include MDT technical metadata and / or MDE header metadata and / or MDI4, MDI5 and / or end-of-file MDF metadata); MD2 metadata stored in the mass memory 3 of the source server 2 but outside the file 61 (for example may include header metadata MDE ', metadata extracted MDEX, and end-of-file metadata MDF' d ' a separate metadata file 64), and MD3 metadata generated by the archive management module 9 and / or by the collection agent 8 (e.g. including contextual metadata MDC, and calculated metadata MDGC, and metadata MDGS entries). The archive management module 9 comprises a file transfer module 15 adapted to control the transfer of all the content data and the metadata of the file collected between each source server 2 and the archive server 5. This module In particular, the file transfer function has the function of controlling the different simultaneous transfers to the same archive server 5, for example according to an FTP type protocol, in particular to limit the number thereof, as well as the data transfer rate. from each source server 2, limiting the transfer rate to a predetermined maximum threshold value, in particular to prevent the excessive occupation of the bandwidth of the link by the archiving function. The content and metadata data transfers can be carried out by the file transfer module 15 according to other types of protocols than the FTP protocol, for example according to protocols specific to the source server 2 concerned and / or to the archive server. 5. In addition, the archive management module 9 includes a metadata association module 16 adapted to gather the content data and the metadata of a collected file (in the case where a file collected containing content data has been transferred to the archive server 5) and to record new metadata corresponding to an already archived file in association with this already archived file. In the case where it receives an archived collected file comprising both content data and metadata, the archiving management module 9 activates a module 17 for validating the calculated code-in particular the identifier code-transmitted prior to the data of content of the collected file. This validation module 17 applies the same calculation method as the module 12 for calculating this code by each source server 2. The archive management module 9 knowing the identity of the source server 2 from which the collected file originates, also knows the calculation method used to obtain the calculated code - in particular the identifier code -. The module 17 for validating the calculated code is adapted to apply this same calculation method to the content data it receives. If the result of the computation performed by the validation module 17 is different from the calculated code-in particular of the identifier code transmitted previously by the source server 2 with regard to the content data of this collected file, an error message is sent and the content data is not archived. On the other hand, if the result of the calculation performed by the validation module 17 corresponds to the same calculated code - in particular to the same identifier code - as that received in association with the content data of the file collected, these content data can be archived, c ie, stored on one of the mass memories 6, 7 associated with the archive server 5. In this case, the MDT, MDE, MDG metadata also received in association with this collected file are related to the data of corresponding content and the calculated code, and the set is recorded as represented in FIG. 1 in mass memory 6, 7. In the example represented in FIG. 1, the calculated code is recorded in a field 18 following the data of the contents of the archived file. Thus, the calculated code is used in a network according to the invention to perform an integrity check of the content data transferred to the archiving server 5. It should be noted that contrary to what is shown schematically in FIG. 1, the metadata MDT, MDE, MDG, the content data and the calculated code are not necessarily saved together in the same location of a mass memory 6, 7 or in the same file. These different information and data can be distributed in different locations of the mass memory, for example recorded in different tables of an archive database. In the case where the archiving management module 9 only receives MDT and / or MDE and / or MDG metadata corresponding to a file collected, the metadata association module 16 records this metadata received in association with the data of the metadata. content of the file collected already previously archived in mass storage memory 6, 7. It should be noted that a network according to the invention is also compatible with the archiving of files whose metadata and / or data are duplicated on different source servers 2 themselves connected via the link system 1, for example by a peer-to-peer communication mode. In this case, the archiving of the same file from each source server 2 on which to associate, with the file data recorded on the archiving server, associated metadata generated by each collection agent 8, including contextual metadata to know the different source servers on which the file is saved. Moreover, an archiving server 5 according to the invention may be provided with indexing features of the content data and / or metadata and / or archived files, so as to facilitate subsequent searches and restorations. The indexing can take place before the recording of the corresponding information in mass storage memory 6, 7 for the new files collected, or on the contrary can be carried out retrospectively on the content data and / or metadata corresponding to files already archives.

Furthermore, in a network according to the invention, the archive management module 9 is adapted to be able to recover any interrupted file collection at a given moment for any reason. Indeed, when restarting the collection, the initialization step 21 of the collection agent 8 replaces this collection agent 8 to the previous initial state. The collection then restarts to the next file not previously processed. In addition, if the interruption of the collection occurs during a file transfer, the module 17 for validating the calculated code of the archiving server 5 will indicate that this file has not been correctly transferred, so that the Content data from these files will be transferred again. The invention can be the subject of many variants of different embodiments and applications. In particular, various software modules may be incorporated, without limitation, for extracting or managing metadata, or for indexing. Moreover, it applies as well regardless of the meaning that the user may want to give to the concept of content data and that of metadata. In any case, the invention makes it possible to archive archive servers in bulk memory only once the data considered as content data in the application in question and defined as such by the fact that the method for calculating the calculated code, and more particularly the identifier code, applies only to these data. Thus, nothing prevents us from predicting that only part of the data of each file collected is treated as content data.

Claims (5)

1 / - Computer network comprising: at least one file server, said source server (2), provided with an agent (8) for file collection to be archived, at least one archive server (5) adapted for allow the archiving of files on at least one mass memory (6, 7) associated with this archiving server, and provided with an archiving management module (9), and at least one digital link (1). ) network adapted to allow the exchange of digital information between the servers, wherein: said archive management module (9) is adapted to be able to issue a collection command to at least one agent (8) for collecting at least one source server (2), this collection command incorporating data representative of predetermined collection criteria, said collection agent (8) being adapted to execute on the source server (2) a search for corresponding files the criteria for collecting an order received by the said ag collection and issued by the archive management module, characterized in that, for each file collected by said collection agent (8), the latter is adapted to: calculate, from data contained in the file collected and stored on the source server (2), and according to a predetermined single calculation method, a numerical code, said calculated code, comprising at least one number, said identifier code, the value of which depends solely on said data and said method of calculation and is specific to said data, transmitting the calculated code to the archiving management module, in that the archiving server (5) is adapted to be able, on receiving a computed code, to find out if it exists, in memory of the archiving server, data corresponding to this calculated code, in that it is adapted to perform a transfer to the archiving server (5) of said data of the source server contained in the file collected and corresponds to nt to the calculated code, if the result of the search performed by the archiving management module (9) is representative of the fact that data corresponding to the calculated code are not present in mass memory of the archive server, said module management system being adapted to memorize in mass memory this data contained in the collected file and the calculated code associated with these data, and in that it is adapted not to transfer to the archiving server (5) said data from the source server contained in the collected file and corresponding to the calculated code, if the result of the search carried out by the archiving management module (9) is representative of the fact that data corresponding to the calculated code are present in mass memory the archive server. 2 / - Network according to claim 1, characterized in that said archive management module (9) is adapted to be able to transmit to the agent (8) for collecting digital data representative of the result of said search for data stored in memory. mass storage of the archiving server and corresponding to the calculated code, and in that the collection agent is adapted to: if the result of the search carried out by the archive management module is representative of the fact that corresponding data to the computed code are not present in the mass memory of the archiving server, generate a transfer authorization command to the archiving server of the data of the source server contained in the collected file, the said archiving management module being adapted to memorize in mass memory this data contained in the collected file and this calculated code associated with these data, if the result of the search carried out by the modu the archiving management is representative of the fact that data corresponding to the calculated code are present in mass memory of the archive server, do not generate an authorization command for transfer of data contained in the file collected. 3 / - Network according to one of claims 1 or 2, characterized in that the archive management module (9) is adapted to perform said search by searching if there is in mass storage of the archive server at least one record containing the identifier code. 4 / - Network according to claim 3, characterized in that it is adapted for: if the result of the search performed by the archive management module (9) is representative of the fact that the identifier code is not present in mass memory of the archiving server, generating a transfer command to the archiving server of the data of the source server contained in the file collected, said archiving management module being adapted to memorize in mass memory the data contained in the file collected and the identifier code associated with these data, if the result of the search carried out by the archiving management module (9) is representative of the fact that the identifier code is present in the mass memory of the archive server, do not generate a command to transfer the data contained in the collected file. 5 / -Network according to one of claims 1 to 4, characterized in that the calculated code is calculated from the collected file data other than the metadata of the collected file, and in that, for each file collected, the agent (8) for collecting files is adapted to collect metadata associated with this file and, if the result of the search carried out by the archiving management module (9) is representative of the fact that data corresponding to the computed code are present in the mass memory of the archiving server, generating a partial transfer authorization command adapted to allow only the transfer of all or part of these metadata to the archiving management module, said archiving management module being adapted to memorize in mass memory these metadata in association with the data of the collected file. 326 / - Network according to one of claims 1 to 5, characterized in that, for each file collected, the file collection agent (8) is adapted to collect and associate with this collected file, data metadata resulting from and / or metadata stored on the source server (2) concerned. 7 / - Network according to one of claims 1 to 6, characterized in that, for each file collected, the file collection agent (8) is adapted to collect and associate with this collected file, metadata chosen from: technical metadata (MDT) describing all or part of the technical content of the file itself and associated with this file; metadata (MDE) developed by a separate software application and recorded according to a predetermined structure; metadata (MDG, MDC) generated according to at least one computer processing predetermined by the collection agent. 8 / - Network according to one of claims 1 to 7, characterized in that for each file collected and transferred to the archive server (5) with at least one calculated code, said archive management module (9) is adapted: to calculate a verification code, from the data contained in the collected file as received by the archive server and according to said predetermined single calculation method used by the collection agent to calculate said corresponding identifier code, to compare the value of this verification code calculated from the received data with the value of this identifier code transmitted with said data of the collected file, and to store in mass memory the data contained in the collected file if and only if these two values of the verification code and the identifier code are equal. 9 / - Network according to one of claims 1 to 8, characterized in that the archive management module (9) comprises a collection management module (10) adapted to centralize communications between the archive management module (9) and each collection agent (8). 10 / - Network according to claim 9, characterized in that the collection management module (10) comprises a file transfer module adapted to control the transfer of data from each file collected between a source server (2) and the server archiving (5). 11 / - Network according to claim 10, characterized in that the file transfer module is adapted to limit the number of simultaneous transfers to the archive server (5) and / or the data transfer rate to a threshold value predetermined maximum.