(51) Int Cl.: G06F 11/14 ( ) G06F 17/30 ( )

Transcription

1 (19) TEPZZ_97799B_T (11) EP B1 (12) EUROPEAN PATENT SPECIFICATION (4) Date of publication and mention of the grant of the patent: Bulletin /19 (1) Int Cl.: G06F 11/14 (06.01) G06F 17/ (06.01) (21) Application number: 08.8 (22) Date of filing: (4) Implementing read/write, multi-versioned file system on top of backup data Einführung von Lese-/Schreib-Dateisystemen mehrerer Versionen am Anfang von Backup-Daten Implementation d un système de fichiers multi-versions en écriture/lecture pour des données de sauvegarde (84) Designated Contracting States: DE GB () Priority: US (43) Date of publication of application: Bulletin 08/ (73) Proprietor: SYMANTEC CORPORATION Cupertino, California 9014 (US) (72) Inventors: Wu, Weibao Vadnais Heights, MN 127 (US) Zeis, Mike Minneapolis, MN 4 (US) Hartnett, Tom St. Paul, MN 8 (US) (74) Representative: Meldrum, David James D Young & Co LLP 1 Holborn London EC1N 2DY (GB) (6) References cited: FR-A US-A US-A US-A EP B1 Note: Within nine months of the publication of the mention of the grant of the European patent in the European Patent Bulletin, any person may give notice to the European Patent Office of opposition to that patent, in accordance with the Implementing Regulations. Notice of opposition shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention). Printed by Jouve, 7001 PARIS (FR)

2 1 EP B1 2 Description BACKGROUND OF THE INVENTION Field of the Invention [0001] This invention relates to computer systems and, more particularly, to backup and restoration of data within computer systems. Description of the Related Art [0002] There is an increasing need for organizations to protect data that resides on a variety of client devices via some type of backup mechanism. For example, numerous client devices may be coupled to a network to which one or more media servers are also coupled. The media servers may include or be further coupled to a storage unit consisting of one or more disk storage devices, tape drives, or other backup media. A backup agent on each client device may convey data files to the media server for storage according to a variety of schedules, policies, etc. For example, large backup datasets may be moved from a client device to a media server configured to store data for later retrieval, thereby protecting data from loss due to user error, system failure, outages, and disasters, etc. as well as archiving information for regulatory compliance, workflow tracking, etc. [0003] In order to make data more readily available, one or more metadata managers may store metadata in a catalog, the metadata describing the associated backup data. Such a catalog may be referred to as a metabase. Metadata managers may be located on separate hosts or co-located on hosts that include backup data. Accordingly, one or more metabases hosted in a variety of locations may contain data describing backup data. [0004] Client devices may backup and restore data on conventional backup media as well as on disk-based backup media. Conventional archival storage techniques where entire datasets are stored, whether or not earlier identical versions have been stored, may be used. In other cases, single-instance storage techniques may be used in which datasets are segmented and the segments are de-duplicated. Conventional and single-instance techniques may be mixed in the same backup environment. To facilitate finding data entities that are stored in backup systems, file system user interfaces may be provided on client devices. For example, files in a backup may be read through interfaces that provide some of the capabilities of Network File System (NFS) or Common Internet File System (CIFS). Unfortunately, these interfaces do not provide the ability to write to, modify, or create multiple versions of files in a backup. [000] In view of the above, an effective system and method for writing and multi-versioning files in a backup stored on backup media that accounts for these issues is desired. [0006] US0/8486 describes, in a back-up storage system, an apparatus and methods for mounting a data volume corresponding to a back-up data set to a host computer. In one example, a method includes mounting a data volume on a host computer, the data volume comprising at least one data file, the data file corresponding to a most recently backed-up version of the at least one data file stored on a backup storage system, and storing, on the backup storage system, data corresponding to a second version of the at least one data file that is more recent than the most recently backed-up version of the at least one data file stored on the backup storage system while preserving the most recently backed-up version of the at least one data file. This document does not however teach or suggest, at least, maintaining a backup catalog configured to store metadata describing a plurality of backed up files, wherein the metadata includes a separate entry for each of the backed up files, and wherein the entry for each respective file includes a filename that corresponds to the respective file and a pointer to the respective data object that references the one or more data segments of the respective file. [0007] US04/ describes a hash-optimized backup system and method takes data blocks and generates a probabilistically unique digital fingerprint of the content of each data block using a substantially collisionfree algorithm. The process compares the generated fingerprint to a database of stored fingerprints and, if the generated fingerprint matches a stored fingerprint, the data block is determined to already have been backed up, and therefore does not need to be backed up again. Only if the generated fingerprint does not match a stored fingerprint is the data block backed up, at which point the generated fingerprint is added to the database of stored fingerprints. Because the algorithm is substantially collision-free, there is no need to compare actual data content if there is a hash-value match. The process can also be used to audit software license compliance, inventory software, and detect computer-file tampering such as viruses and malware. [0008] FR describes a method for saving digital data on a plurality of computer equipment connected to a computer network, characterized in that it does not employ centralized computer server, and comprises: - a prior step of calculation and transmission of the load of the equipment to other network devices, the step being performed by the equipment itself, - a step of saving said distributed data, the selection and distribution of the data being carried out by said equipment, so that the charges relating to the data are distributed in an automated manner and to achieve a charge balance equipment. SUMMARY OF THE INVENTION [0009] The invention is defined in the claims [00] Various embodiments of a computer system and methods are disclosed. In one embodiment, a computer system includes a first host coupled to a backup 2

3 3 EP B1 4 medium and a second host. The backup medium is configured to store data segments and data objects, each data object referencing one or more of the data segments. The second host is configured to convey a write request to the first host identifying a data entity, wherein the data entity comprises data segments referenced by a first data object. The first host is configured to restore a copy of the data entity and create a second data object referencing the data segments that were referenced by the first data object in response to receiving the write request. The second host is further configured to make modifications to the copy, and convey data changes to the first host corresponding to the modifications. The first host is further configured to add one or more data segment references to the second data object in response to receiving a data change that adds data to the copy or deletes a portion of a data segment from the copy. The first host is further configured to remove a data segment reference from the second data object if a data change deletes all of a data segment from the copy. In one embodiment, the backup medium comprises single instance backup data. [0011] In a further embodiment, the computer system includes a catalog configured to store metadata describing the plurality of backup data objects. In a still further embodiment, the host is configured to access the backup medium through a virtual file system interface. The virtual file system interface is configured to enable the second host to write data objects to the backup medium. The virtual file system interface may be further configured to enable the second host to create and store multiple versions of backup data objects on the backup medium. The catalog may be further configured to store metadata describing branches containing multiple version of the plurality of backup data objects. [0012] These and other embodiments will become apparent upon consideration of the following description and accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS [0013] FIG. 1 illustrates one embodiment of a computer system. FIG. 2 is a generalized block diagram of one embodiment of a backup system that may operate within a computer system. FIG. 3 is a more detailed block diagram of one embodiment of a system that may be a portion of a backup system. FIG. 4 illustrates one embodiment of a catalog and its associated data in a storage pool. FIG. illustrates one embodiment of a user interface that may be used by backup agent to view backup data and metadata. FIG. 6 illustrates one embodiment of modifications made during a write operation to data and metadata corresponding to a data entity that is stored in backup storage pool. FIG. 7 illustrates one embodiment of a process for modifying a file in a backup storage pool. FIG. 8 illustrates one embodiment of catalog metadata corresponding to multiple version of a file. [0014] While the invention is susceptible to various modifications and alternative forms, specific embodiments are shown by way of example in the drawings and are herein described in detail. It should be understood, however, that drawings and detailed description thereto are not intended to limit the invention to the particular form disclosed, but on the contrary, the invention is to cover all modifications, falling within the scope of the present invention as defined by the appended claims. DETAILED DESCRIPTION [00] FIG. 1 illustrates one embodiment of a computer system 0. As shown, system 0 includes hosts 1A- 1D and mobile hosts 1A-1D interconnected through a network that includes a local area network (LAN) 1 coupled to a wide area network WAN/Internet 1 and a modem bank 0, which is in turn coupled to a public switched telephone network (PSTN) 160. Hosts 1A-1D are representative of any number of stationary computers. Mobile hosts 1A-1D are representative of any number of mobile client computing devices such as laptops, handheld computers, etc. Both hosts and mobile hosts may operate as peers in a peer-to-peer configuration or as clients and servers in a client/server configuration. [0016] In alternative embodiments, the number and type of hosts, LANs, WANs, and modem banks is not limited to those shown in FIG. 1. Almost any number and combination of server, desktop, and mobile hosts may be interconnected in system 0 via various combinations of modem banks, direct LAN connections, wireless connections, WAN links, etc. Also, at various times one or more hosts may operate offline. In addition, during operation, individual host connection types may change as mobile users travel from place to place connecting, disconnecting, and reconnecting to system 0. [0017] Within system 0, it may be desired to protect data associated with any of hosts 1A-1D and mobile hosts 1A-1D. In order to protect host-associated data, various backup components may operate on hosts 1A-1D and mobile hosts 1A-1D. Turning now to FIG. 2, a generalized block diagram of one embodiment of a backup system 0 that may operate within 3

4 EP B system 0 is shown. System 0 includes hosts 2, 2, 2, 2, and 270 coupled through network 280 to a media server 20 and a metabase server 260. Backup agents 2, 22, and 23 operate on hosts 2, 2, and 2, respectively. In the illustrated embodiment, media server 20 is further coupled to a storage pool 2 in which a copy of data from one or more hosts may be stored. In one embodiment, storage pool 2 may be part of or included in backup server 20. Hosts 2 and 270 may include backup data 24 and backup data 27 respectively, in which a copy of data from one or more hosts may be also be stored. In one embodiment, backup data 24 and/or 27 may be single-instance storage pools. Metabase server 260 includes a catalog 26 on which may be stored metadata describing the data stored in backup data 24 and 27 and in backup medium 2. Additional metabases may be included in system 0 in alternative embodiments. Additional backup data may also be included in system 0, depending on the storage requirements of the system. [0018] During operation, backup agents 2, 22, and 23 may perform data backups. For example, in one embodiment data may be conveyed to one or more backup data locations and associated metadata conveyed to one or more metabases. Backup frequency may depend on a variety of factors including the urgency of data protection, storage pool capacity, network connection state, and enterprise policies. In one embodiment, backups may be done according to a schedule or at other times determined by administrative policy, security policy, or to meet other requirements of an enterprise. [0019] In order to minimize the size of backup data, single-instance storage techniques may be employed. In a single-instance storage pool, data is stored in segments, with each segment having an identifier or fingerprint that may be used to unambiguously identify it. For example, a data file may be segmented, and a fingerprint calculated for each segment. Duplicate copies of data segments are replaced by a single instance of the segment and a set of references to the segment, one for each copy. To retrieve a backup file, a set of fingerprints corresponding to the file s segments may be used as keys to locate the desired segments. Once the segments have been retrieved, they may be used to re-assemble the desired file. [00] FIG. 3 is a more detailed block diagram of one embodiment of a system 0 that may be a portion of system 0. System 0 includes hosts 2 and 2 and metabase server 260. Host 2 includes a local data store 3, applications 3, a virtual file system 3, and a backup agent 2. Host 2 includes backup data 24. Metabase server 260 includes a catalog 26 in which metadata describing the data stored in backup data 24 may be stored. Applications 3 may be any of a variety of conventional applications such as word processing programs, spreadsheets, browsers, etc. Backup agent 2 may backup data from local data store 3 by sending data directly to backup data 24 and associated metadata to catalog 26 using conventional backup techniques. Backup agent 2 may restore data to local data store 3 by retrieving data directly from backup data 24 and associated metadata from catalog 26 using conventional restoration techniques. In addition, applications 3 may perform read and write operations on backup data through an interface provided by virtual file system 3. In one embodiment, virtual file system 3 may provide a CIFS/NFS interface. A variety of alternative interfaces may be provided such as a WebDav interface, pseudo devices interface, etc. To read backup data, virtual file system 3 may present a view of catalog 26, allowing a user or an application to select data files for retrieval, retrieve the files from storage pool 24, and store the retrieved files in local data store 3. Virtual file system 3may allow a user or an application to write backup data, including writing multiple versions of data files, as though writing to a conventional file system. Before describing the processes through which these actions may be performed, a description of the organization of data and metadata stored in backup data 24 and catalog 26 will first be given. [0021] Turning now to FIG. 4, one embodiment of catalog 26 and its associated data in backup data 24 is shown. Catalog 26 may include a database, tables, or other similar data structures. In the illustrated embodiment, catalog 26 includes a table of entries. Each entry includes a filename, a set of attributes, and a data object pointer. More specifically, a first entry includes filename 411A, attributes 411B, and data object pointer 411C, and may be referred to hereinafter as entry 411. Similar filenames, attributes, and data object pointers are shown for entries [0022] A file, as used herein, may comprise one or more files or other data structures. In one embodiment, filenames such as filename 411 A may consist of a userreadable string. However, since multiple files may have the same name, a data object pointer is also stored in each entry to be used as a key to find a corresponding data object in backup data 24. In one embodiment, each data object pointer may be an unambiguous identifier such as a fingerprint. A fingerprint, as used herein, refers to a function of a data entity such as a hash function. In one embodiment, the fingerprints may be encrypted. More particularly, a fingerprint may comprise a Message- Digest algorithm (MD) or other hash function. Alternative hash functions include Secure Hash Algorithm (SHA), a checksum, signature data, and any other suitable function, cryptographic, or otherwise, for identifying a data entity. Copies of data entities such as files or file segments may be identified by comparing a fingerprint of one entity to the fingerprint of another entity. If the fingerprints match, then the two entities are copies of each other. In addition to the filename and data object pointer, additional attributes such as attributes 411B may be included in a file s metadata. Attributes may include a variety of information describing the associated data such as one or more of: a data size, batch number, type, 4

5 7 EP B1 8 version number, ownership, permissions, creation time, error code, etc. Other forms of metadata and/or identifiers will be apparent to those of ordinary skill in the art. [0023] In the illustrated embodiment, backup data 24 includes a data object table and a set of data segments. The data object table includes entries, each of which includes a data object ID and a set of associated data segment pointers. More specifically, a first entry includes data object ID 4 and data segment pointers Additional entries having data object IDs 4, 4,, and 460 are shown. Data object ID 4 is associated with data segment pointers , data object ID 4 is associated with data segment pointer 441, data object ID is associated with data segment pointers 41-44, and data object ID 460 is associated with data segment pointers In one embodiment, each data object ID may be an unambiguous identifier such as a fingerprint. In a further embodiment, each data segment pointer may be an unambiguous identifier such as a fingerprint. Backup data 24 also includes data segments and associated data segment IDs. For example, data segment ID 470A is associated with data segment 470B, 471 A with 471B, etc. In one embodiment, each data segment ID may be an unambiguous identifier such as a fingerprint. In a further embodiment, backup data 24 may include single-instance data objects and single-instance data segments, i.e., both data objects and data segments may be de-duplicated. [0024] The relationships among the data object pointers of catalog 26 and the data object IDs, the data segment pointers, and the data segment IDs of backup data 24 may be arranged to permit files and their data segments to be identified, read, written and versioned. More specifically, as shown via connecting arrows in FIG. 4, data object pointer 411C points to data object ID 4, 412C and 416C point to data object ID 4, 413C and 4C point to data object ID, 414C points to data object ID 4, and 417C points to data object ID 460. In addition, data segment pointer 421 points to data segment ID 470A, data segment pointer 422 points to data segment ID 471A, data segment pointers 423 and 433 point to data segment ID 472A, etc. Having described the structure and organization of one embodiment of a catalog and a backup data storing data for a backup agent, attention will now turn to reading, writing, and versioning backup data files. [002] FIG. illustrates one embodiment of a user interface 00 that may be used by backup agent 2 to view backup data and metadata. Interface 00 may include features such as drop-down menus, a navigation bar, an address field, and so on. Within the interface 00 are shown two panes, a Folders pane 01 and a Details pane 03. Within the Folders pane, a hierarchy of folders and their associated subfolders may be seen and navigated. The illustrated hierarchy includes folders,,,, and 0. Each folder may include one or more sub-folders. For example, folder includes subfolders and folder 0 includes subfolders 1-. As illustrated in FIG., subfolder 1 has been selected as indicated by its highlighted background. [0026] In the Details pane 03, details of the portion of the hierarchy selected in the Folders pane 01 may be listed in rows. Each row may include fields for Name, Size, Type, and Creation date for a particular data entity. For example, subfolder 1 is listed with a Type of "File Folder" and a Creation date of " :18 AM." Subfolder 1 includes a subfolder 60, which includes files File 64 has been selected as indicated by its highlighted background. It is noted that the naming conventions depicted in FIG. (e.g., "Subfolder 1", "File 64") are provided for ease of discussion. In a typical embodiment, names reflective of those used in ordinary business or other environments (e.g., Accounting, Payroll) may be used. [0027] When a desired data entity is selected in the Folders pane 01, backup agent 2 may send a query to catalog 26 to retrieve the names and attributes of each folder and its included files that are part of the data entity. Conventional file manipulation techniques may be used with data entities within interface 00, such as dragand-drop, right-click to open a list of actions, etc. A read operation may be performed when a data entity is selected. A write operation may be performed on a selected data entity through a right-click menu, a drag-and-drop action, from a pull-down menu, etc. When a write operation creates a new data entity such as during backup of a newly created folder, backup agent 2 may add an entry to catalog 26 and corresponding data to backup data 24. However, when a write operation modifies a data entity that is already present in catalog 26, such as during backup of a file within an existing folder or revision of an existing file, a different process that will be described with reference to FIG. 6 may be followed. [0028] FIG. 6 illustrates one embodiment of modifications made during a write operation to data and metadata corresponding to a data entity that is stored in backup data 24. In the illustrated embodiment, entry 4 within catalog 26 corresponds to the data entity to be modified. Entry 4 may generally correspond to the entry 4 described in FIG. 4. Prior to the start of the write operation, depicted by the portion of FIG. 6 labeled "Before Modification," entry 4 includes a data object pointer 4C that points to a data object in backup data 24. Data object is associated with data segment pointers 41-44, which point to the data segment IDs of data segments When a write operation begins, as shown in the portion of FIG. 6 labeled "At Open()," a new catalog entry may be created. The new entry s filename and attributes may remain the same, 4A and 48. However, the new entry is created with a new data object pointer 6C pointing to a new data object 60 that is created in backup data 24. In one embodiment, the new data object 60 may represent a revised version of the data object. The new data object 60 initially receives the same data segment pointers associated with data object ID. In addition to creating the new data

6 9 EP B1 object, the data entity is restored in order to be open for editing. During the write operation, data segment pointers may be added to and/or removed from the new data object as the data entity is edited. Any portion of the data entity that is deleted may cause one or more corresponding data segment pointers to be removed. Any portion of the data entity that is added may cause one or more corresponding data segment pointers to be added. Existing data segments which are modified may cause corresponding data segment pointers to be replaced by pointers to new data segments which reflect the revisions. [0029] For example, as shown in the portion of FIG. 6 labeled "After Modification," data segment pointers have been removed and data segment pointers 62-6 have been added. Data segment pointers 62-6 point to data segments whose IDs are 677A- 680A, respectively. Backup agent 2 also sends data segment 677B-680B to backup data 24. Note that data segments 477B-479B are not removed from backup data 24 since they are still pointed to by at least data segment pointers associated with data object ID. By keeping these data segments in the backup data, an earlier version of the data entity may be preserved, read, and revised further. If backup data 24 is not single-instance backup data, data segment pointers such as pointer 41 may be replaced with new data segment pointers and copies of corresponding data segments stored in backup data 24. In addition, once the write operation is completed, the restored data entity may be sent to backup data 24 and attributes 4B may be replaced with a new set of attributes 6B. For example, the new attributes 6B may include a new timestamp. [00] FIG. 7 illustrates one embodiment of a process 700 for modifying a file within backup data. Process 700 may begin with the identification of a desired file, such as through a browser (block 7). Once a desired file has been identified, it may be opened (block 7) or otherwise made available for access. For example, a copy of the file is restored from backup data. When a file is opened, a new version of a corresponding data object may be created in the data object table of the backup data where the identified file s data segments are stored (block 7) and an associated metadata entry may be added to the catalog (block 73). The data segment pointers that are associated with the file s data object may be added to the new data object in the backup data s data object table (737). The file may then be modified (block 7). Any new data segments that result from the modification of the file may be stored in the backup data (block 70). Data segment pointers corresponding to the new data segments may be added to and data segment pointers corresponding to portions of the file that are deleted may be removed from the new data object (block 760). The file may remain open and blocks 7, 70, and 760 repeated until modifications are complete (decision block 770). For example, the file may remain open during an editing session. Once modifications are complete, the file may be closed (block 780). When the file is closed, the metadata associated with the new data object may be updated in the catalog (block 790), completing process 700 (block 79). [0031] FIG. 8 illustrates one embodiment of catalog metadata corresponding to multiple versions of a file. In the illustrated embodiment, catalog 26 includes entries corresponding to multiple branches and versions of a file with a filename of 8A. Initially, as shown at the upper left portion of FIG. 8, filename 8 A is associated with attributes 8B and data object pointer 8C. Each time the corresponding file is modified, its metadata may change. For example, on the left column of FIG. 8, filename 8A is successively associated with attributes 8B and data object pointer 8C, 8B and data object pointer 8C, 8B and data object pointer 8C, 80B and data object pointer 80C, and 860B and data object pointer 860C. The left column of FIG. 8 may correspond to a first branch of the file. A second branch of the file is shown in the center column and a third branch of the file is shown in the right column of FIG. 8. The second branch may be created from the version of the file whose metadata includes data object pointer 8C as indicated by a connecting arrow. In the second branch, the file may be further modified as indicated by the successive association of filename 8 with attributes 831B and data object pointer 831C, 832B and data object pointer 832C, 833B and data object pointer 833C, 834B and data object pointer 834C, 83B and data object pointer 83C, and 836B and data object pointer 836C. Similarly, the third branch may be created from the version of the file whose metadata includes data object pointer 834C as indicated by a connecting arrow. In the third branch, the file may be further modified as indicated by the successive association of filename 8 with attributes 8341B and data object pointer 8341C, 8342B and data object pointer 8342C, and 8343B and data object pointer 8343C, etc. Similar versioning and branching may be performed for any of the data objects associated with a file, as will be apparent to one of ordinary skill in the art. [0032] It is noted that the above-described embodiments may comprise software. In such an embodiment, the program instructions that implement the methods and/or mechanisms may be conveyed or stored on a computer readable medium. Numerous types of media which are configured to store program instructions are available and include hard disks, floppy disks, CD-ROM, DVD, flash memory, Programmable ROMs (PROM), random access memory (RAM), and various other forms of volatile or non-volatile storage. [0033] Although the embodiments above have been described in considerable detail, numerous variations and modifications will become apparent to those skilled in the art once the above disclosure is fully appreciated. 6

7 11 EP B1 12 Claims 1. A computer system comprising: or more new data segments, add a new pointer referencing the respective data segment to the new data object. a first host (2) configured to store a plurality of backup files (24) on a backup medium, wherein each backup file includes one or more data segments of a plurality of data segments (470B-480B), wherein in said storing the plurality of backup files, the first host is configured to: 2. The system as recited in claim 1, wherein the first data object corresponding to the first backup file includes a first pointer to a first data segment stored on the backup medium, wherein a second data object corresponding to a second backup file includes a second pointer to the same first data segment. store the plurality of data segments on the backup medium; and for each respective backup file of the plurality of backup files, store on the backup medium a respective data object (4, 4, 4,, 460) that references the one or more data segments of the respective file; and a backup catalog (26) configured to store metadata describing the plurality of backed up files, wherein the metadata includes a separate entry (411A-417A) for each of the backed up files, and wherein the entry for each respective file includes a filename that corresponds to the respective file and a pointer (411C-417C) to the respective data object that references the one or more data segments of the respective file; a second host (2) configured to convey to the first host a request to open a first backup file stored on the backup medium, wherein the first backup file corresponds to a first data object; wherein in response to the request to open the first backup file, the first host is configured to: create a new data object on the backup medium corresponding to a new version of the first backup file; and copy each pointer referencing each of the one or more data segments of the first backup file from the first data object to the new data object; and wherein the second host is further configured to: after said conveying the request to open the first backup file, convey to the first host one or more modification requests (7) to modify the first backup file, wherein the one or more modification requests include one or more new data segments; and wherein, in response to receiving the one or more modification requests, the first host is further configured to: store the one or more new data segments on the backup medium (70); and for each respective data segment of the one The system as recited in claim 1 or 2, wherein in response to the request to open the first backup file, the first host is further configured to: add a new entry to the backup catalog that identifies the new version of the first backed up file, wherein the new entry includes a pointer to the new data object. 4. The system as recited in claim 1, wherein the second host is coupled to the first host via a network, wherein the second host comprises a virtual file system interface (3) to the backup medium, and wherein the second host is configured to convey to the first host, via the virtual file system interface, a request to open a first backed up file of the plurality of backed up files.. The system as recited in claim 4, wherein the virtual file system interface is configured to convey the one or more modification requests to the first host. 6. The system as recited in claim 4 or, wherein the virtual file system interface is configured to enable the second host to create and store multiple versions of backup files on the backup medium. 7. The system as recited in claim 1, wherein the backup catalog further includes metadata describing branches containing multiple versions of the plurality of backup files. 8. The system as recited in claim 1, wherein in response to receiving the one or more modification requests, the first host is further configured to modify the new data object to reference the one or more new data segments (760), wherein in said modifying the new data object the first host is configured to perform one or more of: add one or more new pointers to the new data object to reference the one or more new data segments; or remove one or more of the previously existing pointers from the new data object. 9. A method comprising: storing a plurality of backup files on a backup medium, wherein each backup file includes one 7

8 13 EP B1 14 or more data segments of a plurality of data segments (470B-480B), wherein said storing the plurality of backup files includes: storing the plurality of data segments on the backup medium; and for each respective backup file of the plurality of backup files, storing on the backup medium a respective data object (4, 4, 4,, 460) that references the one or more data segments of the respective file; and maintaining a backup catalog (26) configured to store metadata describing the plurality of backed up files, wherein the metadata includes a separate entry (411A-417A) for each of the backed up files, and wherein the entry for each respective file includes a filename that corresponds to the respective file and a pointer to the respective data object that references the one or more data segments of the respective file; wherein the method further comprises: receiving a request to open a first backup file stored on the backup medium, wherein the first backup file corresponds to a first data object; in response to said receiving the request to open the first backup file: creating a new data object on the backup medium corresponding to a new version of the first backup file; and copying each pointer referencing each of the one or more data segments of the first backup file from the first data object to the new data object; and wherein the method further comprises: 2 3. The method as recited in claim 9, wherein the first data object corresponding to the first backup file includes a first pointer to a first data segment stored on the backup medium, wherein a second data object corresponding to a second backup file includes a second pointer to the same first data segment. 11. The method as recited in claim 9 or, further comprising: adding a new entry to the backup catalog that identifies the new version of the first backed up file, wherein the new entry includes a pointer to the new data object. 12. The method as recited in claim 9, further comprising accessing the backup medium through a virtual file system interface. 13. The method as recited in claim 9, further comprising conveying the one or more modification requests through the virtual file system interface 14. The method as recited in claim 9 further comprising storing metadata describing branches containing multiple versions of the plurality of backup files.. A computer readable medium storing computer instructions that are executable by a processor to: store a plurality of backup files on a backup medium, wherein each backup file includes one or more data segments of a plurality of data segments, wherein in said storing the plurality of backup files the instructions are executable by the processor to: store the plurality of data segments on the backup medium; and for each respective backup file of the plurality of backup files, store on the backup medium a respective data object that references the one or more data segments of the respective file; and receiving one or more modification requests to modify the first backup file, wherein the one or more modification requests include one or more new data segments; and in response to said receiving the one or more modification requests: storing the one or more new data segments on the backup medium (70); and for each respective data segment of the one or more new data segments, adding a new pointer referencing the respective data segment to the new data object. 4 0 maintain a backup catalog configured to store metadata describing the plurality of backed up files, wherein the metadata includes a separate entry for each of the backed up files, and wherein the entry for each respective file includes a filename that corresponds to the respective file and a pointer to the respective data object that references the one or more data segments of the respective file wherein the instructions are further executable by the processor to: receive a request to open a first backup file stored on the backup medium, wherein the first 8

9 EP B1 16 backup file corresponds to a first data object; in response to said receiving the request to open the first backup file: Patentansprüche 1. Computersystem, das Folgendes umfasst: create a new data object on the backup medium corresponding to a new version of the first backup file; and copy each pointer referencing each of the one or more data segments of the first backup file from the first data object to the new data object; wherein the instructions are further executable by the processor to: receive one or more modification requests to modify the first backup file, wherein the one or more modification requests include one or more new data segments; in response to said receiving the one or more modification requests: einen ersten Host (2), der dazu ausgelegt ist, mehrere Sicherungsdateien (24) auf einem Sicherungsmedium zu speichern, wobei jede Sicherungsdatei eines oder mehrere Datensegmente von mehreren Datensegmenten (470B- 480B) umfasst, wobei der Host bei dem Speichern der mehreren Sicherungsdateien für Folgendes ausgelegt ist: Speichern der mehreren Datensegmente auf dem Sicherungsmedium; und Speichern eines jeweiligen Datenobjekts (4, 4, 4,, 460) für jede jeweilige Sicherungsdatei der mehreren Sicherungsdateien auf dem Sicherungsmedium, das auf das eine oder die mehreren Datensegmente der jeweiligen Datei verweist; und store the one or more new data segments on the backup medium; and for each respective data segment of the one or more new data segments, add a new pointer referencing the respective data segment to the new data object. 16. The computer readable medium as recited in claim, first data object corresponding to the first backup file includes a first pointer to a first data segment stored on the backup medium, wherein a second data object corresponding to a second backup file includes a second pointer to the same first data segment. 17. The computer readable medium as recited in claim or 16, wherein the instructions are further executable by the processor to: add a new entry to the backup catalog that identifies the new version of the first backed up file, wherein the new entry includes a pointer to the new data object. 18. The computer readable medium as recited in claim 16, wherein the instructions are further executable to access the backup medium through a virtual file system interface. 19. The computer readable medium as recited in claim 17 or 18, wherein the instructions are further executable to convey the one or more modification requests through the virtual file system interface. The computer readable medium as recited in claim, wherein the instructions are further executable to store metadata describing branches containing multiple versions of the plurality of backup files einen Sicherungskatalog (26), der dazu ausgelegt ist, Metadaten zu speichern, die die mehreren gesicherten Dateien beschreiben, wobei die Metadaten einen separaten Eintrag (411A- 417A) für jede der gesicherten Dateien umfassen, und wobei der Eintrag für jede jeweilige Datei einen Dateinamen, der der jeweiligen Datei entspricht, und einen Zeiger (411C-417C) auf das jeweilige Datenobjekt, das auf den einen oder die mehreren Datensegmente der jeweiligen Datei verweist, umfasst; einen zweiten Host (2), der dazu ausgelegt ist, eine Anfrage an den ersten Host zu übermitteln, um eine erste Sicherungsdatei zu öffnen, die auf dem Sicherungsmedium gespeichert ist, wobei die erste Sicherungsdatei einem ersten Datenobjekt entspricht; wobei der erste Host als Antwort auf die Anfrage wegen des Öffnens der ersten Sicherungsdatei für Folgendes ausgelegt ist: Erstellen eines neuen Datenobjekts auf dem Sicherungsmedium, das einer neuen Version der ersten Sicherungsdatei entspricht; und Kopieren jedes Zeigers, der auf jedes des einen oder der mehreren Datensegmente der ersten Sicherungsdatei verweist, von dem ersten Datenobjekt zu dem neuen Datenobjekt; und wobei der zweite Host ferner für Folgendes ausgelegt ist: Übermitteln einer oder mehrerer Modifikationsanfragen (7), um die erste Sicherungsdatei zu modifizieren, nach dem Übermitteln der Anfrage wegen des Öffnens der ersten 9

10 17 EP B1 18 Sicherungsdatei, wobei die eine oder die mehreren Modifikationsanfragen eines oder mehrere neue Datensegmente umfassen; und wobei der erste Host als Antwort auf das Empfangen der einen oder der mehreren Modifikationsanfragen für Folgendes ausgelegt ist: Speichern des einen oder der mehreren Datensegmente auf dem Sicherungsmedium (70); und Hinzufügen eines neuen Zeigers, der auf das jeweilige Datensegment verweist, zu dem neuen Datenobjekt für jedes jeweilige Datensegment von dem einen oder den mehreren neuen Datensegmenten. 2. System nach Anspruch 1, wobei das erste Datenobjekt, das der ersten Sicherungsdatei entspricht, einen ersten Zeiger auf ein erstes Datensegment umfasst, das auf dem Sicherungsmedium gespeichert ist, wobei ein zweites Datenobjekt, das einer zweiten Sicherungsdatei entspricht, einen zweiten Zeiger auf dasselbe erste Datensegment umfasst. 3. System nach Anspruch 1 oder 2, wobei der erste Host als Antwort auf die Anfrage wegen des Öffnens der ersten Sicherungsdatei ferner für Folgendes ausgelegt ist: Hinzufügen eines neuen Eintrags zu dem Sicherungskatalog, der die neue Version der ersten gesicherten Datei identifiziert, wobei der neue Eintrag einen Zeiger auf das neue Datenobjekt umfasst. 4. System nach Anspruch 1, wobei der zweite Host mit dem ersten Host über ein Netz gekoppelt ist, wobei der zweite Host eine virtuelle Dateisystem-Schnittstelle (3) zu dem Sicherungsmedium umfasst und wobei der zweite Host dafür ausgelegt ist, eine Anfrage wegen des Öffnens der ersten gesicherten Datei von den mehreren gesicherten Dateien über die virtuelle Dateisystem-Schnittstelle an den ersten Host zu übermitteln.. System nach Anspruch 4, wobei die virtuelle Dateisystem-Schnittstelle dazu ausgelegt ist, die eine oder die mehreren Modifikationsanfragen an den ersten Host zu übermitteln. 6. System nach Anspruch 4 oder, wobei die virtuelle Dateisystem-Schnittstelle dazu ausgelegt ist, es dem zweiten Host zu ermöglichen, mehrere Versionen von Sicherungsdateien auf dem Sicherungsmedium zu erstellen und zu speichern System nach Anspruch 1, wobei der Sicherungskatalog ferner Metadaten umfasst, die Zweige beschreiben, die mehrere Versionen der mehreren Sicherungsdateien enthalten. 8. System nach Anspruch 1, wobei der erste Host als Antwort auf das Empfangen des einen oder der mehreren Modifikationsanfragen ferner dazu ausgelegt ist, das neue Datenobjekt zu modifizieren, damit es auf das eine oder die mehreren neuen Datensegmente (760) verweist, wobei der erste Host bei dem Modifizieren des neuen Datenobjekts dazu ausgelegt ist, Folgendes auszuführen: Hinzufügen eines oder mehrerer Zeiger zu dem neuen Datenobjekt, um auf das eine oder die mehreren neuen Datensegmente zu verweisen; und/oder Entfernen des einen oder der mehreren vorher vorhandenen Zeiger aus dem neuen Datenobjekt. 9. Verfahren, das Folgendes umfasst: Speichern mehrerer Sicherungsdateien auf einem Sicherungsmedium, wobei jede Sicherungsdatei eines oder mehrere Datensegmente von den mehreren Datensegmenten (470B- 480B) umfasst, wobei das Speichern der mehreren Sicherungsdateien Folgendes umfasst: Speichern der mehreren Datensegmente auf dem Sicherungsmedium; und Speichern eines jeweiligen Datenobjekts (4, 4, 4,, 460) für jede jeweilige Sicherungsdatei von den mehreren Sicherungsdateien auf dem Sicherungsmedium, das auf das eine oder die mehreren Datensegmente der jeweiligen Datei verweist; und Führen eines Sicherungskatalogs (26), der dazu ausgelegt ist, Metadaten zu speichern, die die mehreren gesicherten Dateien beschreiben, wobei die Metadaten einen separaten Eintrag (411A-417A) für jede der gesicherten Dateien umfassen, und wobei der Eintrag für jede jeweilige Datei einen Dateinamen, der der jeweiligen Datei entspricht, und einen Zeiger zu dem jeweiligen Datenobjekt, das auf den einen oder die mehreren Datensegmente der jeweiligen Datei verweist, umfasst; wobei das Verfahren ferner Folgendes umfasst: Empfangen einer Anfrage wegen des Öffnens einer ersten Sicherungsdatei, die auf dem Sicherungsmedium gespeichert ist, wobei die erste Sicherungsdatei einem ersten Datenobjekt entspricht; als Antwort auf die Anfrage wegen des Öffnens der ersten Sicherungsdatei:

11 19 EP B1 Erstellen eines neuen Datenobjekts auf dem Sicherungsmedium, das einer neuen Version der ersten Sicherungsdatei entspricht; und Kopieren jedes Zeigers, der auf jedes des einen oder der mehreren Datensegmente der ersten Sicherungsdatei verweist, von dem ersten Datenobjekt zu dem zweiten Datenobjekt; und wobei das Verfahren ferner Folgendes umfasst: Empfangen einer oder mehrerer Modifikationsanfragen, um die erste Sicherungsdatei zu modifizieren, wobei die eine oder die mehreren Modifikationsanfragen eines oder mehrere neue Datensegmente umfassen; und als Antwort auf das Empfangen der einen oder der mehreren Modifikationsanfragen: Speichern des einen oder der mehreren Datensegmente auf dem Sicherungsmedium (70); und Hinzufügen eines neuen Zeigers, der auf das jeweilige Datensegment verweist, für jedes jeweilige Datensegment von dem einen oder den mehreren neuen Datensegmenten zu dem neuen Datenobjekt.. Verfahren nach Anspruch 9, wobei das erste Datenobjekt, das der ersten Sicherungsdatei entspricht, einen ersten Zeiger auf ein erstes Datensegment umfasst, das auf dem Sicherungsmedium gespeichert ist, wobei ein zweites Datenobjekt, das einer zweiten Sicherungsdatei entspricht, einen zweiten Zeiger auf dasselbe erste Datensegment umfasst. 11. Verfahren nach Anspruch 9 oder, das ferner Folgendes umfasst: Hinzufügen eines neuen Eintrags zu dem Sicherungskatalog, der die neue Version der ersten gesicherten Datei identifiziert, wobei der neue Eintrag einen Zeiger auf das neue Datenobjekt umfasst. 12. Verfahren nach Anspruch 9, das ferner ein Zugreifen auf das Sicherungsmedium durch eine virtuelle Dateisystem-Schnittstelle umfasst. 13. Verfahren nach Anspruch 9, das ferner ein Übermitteln der einen oder der mehreren Modifikationsanfragen durch die virtuelle Dateisystem-Schnittstelle umfasst Verfahren nach Anspruch 9, das ferner ein Speichern von Metadaten umfasst, die Zweige beschreiben, die mehrere Versionen der mehreren Sicherungsdateien enthalten.. Computerlesbares Medium, das Computerbefehle speichert, die von einem Prozessor für Folgendes ausführbar sind: Speichern mehrerer Sicherungsdateien auf einem Sicherungsmedium, wobei jede Sicherungsdatei eine oder mehrere Datensegmente von mehreren Datensegmenten umfasst, wobei die Befehle bei dem Speichern der mehreren Sicherungsdateien von dem Prozessor für Folgendes ausführbar sind: Speichern der mehreren Datensegmente auf dem Sicherungsmedium; und Speichern eines jeweiligen Datenobjekts für jede jeweilige Sicherungsdatei der mehreren Sicherungsdateien auf dem Sicherungsmedium, das auf das eine oder die mehreren Datensegmente der jeweiligen Datei verweist; und Führen eines Sicherungskatalogs, der dazu ausgelegt ist, Metadaten zu speichern, die die mehreren gesicherten Dateien beschreiben, wobei die Metadaten einen separaten Eintrag für jede der gesicherten Dateien umfassen, und wobei der Eintrag für jede jeweilige Datei einen Dateinamen, der der jeweiligen Datei entspricht, und einen Zeiger zu dem jeweiligen Datenobjekt, das auf den einen oder die mehreren Datensegmente der jeweiligen Datei verweist, umfasst; wobei die Befehle von dem Prozessor ferner für Folgendes ausführbar sind: Empfangen einer Anfrage wegen des Öffnens einer ersten Sicherungsdatei, die auf dem Sicherungsmedium gespeichert ist, wobei die erste Sicherungsdatei einem ersten Datenobjekt entspricht; als Antwort auf die Anfrage wegen des Öffnens der ersten Sicherungsdatei: Erstellen eines neuen Datenobjekts auf dem Sicherungsmedium, das einer neuen Version der ersten Sicherungsdatei entspricht; und Kopieren jedes Zeigers, der auf jedes des einen oder der mehreren Datensegmente der ersten Sicherungsdatei verweist, von dem ersten Datenobjekt zu dem zweiten Datenobjekt; und wobei die Befehle von dem Prozessor ferner für Folgendes ausführbar sind: Empfangen einer oder mehrerer Modifikations- 11

12 21 EP B1 22 anfragen, um die erste Sicherungsdatei zu modifizieren, wobei die eine oder die mehreren Modifikationsanfragen eines oder mehrere neue Datensegmente umfassen; und als Antwort auf das Empfangen der einen oder der mehreren Modifikationsanfragen: un support de sauvegarde, où chaque fichier de sauvegarde comprend un ou plusieurs segments de données d une pluralité de segments de données (470B-480B), où dans ledit stockage de la pluralité de fichiers de sauvegarde, le premier hôte est configuré pour : Speichern des einen oder der mehreren Datensegmente auf dem Sicherungsmedium; und Hinzufügen eines neuen Zeigers, der auf das jeweilige Datensegment verweist, für jedes jeweilige Datensegment von dem einen oder den mehreren neuen Datensegmenten zu dem neuen Datenobjekt. 16. Computerlesbares Medium nach Anspruch, wobei das erste Datenobjekt, das der ersten Sicherungsdatei entspricht, einen ersten Zeiger auf ein erstes Datensegment umfasst, das auf dem Sicherungsmedium gespeichert ist, wobei ein zweites Datenobjekt, das einer zweiten Sicherungsdatei entspricht, einen zweiten Zeiger auf dasselbe erste Datensegment umfasst. 17. Computerlesbares Medium nach Anspruch oder 16, wobei die Befehle von dem Prozessor ferner für Folgendes ausführbar sind: Hinzufügen eines neuen Eintrags zu dem Sicherungskatalog, der die neue Version der ersten gesicherten Datei identifiziert, wobei der neue Eintrag einen Zeiger auf das neue Datenobjekt umfasst. 18. Computerlesbares Medium nach Anspruch 16, wobei die Befehle ferner für Folgendes ausführbar sind: Zugreifen auf das Sicherungsmedium durch die virtuelle Dateisystem-Schnittstelle. 19. Computerlesbares Medium nach Anspruch 17 oder 18, wobei die Befehle ferner für Folgendes ausführbar sind: Übermitteln der einen oder der mehreren Modifikationsanfragen durch die virtuelle Dateisystem-Schnittstelle.. Computerlesbares Medium nach Anspruch, wobei die Befehle ferner für Folgendes ausführbar sind: Speichern von Metadaten, die Zweige beschreiben, die mehrere Versionen der mehreren Sicherungsdateien enthalten. Revendications 1. Système informatique comprenant : un premier hôte (2) configuré pour stocker une pluralité de fichiers de sauvegarde (24) sur stocker la pluralité de segments de données sur le support de sauvegarde ; et pour chaque fichier de sauvegarde respectif de la pluralité de fichiers de sauvegarde, stocker sur le support de sauvegarde un objet de données respectif (4, 4, 4,, 460) qui référence le ou les segments de données du fichier respectif ; et un catalogue de sauvegarde (26), configuré pour stocker des métadonnées décrivant la pluralité de fichiers sauvegardés, où les métadonnées comprennent une entrée séparée (411A-417A) pour chacun des fichiers sauvegardés, et où l entrée pour chaque fichier respectif comprend un nom de fichier qui correspond au fichier respectif et un pointeur (411C-417C) vers l objet de données respectif qui référence le ou les segments de données du fichier respectif ; un second hôte (2), configuré pour acheminer vers le premier hôte une demande pour ouvrir un premier fichier de sauvegarde stocké sur le support de sauvegarde, où le premier fichier de sauvegarde correspond à un premier objet de données ; où, en réponse à la demande d ouvrir le premier fichier de sauvegarde, le premier hôte est configuré pour : créer un nouvel objet de données sur le support de sauvegarde correspondant à une nouvelle version du premier fichier de sauvegarde ; et copier chaque pointeur référençant chacun du ou des segments de données du premier fichier de sauvegarde du premier objet de données au nouvel objet de données ; et où le second hôte est en outre configuré pour : après ledit acheminement de la demande pour ouvrir le premier fichier de sauvegarde, acheminer vers le premier hôte une ou plusieurs demandes de modification (7) pour modifier le premier fichier de sauvegarde, où la ou les demandes de modification comprennent un ou plusieurs nouveaux segments de données ; et où, en réponse à la réception de la ou des demandes de modification, le premier hôte est en outre configuré pour : stocker le ou les nouveaux segments de données sur le support de sauvegarde (70) ; et 12

13 23 EP B1 24 pour chaque segment de données respectif du ou des nouveaux segments de données, ajouter un nouveau pointeur référençant le segment de données respectif vers le nouvel objet de données. 2. Système tel que décrit dans la revendication 1, dans lequel le premier objet de données correspondant au premier fichier de sauvegarde comprend un premier pointeur vers un premier segment de données stocké sur le support de sauvegarde, où un second objet de données correspondant à un second fichier de sauvegarde comprend un second pointeur vers le même premier segment de données. 3. Système tel que décrit dans la revendication 1 ou la revendication 2, dans lequel, en réponse à la demande d ouvrir le premier fichier de sauvegarde, le premier hôte est en outre configuré pour : ajouter une nouvelle entrée au catalogue de sauvegarde qui identifie la nouvelle version du premier fichier sauvegardé, où la nouvelle entrée comprend un pointeur vers le nouvel objet de données. 4. Système tel que décrit dans la revendication 1, dans lequel le second hôte est couplé au premier hôte par l intermédiaire d un réseau, où le second hôte comprend une interface de système de fichiers virtuels (3) vers le support de sauvegarde, et où le second hôte est configuré pour acheminer vers le premier hôte, par l intermédiaire de l interface de système de fichiers virtuels, une demande pour ouvrir un premier fichier sauvegardé de la pluralité de fichiers sauvegardés.. Système tel que décrit dans la revendication 4, dans lequel l interface de système de fichiers virtuels est configurée pour acheminer la ou les demandes de modification vers le premier hôte. 6. Système tel que décrit dans la revendication 4 ou la revendication, dans lequel l interface de système de fichiers virtuels est configurée pour permettre au second hôte de créer et de stocker plusieurs versions de fichiers de sauvegarde sur le support de sauvegarde. 7. Système tel que décrit dans la revendication 1, dans lequel le catalogue de sauvegarde comprend en outre des métadonnées décrivant des branches contenant plusieurs versions de la pluralité de fichiers de sauvegarde. 8. Système tel que décrit dans la revendication 1, dans lequel, en réponse à la réception de la ou des demandes de modification, le premier hôte est en outre configuré pour modifier le nouvel objet de données pour référencer le ou les nouveaux segments de données (760), où dans ladite modification du nouvel objet de données, le premier hôte est configuré pour exécuter une ou plusieurs des étapes suivantes : ajouter un ou plusieurs nouveaux pointeurs vers le nouvel objet de données pour référencer le ou les nouveau segments de données ; ou supprimer un ou plusieurs des pointeurs préalablement existants du nouvel objet de données. 9. Procédé comprenant les étapes suivantes : stocker une pluralité de fichiers de sauvegarde sur un support de sauvegarde, où chaque fichier de sauvegarde comprend un ou plusieurs segments de données d une pluralité de segments de données (470B-480B), où ledit stockage de la pluralité de fichiers de sauvegarde comprend les étapes suivantes : stocker la pluralité de segments de données sur le support de sauvegarde ; et pour chaque fichier de sauvegarde respectif de la pluralité de fichiers de sauvegarde, stocker sur le support de sauvegarde un objet de données respectif (4, 4, 4,, 460) qui référence le ou les segments de données du fichier respectif ; et maintenir un catalogue de sauvegarde (26) configuré pour stocker des métadonnées décrivant la pluralité de fichiers sauvegardés, et où les métadonnées comprennent une entrée séparée (411A-417A) pour chacun des fichiers sauvegardés, et où l entrée pour chaque fichier respectif comprend un nom de fichier qui correspond au fichier respectif et un pointeur vers l objet de données respectif qui référence le ou les segments de données du fichier respectif ; où le procédé comprend en outre les étapes suivantes : recevoir une demande pour ouvrir un premier fichier de sauvegarde stocké sur le support de sauvegarde, où le premier fichier de sauvegarde correspond à un premier objet de données ; en réponse à ladite réception de la demande d ouvrir le premier fichier de sauvegarde : créer un nouvel objet de données sur le support de sauvegarde correspondant à une nouvelle version du premier fichier de sauvegarde ; et copier chaque pointeur référen- 13

14 2 EP B1 26 çant chacun du ou des segments de données du premier fichier de sauvegarde du premier fichier de sauvegarde depuis le premier objet de données vers le nouvel objet de données ; et où le procédé comprend en outre les étapes suivantes : recevoir une ou plusieurs demandes de modification pour modifier le premier fichier de sauvegarde, où les une ou plusieurs demandes de modification comprennent un ou plusieurs nouveaux segments de données ; et en réponse à ladite réception des une ou plusieurs demandes de modification : stocker le ou les nouveaux segments de données sur le support de sauvegarde (70) ; et pour chaque segment de données respectif du ou des nouveaux segments de données, ajouter un nouveau pointeur référençant le segment de données respectif vers le nouvel objet de données.. Procédé tel que décrit dans la revendication 9, dans lequel le premier objet de données correspondant au premier fichier de sauvegarde comprend un premier pointeur vers un premier segment de données stocké sur le support de sauvegarde, où un second objet de données correspondant à un second fichier de sauvegarde comprend un second pointeur vers le même premier segment de données. 11. Procédé tel que décrit dans la revendication 9 ou la revendication, comprenant en outre l étape suivante : ajouter une nouvelle entrée au catalogue de sauvegarde qui identifie la nouvelle version du premier fichier sauvegardé, où la nouvelle entrée comprend un pointeur vers le nouvel objet de données Procédé tel que décrit dans la revendication 9, comprenant en outre d acheminer la ou les demandes de modification par l intermédiaire de l interface de système de fichiers virtuels. 14. Procédé tel que décrit dans la revendication 9, comprenant en outre de stocker des métadonnées décrivant des branches contenant plusieurs versions de la pluralité de fichiers de sauvegarde.. Support lisible par ordinateur stockant des instructions informatiques qui sont exécutables par un processeur pour : stocker une pluralité de fichiers de sauvegarde sur un support de sauvegarde, où chaque fichier de sauvegarde comprend un ou plusieurs segments de données d une pluralité de segments de données, où dans ledit stockage de la pluralité de fichiers de sauvegarde, les instructions sont exécutables par le processeur pour : stocker la pluralité de segments de données sur le support de sauvegarde ; et pour chaque fichier de sauvegarde respectif de la pluralité de fichiers de sauvegarde, stocker sur le support de sauvegarde un objet de données respectif qui référence le ou les segments de données du fichier respectif ; et maintenir un catalogue de sauvegarde configuré pour stocker des métadonnées décrivant la pluralité de fichiers sauvegardés, où les métadonnées comprennent une entrée séparée pour chacun des fichiers sauvegardés, et où l entrée pour chaque fichier respectif comprend un nom de fichier qui correspond au fichier respectif et un pointeur vers l objet de données respectif qui référence le ou les segments de données du fichier respectif ; où les instructions sont en outre exécutables par le processeur pour : recevoir une demande pour ouvrir un premier fichier de sauvegarde stocké sur le support de sauvegarde, où le premier fichier de sauvegarde correspond à un premier objet de données ; en réponse à ladite réception de la demande d ouvrir le premier fichier de sauvegarde : 12. Procédé tel que décrit dans la revendication 9, comprenant en outre d accéder au support de sauvegarde par l intermédiaire d une interface de système de fichiers virtuels. créer un nouvel objet de données sur le support de sauvegarde correspondant à une nouvelle version du premier fichier de sauvegarde ; et 14

15 27 EP B1 28 copier chaque pointeur référençant chacun du ou des segments de données du premier fichier de sauvegarde du premier objet de données au nouvel objet de données ; où les instructions sont en outre exécutables par le processeur pour : recevoir une ou plusieurs demandes de modification pour modifier le premier fichier de sauvegarde, où les une ou plusieurs demandes de modification comprennent un ou plusieurs nouveaux segments de données ; en réponse à ladite réception des une ou plusieurs demandes de modification : sauvegarde par l intermédiaire d une interface de système de fichiers virtuels. 19. Support lisible par ordinateur tel que décrit dans la revendication 17 ou la revendication 18, dans lequel les instructions sont en outre exécutables pour acheminer la ou les demandes de modification par l intermédiaire de l interface de système de fichiers virtuels.. Support lisible par ordinateur tel que décrit dans la revendication, dans lequel les instructions sont en outre exécutables pour stocker des métadonnées décrivant des branches contenant plusieurs versions de la pluralité de fichiers de sauvegarde. stocker le ou les nouveaux segments de données sur le support de sauvegarde ; et pour chaque segment de données respectif du ou des nouveaux segments de données, ajouter un nouveau pointeur référençant le segment de données respectif vers le nouvel objet de données. 16. Support lisible par ordinateur tel que décrit dans la revendication, le premier objet de données correspondant au premier fichier de sauvegarde comprend un premier pointeur vers un premier segment de données stocké sur le support de sauvegarde, où un second objet de données correspondant à un second fichier de sauvegarde comprend un second pointeur vers le même premier segment de données Support lisible par ordinateur tel que décrit dans la revendication ou la revendication 16, dans lequel les instructions sont en outre exécutables par le processeur pour : 4 ajouter une nouvelle entrée au catalogue de sauvegarde qui identifie la nouvelle version du premier fichier sauvegardé, où la nouvelle entrée comprend un pointeur vers le nouvel objet de données. 18. Support lisible par ordinateur tel que décrit dans la revendication 16, dans lequel les instructions sont en outre exécutables pour accéder au support de 0

16 EP B1 16

17 EP B1 17

18 EP B1 18

19 EP B1 19

20 EP B1