Patent Grant
8966288
The invention relates to data transfer mechanisms, and in particular, to a software-based, high speed data pipe for providing high speed and reliable data transfer between computers.
Data, in the process of being archived or transferred from one location to another, will pass through various phases where different operations such as compression, network transfer, storage, etc. will take place on it. There are essentially two approaches that can be taken when implementing such a transfer mechanism. One would be to split the archival process into sub-tasks, each of which would perform a specific function (e.g. Compression). This would then require copying of data between the sub-tasks, which could prove processor intensive. The other method would be to minimize copies, and have a monolithic program performing all of the archival functions. The downside to this would be loss of parallelism. A third alternative would of course be to use threads to do these tasks and use thread-signaling protocols, however, it is realized that this would not be entirely practical since threads are not fully supported on many computing platforms.
Accordingly, it is highly desirable to obtain a high-speed data transfer mechanism implemented in software and developed for the needs of high speed and reliable data transfer between computers. It is also desirable to provide a mechanism to encrypt the data being transferred.
In accordance with embodiments of the invention, a method is provided for performing a storage operation in a pipeline storage system in which one or more data streams containing data to be stored are written into data chunks. The method includes generating an encryption key associated with a first archive file to be stored when encryption is requested for the storage operation, encrypting the archive data from the data stream using the encryption key to create an encrypted data chunk when a data stream containing the archive file is processed in the pipeline storage system, storing the encrypted data chunk on a storage medium, and storing the encryption key in a manner accessible during a restore operation of the encrypted data chunk.
The invention will be better understood with reference to the following drawings, in which:
The present invention includes methods and systems operating in conjunction with a modular storage system to enable computers on a network to share storage devices on a physical and logical level. An exemplary modular storage system is the GALAXY™ backup and retrieval system and QiNetix™ storage management system available from CommVault Systems of New Jersey. The modular architecture underlying this system is described in the above referenced patent applications, each of which is incorporated herein.
Preferred embodiments of the invention are now described with reference to the drawings. An embodiment of the system of the present invention is shown in
A client 8 can be any networked client 8 and preferably includes at least one attached information store 90A. The information store 90A may be any memory device or local data storage device known in the art, such as a hard drive, CD-ROM drive, tape drive, RAM, or other types of magnetic, optical, digital and/or analog local storage. In some embodiments of the invention, the client 8 includes at least one data agent 95A, which is a software module that is generally responsible for performing storage operations on data of a client 8 stored in information store 90A or other memory location. Storage operations include, but are not limited to, creation, storage, retrieval, migration, deletion, and tracking of primary or production volume data, secondary volume data, primary copies, secondary copies, auxiliary copies, snapshot copies, backup copies, incremental copies, differential copies, synthetic copies, HSM copies, archive copies, Information Lifecycle Management (“ILM”) copies, and other types of copies and versions of electronic data. In some embodiments of the invention, the system provides at least one, and typically a plurality of data agents 95A for each client, each data agent 95A is intended to backup, migrate, and recover data associated with a different application. For example, a client 8 may have different individual data agents 95A designed to handle Microsoft Exchange data, Lotus Notes data, Microsoft Windows file system data, Microsoft Active Directory Objects data, and other types of data known in the art.
The storage manager 100A is generally a software module or application that coordinates and controls the system, for example, the storage manager 100A manages and controls storage operations performed by the system. The storage manager 100A communicates with all components of the system including client 8, data agent 95A, media agent 105A, and storage devices 115A to initiate and manage storage operations. The storage manager 100A preferably has an index 107A, further described herein, for storing data related to storage operations. In general, the storage manager 100A communicates with storage devices 115A via a media agent 105A. In some embodiments, the storage manager 100A communicates directly with the storage devices 115A.
The system includes one or more media agent 105A. The media agent 105A is generally a software module that conducts data, as directed by the storage manager 100A, between the client 8 and one or more storage devices 115A, for example, a tape library, a hard drive, a magnetic media storage device, an optical media storage device, or other storage device. The media agent 105A is communicatively coupled with and controls the storage device 115A. For example, the media agent 105A might instruct a storage device 115A to perform a storage operation, e.g., archive, migrate, or restore application specific data. The media agent 105A generally communicates with the storage device 115A via a local bus such as a SCSI adaptor.
Each media agent 105A maintains an index cache 110A which stores index data that the system generates during storage operations as further described herein. For example, storage operations for Microsoft Exchange data generate index data. Media management index data includes, for example, information regarding the location of the stored data on a particular media, information regarding the content of the information stored such as file names, sizes, creation dates, formats, application types, and other file-related criteria, information regarding one or more clients associated with the information stored, information regarding one or more storage policies, storage criteria, or storage preferences associated with the information stored, compression information, retention-related information, encryption-related information, stream-related information, and other types of information. Index data thus provides the system with an efficient mechanism for performing storage operations including locating user files for recovery operations and for managing and tracking stored data.
The system generally maintains two copies of the media management index data regarding particular stored data. A first copy is generally stored with the data copied to a storage device 115A. Thus, a tape may contain the stored data as well as index information related to the stored data. In the event of a system restore, the index information stored with the stored data can be used to rebuild a media agent index 110A or other index useful in performing storage operations. In addition, the media agent 105A that controls the storage operation also generally writes an additional copy of the index data to its index cache 110A. The data in the media agent index cache 110A is generally stored on faster media, such as magnetic media, and is thus readily available to the system for use in storage operations and other activities without having to be first retrieved from the storage device 115A.
The storage manager 100A also maintains an index cache 107A. Storage manager index data is used to indicate, track, and associate logical relationships and associations between components of the system, user preferences, management tasks, and other useful data. For example, the storage manager 100A might use its index cache 107A to track logical associations between media agent 105A and storage devices 115A. The storage manager 100A may also use its index cache 107A to track the status of storage operations to be performed, storage patterns associated with the system components such as media use, storage growth, network bandwidth, service level agreement (“SLA”) compliance levels, data protection levels, storage policy information, storage criteria associated with user preferences, retention criteria, storage operation preferences, and other storage-related information.
A storage policy is generally a data structure or other information which includes a set of preferences and other storage criteria for performing a storage operation. The preferences and storage criteria may include, but are not limited to: a storage location, relationships between system components, network pathway to utilize, retention policies, data characteristics, compression or encryption requirements, preferred system components to utilize in a storage operation, and other criteria relating to a storage operation. A storage policy may be stored to a storage manager index, to archive media as metadata for use in restore operations or other storage operations, or to other locations or components of the system.
Index caches 107A and 110A typically reside on their corresponding storage component's hard disk or other fixed storage device. For example, the jobs agent 102A of a storage manager 100A may retrieve storage manager index 107A data regarding a storage policy and storage operation to be performed or scheduled for a particular client 8. The jobs agent 102A, either directly or via another system module, communicates with the data agent 95A at the client 8 regarding the storage operation. In some embodiments, the jobs agent 102A also retrieves from the index cache 107A a storage policy associated with the client 8 and uses information from the storage policy to communicate to the data agent 95A one or more media agents 105A associated with performing storage operations for that particular client 8 as well as other information regarding the storage operation to be performed such as retention criteria, encryption criteria, streaming criteria, etc. The data agent 95A then packages or otherwise manipulates the client information stored in the client information store 90A in accordance with the storage policy information and/or according to a user preference, and communicates this client data to the appropriate media agent(s) 100A for processing. The media agent(s) 105A store the data according to storage preferences associated with the storage policy including storing the generated index data with the stored data, as well as storing a copy of the generated index data in the media agent index cache 110A.
In some embodiments, components of the system may reside and execute on the same computer. In some embodiments, a client component such as a data agent 95A, a media agent 105A, or a storage manager 100A coordinates and directs local archiving, migration, and retrieval application functions as further described in application Ser. No. 09/610,738. These client components can function independently or together with other similar client components.
Data and other information is transported throughout the system via buffers and network pathways including, among others, a high-speed data transfer mechanism, such as the CommVault DataPipe™, as further described in U.S. Pat. No. 6,418,478 and application Ser. No. 09/495,751, each of which is hereby incorporated herein by reference in its entirety. Self describing tag headers are disclosed in these applications wherein data is transferred between a flexible grouping of data transport modules each supporting a separate function and leveraging buffers in a shared memory space. Thus, a data transport module receives a chunk of data and decodes how the data should be processed according to information contained in the chunk's header, and in some embodiments, the chunk's trailer. U.S. Pat. No. 6,418,478 and application Ser. No. 09/495,751 generally address “logical data” transported via TCP/IP, however, embodiments of the invention herein are also contemplated which are directed to transporting, multiplexing, encrypting, and generally processing block level data as disclosed, for example, in pending application Ser. No. 10/803,542, titled Method And System For Transferring Data In A Storage Operation, which is hereby incorporated herein by reference in its entirety.
As discussed, these applications generally disclose systems and methods of processing logical data. Thus, for example, contiguous blocks of data from a file might be written on a first volume as blocks 1, 2, 3, 4, 5, etc. The operating system of the host associated with the first volume would assist in packaging the data adding additional OS-specific information to the chunks. Thus, when transported and stored on a second volume, the blocks might be written to the second in a non-contiguous order such as blocks 2, 1, 5, 3, 4. On a restore storage operation, the blocks could (due to the OS-specific information and other information) be restored to the first volume in contiguous order, but there was no control over how the blocks were laid out or written to the second volume. Incremental block level backups of file data was therefore extremely difficult if not impossible in such a system since there was no discernable relationship between how blocks were written on the first volume and how they were written on the second volume.
Thus, in some embodiments, the system supports transport and incremental backups (and other storage operations) of block level data via a TCP/IP (and other transport protocols) over a LAN, WAN, SAN, etc. Additional data is added to the multi-tag header discussed in the applications referenced above which communicates how each block was written on the first volume. Thus, for example, a header might contain a file map of how the blocks were written on the first volume and the map could be used to write the blocks in similar order on the second volume. In other embodiments, each chunk header might contain a pointer or other similar data structure indicating the chunk's position relative to other chunks in the file. Thus, when a file block or other block changed on the first volume, the system could identify and update the corresponding copy of the block located on the second volume and effectively perform an incremental backup or other storage operation.
In the system, for example as in the CommVault Galaxy system, archives are grouped by Storage Policy. Many clients/sub clients can point to the same Storage Policy. Each Storage Policy has a Primary copy and zero or more Secondary copies. Each Copy has one or more streams related to the number of Drives in a Drive Pool.
The system uses a tape media to its maximum capacity and throughput by multiplexing data from several clients onto the same media at the same time. The system allows for a stream to be reserved more than once by different clients and have multiple data movers write to this same piece of media.
During backup or other storage operations, data from a data agent to a media agent is transferred over a “Data pipeline” as further described herein and in U.S. Pat. No. 6,418,478 and application Ser. No. 09/495,751. One or more transport processes or modules, such as the Dsbackup in the CommVault Galaxy system, form the tail end on the Media Agent for the pipeline. For example, in the Galaxy system, the Datamover process running as part of Dsbackup is responsible for writing data to the media. For data multiplexing, many such Data movers belonging to different pipelines have to write to the same piece of media. This can be achieved by splitting the Datamover pipeline process into multiple components including a data receiver, a data writer, and other modules as necessary.
The DataPipe
A DataPipe comprises a named set of tasks executing within one or more computers that cooperate with each other to transfer and process data in a pipelined manner. Within a DataPipe, a pipeline concept is used to improve performance of data transfer across multiple computers in a network. However, within a DataPipe, any stage within the pipeline may have multiple instances, thus greatly increasing the scaleability and performance of the basic pipeline concept.
The DataPipe mechanism processes data by dividing its processing into logical tasks that can be performed in parallel. It then sequences those tasks in the order in which they are to act on the data. For example, a head task may extract data from a database, a second task may encrypt it, a third may compress it, a fourth may send it out over the network, a fifth may receive it from the network, and a sixth may write it to a tape. The latter two tasks may reside on a different computer than the others, for example.
All of the tasks that comprise a single DataPipe on a given computer have access to a segment of shared memory that is divided into a number of buffers. A small set of buffer manipulation primitives is used to allocate, free, and transfer buffers between tasks.
Semaphores (or other OS specific mutual exclusion or signaling primitives) are used to coordinate access to buffers between tasks on a given computer. Special tasks, called network agents, send and receive data across network connections using standard network protocols. These agents enable a DataPipe to connect across multiple computer systems. A single DataPipe can therefore reside on more than one computer and could reside on computers of different types.
Each task may be implemented as a separate thread, process, or as a procedure depending on the capabilities of the computing system on which the DataPipe is implemented.
As mentioned previously, each task may be implemented as a separate thread, or process, or as a procedure in a monolithic process (in cases where native platforms don't support any forms of parallel execution or multi processing). For data transfer across network, dedicated network readers and writers ensure communication across the net.
Referring to
As shown in
From the preceding discussion, one can ascertain that a pipeline or DataPipe 10 comprises a head task 15 that generates the data to be archived or transferred from store 50, and a tail task 40 which accomplishes the final task of storing or writing the data to store 60, including archiving or restoring on the data as shown in
A pipeline on a particular machine can be arranged to provide a feed to another different machine. A schematic diagram is illustrated in
In addition to the transferring of data from one computer to another, a unique capability of the datapipe invention is the ability to scale to enable full utilization of the bandwidth of a network, and to fully utilize the number of peripheral devices such as tape drives, or fully utilize other hardware components such as CPUs. The scaleability of a DataPipe is achieved by using multiple instances of each task in the pipeline.
For example, multiple head tasks operating in parallel may gather data from a database and deposit it into buffers. Those buffers may then be processed by several parallel tasks that perform a function such as encryption. The encryption tasks in turn may feed several parallel tasks to perform compression, and several parallel tasks may perform network send operations to fully exploit network bandwidth. On the target computer, several network reader tasks may receive data, which is written to multiple tape units by several tasks. All of these tasks on both computers are part of the same DataPipe and collectively perform the job of moving data from the database to tape units. They do this job extremely efficiently by fully utilizing all available bandwidth and hardware allocated to the DataPipe while also minimizing CPU cycles by avoiding unnecessary copying of the data as it moves from one stage of the DataPipe to the next.
In general, there could be N stages in a given DataPipe pipeline. At each stage of the pipeline, there could be p instances of a given module task. These N stages could all be on the local machine or could be split across two different machines in which case there are network writers and network readers (i.e. pseudo tail and head network agents) which work together to ensure continuity in the pipeline.
Referring to
Buffer Manipulation Primitives
Referring now to
Master_Monitor is connected to a predefined port, to enable it to communicate with its peers on other computer systems. Master_Monitor monitors the status of all DataPipes under its control at all times and is able to provide status of the DataPipe to the application software that uses the DataPipe.
To accomplish these above tasks, a master manager program called Master_Monitor executes in the preferred embodiment as a daemon on all process machines, listening on a well-known port, to serve requirements of pipeline operations. Master_Monitor functions to monitor status of all pipelines under its control at all times and reports status of the pipeline to all its sub-modules. As shown in
DataPipe Initiation
Referring now to
Referring now to
Identification
The process responsible for initiation of the pipeline constructs a name for the pipeline using its own process Id, a time stamp, and the name of the machine where the initiator process is running. This pipeline name is passed along with both the Initiate-Pipe as well as the EXTEND_Pipe message so that the pipeline is identified with the same name on all computers on which it is operating (i.e. both the remote as well as the local machine). All shared memory segments and semaphores (reference numeral 85 of
Data Transfer Implementation
Allocation: Receive: Send: Free
Directing attention to
As shown in
All FreeBuf( ) calls free their buffers into the input queue of first module. By the same rule, first stage modules are never permitted to do a ReceiveBuf( ) but are permitted to do AllocBuf( ). On the other hand, tail processes are permitted to perform only FreeBuf( ) and never permitted to do a SendBuf( ). All other modules can Receive, Allocate, Send, and Free buffers. First stage modules always perform SendBuf( ) after they execute each AllocBuf( ).
Each queue 95 is associated with a semaphore to guarantee orderly access to shared memory and which gets triggered upon actions such as AllocBuf( ), ReceiveBuf( ), SendBuf( ) and FreeBuf( ). Dedicated network agents thus map themselves across any network interface on the system, as long as data propagation is ensured. The number of network agents per pipeline is a configurable parameter, which helps this mechanism exploit maximum data transfer bandwidth available on the network over which it is operating. A single dedicated parent network thread/process monitors performance and status of all network agents on that particular machine for a particular pipeline.
Referring again to
Attachments
As the identification process is completed, all modules attach themselves to a specific shared memory space segment that is shared among modules on that machine for this particular pipeline. This shared memory segment has many data buffers, input queues for all stages on the pipeline, and their initial values. Each module identifies its own input queues and output queues depending on the stage that module is supposed to run at, and initial queue (first stage) is populated with number of data segments for sharing on this particular pipeline. Also all modules attach themselves to an allocator semaphore array, which controls the number of buffers allocated by a specific module that can be active in the pipeline.
Data Integrity
Integrity of the data passed along and the sequencing of data are maintained in part by a pair of special purpose modules termed sequencer and resequencer processes.
Referring now to
Hence, in the preferred embodiment, all data pipe transfers employing multi-instance stages via the sequencer/resequencer processes ensure that the input sequence of sequence numbers are not violated for each instance of the module. Further, the restriction that all modules of a specific multi-instance stage should be of the same type eliminates the chances for preferential behavior.
Fairness
The concept of fairness means that each task will be assured of getting the input buffers it needs to operate on without waiting longer than necessary. Fairness among the modules in a given DataPipe where no stage of the pipeline has more than one instance is automatic. As the tail task frees a buffer it enters the free buffer pool where it may enable the head task to allocate it and begin processing. All tasks in the DataPipe operate a maximum speed overlapping the processing done by other tasks in the preceding or following stage of the pipeline.
If a DataPipe has stages consisting of parallel instances of a task, fairness among those tasks is assured by using an allocator semaphore which counts from Max_Buffers/NA (where NA is the number of allocators for this DataPipe on this particular machine) downward to zero. All FreeBuf( )s increment this semaphore back, however, there could be only Max_Buffers/NA buffers allocated by any allocator module in this DataPipe. This ensures that all allocators get a fair share of the available total number of input buffers. If a particular process attempts to allocate more buffers than it is allowed, the master_monitor process prevents such allocation, causing the process to either terminate or wait until a buffer currently allocated to the process becomes freed thereby incrementing the semaphore back up to allow the process to allocate another buffer.
Control Messages
All instances of all modules have a control socket to Master_Monitor over which control messages are exchanged. All network readers/writers have an analogous control socket to their parent network agent. The parent network agent itself has a control socket to Master_Monitor. Each module periodically checks its control socket for any messages from Master_Monitor. Critical information such as a STOP_PIPE message is passed to Master_Monitor via this mechanism.
Status Monitoring
Each module initiated by Master_Monitor on a given machine is monitored by either a parent network process (in the case of network reader or writer), or by Master_Monitor itself, for states of execution. In case any module is reported as having terminated abnormally, Master_Monitor identifies this exception, and signals all the modules on that particular pipeline to stop. This is done by means of control messages through control sockets as described previously. Upon safely stopping all modules pertaining to this particular pipeline, it signals the remote machine's Master_Monitor to stop the remote side of this particular pipeline and entire pipeline is shut down safely by means of control message signaling.
Implementation
In a preferred embodiment, DataPipe is implemented on Sun Solaris or HP-UX operating systems and incorporated into Release 2.7 of CommVault System's Vault98 storage management product.
To set up the DataPipe the Master_Monitor for this is called giving it the name of the DataPipe and the names of the modules that will use the pipe (module 10).
Master_Monitor (Initiate_Pipe(Sample_pipe,A,B,C)).
Within the logic of module A, Alloc_Buf( ) function is then called to obtain a buffer (20). The logic of module A may perform any actions it wants to fill the buffer with useful data. When it has completed its processing of the buffer (30), it calls SendBuf( ) to send the buffer to module B for processing (40). Module A then repeats its function by again calling Alloc_Buf( ) to obtain the next buffer.
The logic of module B calls ReceiveBuf( ) to obtain a buffer of data from module A (50). It then operates on the buffer by performing processing as required (60). When it is finished with the buffer it calls SendBuf( ) to send that buffer to module C (70).
Module B then repeats if function by again calling ReceiveBuf( ) to obtain the next buffer from module A.
Module C obtains a buffer of data from module B by calling ReceiveBuf( ). When it has completed its processing of the data in that buffer (90), it calls FreeBuf( ) to release the buffer (100). Like the other two modules, it loops back to receive the next buffer form module B.
The primitives used to allocate, free, send, and receive buffers are synchronized by the use of semaphores. This ensures coordination between the modules so that the receiving module does not start processing data before the sending module has finished with it. If no buffer is available, the AllocBuf or ReceiveBuf primitives will wait until one is available. All three modules operate in parallel as separate tasks. The order of processing from A to B to C is established in the initial call to Master_Monitor that established the DataPipe.
Referring now to
Encryption
As discussed above, the system also supports encrypted pipelined data transfer by allowing for encryption to be one of the processes or tasks performed in the datapipe.
Data protection in storage management systems is a tradeoff between user's convenience and security, speed of operation and capabilities of the encryption algorithm, length of the encryption keys, government restrictions, and other elements known in the art. There are many encryption algorithms available that vary by strength, speed and other parameters. Most encryption algorithms, however, offer various ways to manage the encryption keys. For example, some implementations include hardware USB devices that can store user's private keys. Whenever that user needs an access to some encrypted material, the hardware unit is inserted into the USB slot, and the key is retrieved from the unit. Some units have built-in encrypting capabilities, providing additional security: the key no longer has to travel over the USB bus. All crypto operations are conducted within the unit itself.
More conventional implementations involve storing secret keys in so-called key rings (technically just binary files with some specific format) protected with a user's pass-phrase. The user's pass-phrase is stored nowhere but in the user's head, so the secret keys can be considered to be almost secure. “Almost” because the security of the keys now depend on a human-selected word or phrase, and human languages are known to be quite redundant (1.3 bits for a letter in average English text), plus some sort of dictionary attack is possible. Thus, users and system administrators must chose a system of key management that best suits their particular needs.
Users and system administrators also confront the problem of key distribution. If there is more than one computer involved, there will be need for transferring keys from one machine to the other. One can say that a “secure” link is needed. But the security of such “secure” link has to be guaranteed by some other key, which should have been distributed first, but for distribution of which another secure session would be needed, etc. etc.
When transferring encrypted data, users generally must confront key management issues and often will want to have precise control over where sensitive information is stored and how this information is stored. In some embodiments, users only want some minimum scrambling, or want only the security of the pipeline connection for secure over-the-network data transfer, and prefer not to enter pass-phrases or use other methods every time they wish to encrypt or decrypt data. Such users will probably be satisfied storing the keys in some scrambled form on the CommServe, media agents, storage media, or other elements of the system. Thus, in some embodiments, for example, in the CommVault Galaxy system, the key management problem divides in two: key management on the CommServe or storage manager and key management on the backup media.
To be able to restore encrypted data back, the data encryption keys must generally be stored somewhere. While it is possible to store keys on the media itself where encrypted data is being stored, keys are generally stored in the storage manager or CommServe database/index cache. The CommServe can be configured to trust sensitive data to it unconditionally, or users may agree to store such data on the CommServe, provided that some additional protection is involved. For example, additional protection could be a pass-phrase that only customer knows.
Thus, as far as key storage on the CommServe is concerned, we generally have two cases: strong (where keys are encrypted with a pass-phrase) and weak (where keys are simply scrambled in the index cache)
With strong encryption key management (also referred to herein as “CS_KM_STRONG”), the data encryption keys are stored on the CommServe protected by some sort of a pass-phrase. For example, the pass-phrase may exist only in the customer's head. Such an encryption scheme offers many benefits. For example, even though the data encryption keys are kept on the CommServe and can be accessed by the storage management software, such as CommVault's Galaxy software, when needed, the storage manager still lacks one important piece of information without which the encryption keys cannot be reconstructed—the user's pass-phrase. Without this pass-phrase the keys are unusable, and the data is unrecoverable.
In some embodiments, the system prompts the user to enter the pass-phrase every time when a restore is attempted. In other embodiments, the system does not prompt users to enter pass-phrases during the backup (so that Galaxy could get the data encryption key to perform the backup encryption).
Asymmetric public-key cryptography is used to facilitate this latter method. Asymmetric algorithms use two keys instead of one. The first key (called public) is not protected, and is used to encrypt the data. The second key (called private) is guarded by all means, and can be used to decrypt the data. Thus, in some embodiments, the system encrypts backup data with the public key (which can be stored unprotected in the CS database), and decrypt backup data with the private key (which will be protected by user's pass-phrase). In some embodiments as further described herein, poor performance of asymmetric crypto algorithms may avoided by using symmetric cipher to perform data encryption, and storing the symmetric data encryption key encrypted with the asymmetric public key.
With weak encryption key management (also referred to herein as “CS_KM_WEAK”), keys are merely scrambled in the storage manager index cache and do not generally require a pass-phrase. For example, in some embodiments, users may consider their CommServes to be secure or at minimal risk and thus not require a strong encryption key management scheme as discussed above. Also, users may dislike the additional inconvenience of having a pass-phrase to remember and enter during restores.
Thus, the data encryption key is stored in a scrambled form in the database. Something is generally referred to as “scrambled” if it's made unintelligible by some sort of built-in algorithm, which is not controlled by any key or pass-phrase that would exist separately from this algorithm. Clearly, scrambling is potentially less secure than strong encryption key management, because by isolating the scrambling/descrambling code in the Galaxy binaries, any scrambled information can be restored to its original form. The advantage of scrambling (weak) over pass-phrase (strong) encryption is that both backups and restores will not require user to provide any extra information (such as the pass-phrase).
In some embodiments, for example, in an application service provider (“ASP”) setting or other similar setting, trust level varies between components of the system. For example, an ASP might maintain Media Agents and CommServes in an ASP-controlled data center, but the system's Data Agents might belong to the ASP's customers. Thus, the Data Agents may or may not be configured to fully trust the ASP to handle their data. In this situation the data being backed up belongs to the customers, and the customers completely trust Data Agents (because they're in customer's physical control), but Media Agents and CommServe databases are handled by ASP, so customers don't really trust either of them.
One possible solution is to protect everything with a pass-phrase, which the ASP's customers can set without the ASP's knowledge. There is no real problem here except for the customer now having to specify pass-phrase each time when they perform restore operation. In some embodiments, however, this minor inconvenience can be worked around by means of pass-phrase export files. These files are kept on the Data Agent in a dedicated directory (e.g. /opt/galaxy/PF or some other similar directory) and contain Data Agent's pass-phrase in some scrambled form. Thus, each time a restore starts, the restore process looks for the pass-phrase export files on the destination machine, and if such file is found, use the enclosed pass-phrase to unlock the encryption keys. Thus, the customer can restore his data to his machines w/o having to provide a pass-phrase, but for anyone else (including the ASP), data restoration is impossible without the pass-phrase.
In some embodiments, unattended Synthetic Full backups present a different problem. Synthetic Full backups combine a full backup with several incrementals to produce a new full backup. This combining involves running backup and restore pipelines. Since restoring encrypted data generally requires a pass-phrase, unattended SynthFull backups are often impossible when CommServe security is CS_KM_STRONG.
One possible work around this problem is to have a copy of asymmetric public key stored scrambled (rather than encrypted with the user pass-phrase) specially for running SynthFull backup processes. The hack is gross, because in theory, the system could be directed to use the same key to run restores as well. Cryptographic protection thus gets reduced to protection by code.
Generally, encryption keys are not stored on backup media with the information that they protect since doing so is somewhat akin to locking a house and then putting the keys under the doormat. Yet, if the system doesn't store any information on the backup media, the recovery of data in case of disasters (such as a CommServe failure) will generally be extremely difficult if not impossible. Again, there is a tradeoff here between the overall security of the storage management system and the user's convenience. Thus, in some embodiment, key management on the backup media does occur.
For the key management on the media, there are a number of distinct security levels. The first is merely scrambling keys on the backup media. Due to its potential weaknesses as further described below, this method of media key management is referred to herein as MM_KM_WEAK throughout the rest of the document. One weakness to this method is that anyone able to figure out the scrambling algorithm (or anyone able to invoke unscrambling code) will be able to fully recover the backup media without any additional knowledge required. All the strengths of encryption algorithms employed by the system are thus likely nullified by this loophole. Yet, this scheme has some advantages, and some uses: (1) The user never has to remember or enter a pass-phrase. All backups/restores remain fully automatic, and the encryption is 100% transparent to the operator. (2) The data traveling over the pipeline is still encrypted and protected against interception by an eavesdropper. Thus, in some situations, MM_KM_WEAK may be desirable.
Another scheme is strong media key management (“MM_KM_STRONG”). In this embodiment, data encryption keys are stored on the media, but we additionally encrypt them with the user's pass-phrase. The pass-phrase becomes the crucial information that will exist only in the customer head, without which the data cannot generally be reconstructed by third party or even by an encrypted data recovery tool such as CommVault's Galaxy DrTool.
The final method of media key management is referred to herein as paranoid (“MM_KM_PARANOID”). In this case there are NO keys stored on the media at all. Data recovery without the CommServe database will generally be impossible and data recover tools such as DrTool will not work since the encrypted data on the media will not contain any additional information these tools require to decrypt and recover the data.
The tables below summarize various advantages and disadvantages of key management schemes on the storage manager and on backup media according to embodiments of the invention:
Besides the encryption key management/storage problem discussed above, there is also a key exchange problem: even if the keys are stored on the CommServe in a secure way, they generally must be somehow transferred to the IDA and MA—the places where the real data encryption or decryption generally takes place.
If the keys are distributed in a clear text form, they can easily be intercepted by an eavesdropper, and used later to restore backup data. This method is the least secure and should generally be avoided if possible.
If keys are distributed in “scrambled” form, the eavesdropper's task becomes more difficult, but still possible via the usual scrambling drawback. Once an entry point to the system's unscrambling routing is found and negotiated, for example to the Galaxy CvLib DLL, any scrambled message can be recovered.
If keys are distributed encrypted with some user's chosen password, the user would have to enter that password twice for each client/MA: once on the client itself, second time on the CommServe, so that the password would be stored somewhere in registry for further usage, and would never appear on the network. While providing an illusion of security, this third method is, however, potentially inconvenient for the user (too many passwords to enter), and the security gain is not that great: the passwords will have to be stored in file or registry anyway, and they can appear there only in the “scrambled” form (unless we ask user to enter yet another password, etc. etc.). Thus in some embodiments, a variant of this scheme uses an automatically chosen password that requires no user interaction and still yields a good security. For example, the system uses a session network password to encrypt encryption keys when they're sent between the machines. Each client computer has a unique session network password, and CommServe knows all these passwords, so they become a very good candidate for data key encryption before sending them over the network. Since IDAs and MAs don't know each other's network password, they cannot easily exchange keys, so, in some embodiments, it becomes the responsibility of the CommServe to generate random data encryption key and to pass it to IDA and/or MA—depending on where the encryption should take place.
This section describes various aspects of data encryption implementation according to embodiments of the invention, and relies upon the following variables:
Thus, exemplary encryption schemes and methods according to embodiments of the invention can be represented as follows:
In some embodiments, the KRSApri key encrypted with a user-selectable pass-phrase according to this equation:
KRsAprienc=BF(KRsApri,MD5(PassPhrase))
Before storing into the database it is scrambled and converted to ASCII:
KRSAprienc,db=Scramble(KRSAprienc)
In some embodiments, during restores the KRSAprienc, db key is decrypted back to KRSAPri according to this formula:
KRSApri=BF(Scramble−1(KRSAprienc,db),MD5(PassPhrase))
In some embodiments, the KRSApub key is stored in the database in this scrambled ASCII format:
KRSApubdb=Scramble(KRSApub)
And the reverse transformation is represented as follows:
KRSApub=Scramble−1(KRSApubdb)
In some embodiments, the backup data text is encoded as follows:
Ei=BF(Di,KBF)ECB(Electronic Codeblock Mode)
or
E0=BF(D0{circle around (x)}IVBF,KBF)CBC(Cipher Block Chaining Mode)
Ei=BF(Di{circle around (x)}Ei-1,KBF)
In some embodiments, during restores, the encoded data will be decrypted as follows:
Di=BF(Ei,KBF)ECB Mode
or
D0=BF(E0,KBF){circle around (x)}IVBFCBCMode
Di=BF(Ei,KBF){circle around (x)}Ei-1
Before being stored into the database, the backup data key KBF (chosen randomly for every backup) is encrypted according to this formula in some embodiments:
KBFenc=RSA(KBF,KRSApub)
In the database it is stored in the scrambled ASCII format in some embodiments:
KBFenc,db=Scramble(KBFenc)
In some embodiments, during restores the KBF key will be recovered from the database according to this formula:
KBF=RSA−1(Scramble−1(KBFenc,db),KRSApri)
And KRSApri is decrypted from KRSAprienc,db as described above
In some embodiments, before being transmitted from CS to IDA or MA, the KBF m key is encrypted with the client's network password according to this formula:
KBFnetenc=BF(KBF,NetPass)
In some embodiments, the client obtains KBF from KBFnetenc using this formula:
KBF=BF(KBFnetenc,NetPass)
There are three categories of encryption-related settings that are stored in the storage manager database/index cache:
The KBF keys are generated randomly for every archive file. The KRSApub and KRSApri are created once per client. All subclients will share the same RSA keys, the same pass-phrase and the same key management settings. To be able to turn encryption ON/OFF individually for every subclient, the “encryption ON” flag should be stored per subclient.
Encryption settings and their related GUIs are generally split into two groups: those that are specified for the entire client computer, and those that can be customized on a per-subclient basis.
The settings specified for the entire client computer (all subclients) are summarized in the screenshot in
When “Pass-phrase: <Export>” button is pressed in
There is minimal space in the self describing multi-tag headers to specify additional information regarding the encryption functionality disclosed herein. In some embodiments, the system also uses a new header variable or field, but this results in a loss of compatibility with earlier clients.
Unlike compression, when encrypting data an expansion of the data size is possible and in some cases expected. For example, since the Blowfish algorithm is a block cipher, it will pad data to the block boundary (64 bits). Thus, it can add up to 7 bytes to data associated with every tag header. This padding is not an issue though because when backup data is put into the pipeline buffer, it's aligned on the boundary of 8. And since 64 bits constitute exactly 8 bytes, Blowfish will merely consume the unused alignment bytes, and no data expansion will occur. Unfortunately, lack of data expansion described above is true only for the simplest mode of Blowfish operation. If one implements a more secure CBC mode, or even just adds a 32-bit CRC32 checksum to guarantee data consistency, we will observe expansion of up to 4+8 bytes per tag header (CRC32+IV).
Current data pipeline code already implements a failsafe margin (at least 2 KB in some embodiments) used by the compression module during data compression (and released afterwards), so the same can be done for encryption as well. In fact, the encryption module uses the same margin, which is already there for compression. If user backs up a great deal of tiny files, there will be 12*N bytes expansion, where N is the number of tag headers that fit one pipeline buffer (32K). The 2 KB failsafe buffer will be exhausted if average file size is 96 bytes (32K/(size+96)*12=2K). Thus, the appropriate fall-back mechanism will have to be implemented in the encryption module, which would automatically allocate a new pipeline buffer, should this become necessary.
Each tag header will have a flag indicating whether the data has been encrypted or not. Since there is a 32-bit “compressed_data” flag already present in that structure, and that flag can take only values of 0 and 1, we can use the second byte of the integer to specify encryption algorithm to use. Other values of that byte can be reserved for encryption algorithms that may be implemented.
It may also be necessary to track the size of the data after it has been encrypted. For compression this is done by utilizing a second field of the tag header. For encryption another field out of the tag header is allocated for this purpose, or the compressed size replaced with the encrypted size and save the compressed size in an encryption header that would follow every tag header.
The system also uses chunk trailers from the data pipeline to store encryption information. Every chunk is followed by a chunk trailer that contains information about archive files encoded in the chunk. This information is used by data recovery tools, such as DrTool, to recover data in the absence of the CommServe database. For example, in some embodiments, the Chunk Trailer is the natural place where encryption keys can be stored to allow DrTool successfully decrypt the data without contacting the CS for the KBF.
In some embodiments, the Chunk Trailer is an ASCII entity comprised of two columns (one for variable names, the other one for the values). Depending on the media key management security level, the following information may be stored in the Chunk Trailer:
In the course of a backup, the following encryption-related events generally occur. Note that encryption can take place on a data agent or a media agent. Moreover, if network encryption is ON, but media encryption is OFF (ENC_NETWORK_ONLY), decryption may be happening as well:
The following events generally occur in a restore operation:
7. The decrypt/encrypt modules retrieve the appropriate key to process the archive file buffers. For example, the decrypt/encrypt modules intercept this buffer and issue CVA_GET AFILE_RESTORE KEY to ArchiveManager in order to retrieve KBF for this archive file.
The following sequence occurs during disaster recovery operations, for example, when the storage manager is unavailable or at other times:
In some embodiments, the system employs an encryption API on top of OpenSSL, for example, CommVault's CvDataCrypt API, that implements appropriate format of tag data encryption, scrambling, etc.
All keys are generally converted to ASCII form before storing them in the database.
The scrambler function randomizes the binary data, computes checksum, and encrypts the whole thing using some built-in key.
Systems and modules described herein may comprise software, firmware, hardware, or any combination(s) of software, firmware, or hardware suitable for the purposes described herein. Software and other modules may reside on servers, workstations, personal computers, computerized tablets, PDAs, and other devices suitable for the purposes described herein. Software and other modules may be accessible via local memory, via a network, via a browser or other application in an ASP context, or via other means suitable for the purposes described herein. Data structures described herein may comprise computer files, variables, programming arrays, programming structures, or any electronic information storage schemes or methods, or any combinations thereof, suitable for the purposes described herein. User interface elements described herein may comprise elements from graphical user interfaces, command line interfaces, and other interfaces suitable for the purposes described herein. Screenshots presented and described herein can be displayed differently as known in the art to input, access, change, manipulate, modify, alter, and work with information.
While the invention has been described and illustrated in connection with preferred embodiments, many variations and modifications as will be evident to those skilled in this art may be made without departing from the spirit and scope of the invention, and the invention is thus not to be limited to the precise details of methodology or construction set forth above as such variations and modification are intended to be included within the scope of the invention.
This application is a continuation of U.S. patent application Ser. No. 12/796,007 titled SYSTEM AND METHOD FOR PROVIDING ENCRYPTION IN STORAGE OPERATIONS IN A STORAGE NETWORK, SUCH AS FOR USE BY APPLICATION SERVICE PROVIDERS THAT PROVIDE DATA STORAGE SERVICES, filed Jun. 8, 2010, now U.S. Pat. No. 8,429,428, which is a continuation of U.S. patent application Ser. No. 11/843,453 titled SYSTEM AND METHOD FOR PROVIDING ENCRYPTION IN PIPELINED STORAGE OPERATIONS IN A STORAGE NETWORK, filed Aug. 22, 2007, now U.S. Pat. No. 7,739,381, which is continuation of U.S. patent application Ser. No. 10/990,284, titled SYSTEM AND METHOD FOR PROVIDING ENCRYPTION IN A STORAGE NETWORK BY STORING A SECURED ENCRYPTION KEY WITH ENCRYPTED ARCHIVE DATA IN AN ARCHIVE STORAGE DEVICE, filed Nov. 15, 2004, now U.S. Pat. No. 7,277,941, which is a continuation-in-part of U.S. patent application Ser. No. 10/144,683 titled PIPELINED HIGH SPEED DATA TRANSFER MECHANISM, filed May 13, 2002, now U.S. Pat. No. 7,401,154, which was a continuation of U.S. patent application Ser. No. 09/038,440 filed Mar. 11, 1998, now U.S. Pat. No. 6,418,478; all of which are hereby incorporated by reference in their entireties. Application Ser. No. 10/990,284 also claims the benefit of U.S. Provisional Patent Application No. 60/519,526 titled SYSTEM AND METHOD FOR PERFORMING PIPELINED STORAGE OPERATIONS IN A STORAGE NETWORK, filed Nov. 13, 2003, which application is incorporated herein by reference in its entirety. This application is related to the following patents and applications, each of which is hereby incorporated herein by reference in its entirety: Application Ser. No. 10/990,357 titled SYSTEM AND METHOD FOR COMBINING DATA STREAMS IN PIPELINED STORAGE OPERATIONS IN A STORAGE NETWORK, filed Nov. 15, 2004, now U.S. Pat. No. 7,315,923;Application Ser. No. 09/495,751, titled HIGH SPEED TRANSFER MECHANISM, filed Feb. 1, 2000, now U.S. Pat. No. 7,209,972;Application Ser. No. 09/610,738, titled MODULAR BACKUP AND RETRIEVAL SYSTEM USED IN CONJUNCTION WITH A STORAGE AREA NETWORK, filed Jul. 6, 2000, now U.S. Pat. No. 7,035,880;Application Ser. No. 09/774,268, titled LOGICAL VIEW AND ACCESS TO PHYSICAL STORAGE IN MODULAR DATA AND STORAGE MANAGEMENT SYSTEM, filed Jan. 30, 2001, now U.S. Pat. No. 6,542,972;Application Ser. No. 10/658,095, titled DYNAMIC STORAGE DEVICE POOLING IN A COMPUTER SYSTEM, filed Sep. 9, 2003, now U.S. Pat. No. 7,130,970; andApplication Ser. No. 60/460,234, titled SYSTEM AND METHOD FOR PERFORMING STORAGE OPERATIONS IN A COMPUTER NETWORK, filed Apr. 3, 2003.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4296465 | Lemak | Oct 1981 | A |
| 4686620 | Ng | Aug 1987 | A |
| 4995035 | Cole et al. | Feb 1991 | A |
| 5005122 | Griffin et al. | Apr 1991 | A |
| 5093912 | Dong et al. | Mar 1992 | A |
| 5133065 | Cheffetz et al. | Jul 1992 | A |
| 5193154 | Kitajima et al. | Mar 1993 | A |
| 5212772 | Masters | May 1993 | A |
| 5226157 | Nakano et al. | Jul 1993 | A |
| 5239647 | Anglin et al. | Aug 1993 | A |
| 5241668 | Eastridge et al. | Aug 1993 | A |
| 5241670 | Eastridge et al. | Aug 1993 | A |
| 5276860 | Fortier et al. | Jan 1994 | A |
| 5276867 | Kenley et al. | Jan 1994 | A |
| 5287500 | Stoppani, Jr. | Feb 1994 | A |
| 5301351 | Jippo et al. | Apr 1994 | A |
| 5311509 | Heddes et al. | May 1994 | A |
| 5321816 | Rogan et al. | Jun 1994 | A |
| 5333315 | Saether et al. | Jul 1994 | A |
| 5347653 | Flynn et al. | Sep 1994 | A |
| 5410700 | Fecteau et al. | Apr 1995 | A |
| 5448724 | Hayashi | Sep 1995 | A |
| 5491810 | Allen | Feb 1996 | A |
| 5495607 | Pisello et al. | Feb 1996 | A |
| 5504873 | Martin et al. | Apr 1996 | A |
| 5544345 | Carpenter et al. | Aug 1996 | A |
| 5544347 | Yanai et al. | Aug 1996 | A |
| 5559957 | Balk | Sep 1996 | A |
| 5559991 | Kanfi | Sep 1996 | A |
| 5598546 | Blomgren | Jan 1997 | A |
| 5619644 | Crockett et al. | Apr 1997 | A |
| 5638509 | Dunphy et al. | Jun 1997 | A |
| 5673381 | Huai et al. | Sep 1997 | A |
| 5682513 | Candelaria et al. | Oct 1997 | A |
| 5683513 | Fujimaki | Nov 1997 | A |
| 5699361 | Ding et al. | Dec 1997 | A |
| 5729743 | Squibb | Mar 1998 | A |
| 5751997 | Kullick et al. | May 1998 | A |
| 5752041 | Fosdick | May 1998 | A |
| 5758068 | Brandt et al. | May 1998 | A |
| 5758359 | Saxon | May 1998 | A |
| 5761677 | Senator et al. | Jun 1998 | A |
| 5761734 | Pfeffer et al. | Jun 1998 | A |
| 5764972 | Crouse et al. | Jun 1998 | A |
| 5778395 | Whiting et al. | Jul 1998 | A |
| 5805920 | Sprenkle et al. | Sep 1998 | A |
| 5812398 | Nielsen | Sep 1998 | A |
| 5813009 | Johnson et al. | Sep 1998 | A |
| 5813017 | Morris | Sep 1998 | A |
| 5860104 | Witt et al. | Jan 1999 | A |
| 5875478 | Blumenau | Feb 1999 | A |
| 5887134 | Ebrahim | Mar 1999 | A |
| 5901327 | Ofek | May 1999 | A |
| 5903646 | Rackman | May 1999 | A |
| 5924102 | Perks | Jul 1999 | A |
| 5950205 | Aviani, Jr. | Sep 1999 | A |
| 5956519 | Wise et al. | Sep 1999 | A |
| 5970233 | Liu et al. | Oct 1999 | A |
| 5970255 | Tran et al. | Oct 1999 | A |
| 5974563 | Beeler, Jr. | Oct 1999 | A |
| 5999629 | Heer et al. | Dec 1999 | A |
| 6003089 | Shaffer et al. | Dec 1999 | A |
| 6009274 | Fletcher et al. | Dec 1999 | A |
| 6012090 | Chung et al. | Jan 2000 | A |
| 6021415 | Cannon et al. | Feb 2000 | A |
| 6026414 | Anglin | Feb 2000 | A |
| 6052735 | Ulrich et al. | Apr 2000 | A |
| 6076148 | Kedem | Jun 2000 | A |
| 6094416 | Ying | Jul 2000 | A |
| 6094684 | Pallmann | Jul 2000 | A |
| 6105129 | Meier et al. | Aug 2000 | A |
| 6131095 | Low et al. | Oct 2000 | A |
| 6131190 | Sidwell | Oct 2000 | A |
| 6148412 | Cannon et al. | Nov 2000 | A |
| 6154787 | Urevig et al. | Nov 2000 | A |
| 6161111 | Mutalik et al. | Dec 2000 | A |
| 6167402 | Yeager | Dec 2000 | A |
| 6169976 | Colosso | Jan 2001 | B1 |
| 6212512 | Barney et al. | Apr 2001 | B1 |
| 6260069 | Anglin | Jul 2001 | B1 |
| 6269431 | Dunham | Jul 2001 | B1 |
| 6275953 | Vahalia et al. | Aug 2001 | B1 |
| 6292783 | Rohler et al. | Sep 2001 | B1 |
| 6301592 | Aoyama et al. | Oct 2001 | B1 |
| 6324581 | Xu et al. | Nov 2001 | B1 |
| 6328766 | Long | Dec 2001 | B1 |
| 6330570 | Crighton | Dec 2001 | B1 |
| 6330642 | Carteau | Dec 2001 | B1 |
| 6343324 | Hubis et al. | Jan 2002 | B1 |
| RE37601 | Eastridge et al. | Mar 2002 | E |
| 6356801 | Goodman et al. | Mar 2002 | B1 |
| 6381331 | Kato et al. | Apr 2002 | B1 |
| 6389432 | Pothapragada et al. | May 2002 | B1 |
| 6418478 | Ignatius et al. | Jul 2002 | B1 |
| 6421711 | Blumenau et al. | Jul 2002 | B1 |
| 6487561 | Ofek et al. | Nov 2002 | B1 |
| 6519679 | Devireddy et al. | Feb 2003 | B2 |
| 6538669 | Lagueux, Jr. et al. | Mar 2003 | B1 |
| 6542972 | Ignatius et al. | Apr 2003 | B2 |
| 6564228 | O'Connor | May 2003 | B1 |
| 6577734 | Etzel et al. | Jun 2003 | B1 |
| 6604149 | Deo et al. | Aug 2003 | B1 |
| 6654825 | Clapp et al. | Nov 2003 | B2 |
| 6658526 | Nguyen et al. | Dec 2003 | B2 |
| 6757680 | Choy | Jun 2004 | B1 |
| 6772332 | Boebert et al. | Aug 2004 | B1 |
| 6898286 | Murray | May 2005 | B2 |
| 6973621 | Sie et al. | Dec 2005 | B2 |
| 7035880 | Crescenti et al. | Apr 2006 | B1 |
| 7130970 | Devassy et al. | Oct 2006 | B2 |
| 7209972 | Ignatius et al. | Apr 2007 | B1 |
| 7213269 | Orthlieb et al. | May 2007 | B2 |
| 7277941 | Ignatius et al. | Oct 2007 | B2 |
| 7287045 | Saika et al. | Oct 2007 | B2 |
| 7287047 | Kavuri | Oct 2007 | B2 |
| 7315923 | Retnamma et al. | Jan 2008 | B2 |
| 7320068 | Zimniewicz et al. | Jan 2008 | B2 |
| 7328189 | Ling | Feb 2008 | B2 |
| 7360252 | Torrubia-Saez et al. | Apr 2008 | B1 |
| 7389273 | Irwin et al. | Jun 2008 | B2 |
| 7395282 | Crescenti et al. | Jul 2008 | B1 |
| 7401154 | Ignatius et al. | Jul 2008 | B2 |
| 7506102 | Lev-Ran et al. | Mar 2009 | B2 |
| 7519827 | Anderson et al. | Apr 2009 | B2 |
| 7581077 | Ignatius et al. | Aug 2009 | B2 |
| 7620976 | Low et al. | Nov 2009 | B2 |
| 7627569 | Gafter | Dec 2009 | B2 |
| 7627776 | Petruzzo | Dec 2009 | B2 |
| 7702693 | Aiyagari et al. | Apr 2010 | B1 |
| 7739381 | Ignatius et al. | Jun 2010 | B2 |
| 7748027 | Patrick | Jun 2010 | B2 |
| 7761713 | Baar | Jul 2010 | B2 |
| 7782742 | Park | Aug 2010 | B2 |
| 7805600 | Bucher et al. | Sep 2010 | B2 |
| 7818262 | Kavuri et al. | Oct 2010 | B2 |
| 7840537 | Gokhale et al. | Nov 2010 | B2 |
| 7882315 | Tsai et al. | Feb 2011 | B2 |
| 7926087 | Holl, II et al. | Apr 2011 | B1 |
| 8108427 | Prahlad et al. | Jan 2012 | B2 |
| 8131648 | Barton | Mar 2012 | B2 |
| 8140786 | Bunte et al. | Mar 2012 | B2 |
| 8165221 | Zheng et al. | Apr 2012 | B2 |
| 8219524 | Gokhale | Jul 2012 | B2 |
| 20020007347 | Blumenthal et al. | Jan 2002 | A1 |
| 20020007351 | Hillegass et al. | Jan 2002 | A1 |
| 20020077988 | Sasaki et al. | Jun 2002 | A1 |
| 20020120726 | Padole et al. | Aug 2002 | A1 |
| 20020128976 | O'Connor et al. | Sep 2002 | A1 |
| 20020147734 | Shoup et al. | Oct 2002 | A1 |
| 20020174011 | Sanchez et al. | Nov 2002 | A1 |
| 20020178366 | Ofir | Nov 2002 | A1 |
| 20030005428 | Roman | Jan 2003 | A1 |
| 20030200104 | Heming et al. | Oct 2003 | A1 |
| 20040093229 | Plain | May 2004 | A1 |
| 20040204949 | Shaji et al. | Oct 2004 | A1 |
| 20040210509 | Eder | Oct 2004 | A1 |
| 20040249759 | Higashi et al. | Dec 2004 | A1 |
| 20040255143 | Wemyss et al. | Dec 2004 | A1 |
| 20050004951 | Ciaramitaro et al. | Jan 2005 | A1 |
| 20050027657 | Leontiev et al. | Feb 2005 | A1 |
| 20050091655 | Probert et al. | Apr 2005 | A1 |
| 20050097440 | Lusk et al. | May 2005 | A1 |
| 20050108526 | Robertson | May 2005 | A1 |
| 20060147043 | Mann et al. | Jul 2006 | A1 |
| 20060224846 | Amarendran et al. | Oct 2006 | A1 |
| 20060242296 | Woolard et al. | Oct 2006 | A1 |
| 20060242558 | Racovolis et al. | Oct 2006 | A1 |
| 20070050362 | Low et al. | Mar 2007 | A1 |
| 20070198421 | Muller et al. | Aug 2007 | A1 |
| 20070198422 | Prahlad et al. | Aug 2007 | A1 |
| 20080005380 | Kawasaki et al. | Jan 2008 | A1 |
| 20080091747 | Prahlad et al. | Apr 2008 | A1 |
| 20080243795 | Prahlad et al. | Oct 2008 | A1 |
| 20080307020 | Ko et al. | Dec 2008 | A1 |
| 20080320319 | Muller et al. | Dec 2008 | A1 |
| 20090222907 | Guichard | Sep 2009 | A1 |
| 20090319534 | Gokhale | Dec 2009 | A1 |
| 20090319585 | Gokhale | Dec 2009 | A1 |
| 20100031017 | Gokhale et al. | Feb 2010 | A1 |
| 20100242096 | Varadharajan et al. | Sep 2010 | A1 |
| 20100313039 | Ignatius et al. | Dec 2010 | A1 |
| 20120023140 | Prahlad et al. | Jan 2012 | A1 |
| 20140101451 | Chan et al. | Apr 2014 | A1 |
| Number | Date | Country |
|---|---|---|
| 0259912 | Mar 1988 | EP |
| 0405926 | Jan 1991 | EP |
| 0467546 | Jan 1992 | EP |
| 0774715 | May 1997 | EP |
| 0809184 | Nov 1997 | EP |
| 0862304 | Sep 1998 | EP |
| 0899662 | Mar 1999 | EP |
| 0981090 | Feb 2000 | EP |
| WO-9513580 | May 1995 | WO |
| WO-9912098 | Mar 1999 | WO |
| Entry |
|---|
| U.S. Appl. No. 13/874,358, filed Apr. 30, 2013, Varadharajan et al. |
| U.S. Appl. No. 13/899,384, filed May 21, 2013, Prahlad et al. |
| Armstead et al., “Implementation of a Campus-wide Distributed Mass Storage Service: The Dream vs. Reality,” IEEE, 1995, pp. 190-199. |
| Arneson, “Mass Storage Archiving in Network Environments,” Digest of Papers, Ninth IEEE Symposium on Mass Storage Systems, Oct. 31, 1988-Nov. 3, 1988, pp. 45-50, Monterey, CA. |
| Cabrera et al., “ADSM: A Multi-Platform, Scalable, Backup and Archive Mass Storage System,” Digest of Papers, Compcon '95, Proceedings of the 40th IEEE Computer Society International Conference, Mar. 5, 1995-Mar. 9, 1995, pp. 420-427, San Francisco,CA. |
| CommVault Systems, Inc., “QiNetix Books Online Documentation,” submitted on CD-ROM, released Dec. 2005. |
| CommVault, “Firewall Considerations—How to,” <http://documentation.commvault.com/commvault/release—8—0—0/books—online—1/english—us/features/firewall/firewall—how—to.htm>, internet accessed on Feb. 27, 2009, 11 pages. |
| CommVault, “Firewall Considerations,” <http://documentation.commvault.com/commvault/release—8—0—0/books—online—1/english—us/features/firewall/firewall.htm>, internet accessed on Feb. 27, 2009, 8 pages. |
| Eitel, “Backup and Storage Management in Distributed Heterogeneous Environments,” IEEE, 1994, pp. 124-126. |
| Jander, M., “Launching Storage-Area Net,” Data Communications, US, McGraw Hill, NY, vol. 27, No. 4 (Mar. 21, 1998), pp. 64-72. |
| Jason Gait, “The Optical File Cabinet: A Random-Access File System for Write-Once Optical Disks,” IEEE Computer, vol. 21, No. 6, pp. 11-22 (1988) (see in particular figure 5 in p. 15 and recitation in claim 5). |
| Kwok, S. H., Digital rights management for the online music business, SIGecom Exch. 3, 3 (Jun. 2002), available at <http://doi.acm.org/10.1145/844339.844347>, 8 pages. |
| Microsoft SQL Server Documentation, “Adding a Member to a Predefined Role,” 2000. |
| Microsoft SQL Server Documentation, “Adding a Member to a SQL Server Database Role,” 2000. |
| Microsoft SQL Server Documentation, “Adding a Windows User or Group,” 2000. |
| Microsoft SQL Server Documentation, “Authentication Modes,” 2000. |
| Microsoft SQL Server Documentation, “Creating User-Defined SQL Server Database Roles,” 2000. |
| Microsoft SQL Server Documentation, “Granting a Windows User or Group Access to a Database,” 2000. |
| Microsoft SQL Server Documentation, “Security Architecture,” 2000. |
| PCT International Search Report for Application No. PCT/US01/02931, Mar. 6, 2002. |
| Ron White, “How Computers Work”, Sixth Edition, Que Corporation, Jun. 26, 2002, 7 pages. |
| Rosenblum et al., “The Design and Implementation of a Log-Structured File System,” Operating Systems Review SIGPOPS, vol. 25, No. 5, New York, US, pp. 1-15 (May 1991). |
| Number | Date | Country | |
|---|---|---|---|
| 20130311785 A1 | Nov 2013 | US |
| Number | Date | Country | |
|---|---|---|---|
| 60519526 | Nov 2003 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 12796007 | Jun 2010 | US |
| Child | 13868320 | US | |
| Parent | 11843453 | Aug 2007 | US |
| Child | 12796007 | US | |
| Parent | 10990284 | Nov 2004 | US |
| Child | 11843453 | US | |
| Parent | 09038440 | Mar 1998 | US |
| Child | 10144683 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10144683 | May 2002 | US |
| Child | 10990284 | US |
