White Paper

TapeWare ® - Secure Data Protection for Windows, NetWare, Linux, and UNIX Server and Workstation Environments

Source: InoStor - A Tandberg Data Company
Overview
TapeWare by Yosemite Technologies offers a scalable data protection solution that is recognized by top manufacturers. TapeWare provides a reliable, affordable, and easy to implement solution for small office and enterprise-level applications alike. TapeWare supports mixed-platform data protection of any combination of Windows, NetWare, Linux and UNIX based systems, through an intuitive, easy-to-use interface. Essential features like remote administration, distributed backup processing, redundant tape drives and disaster recovery are seamlessly integrated. Storage management services for Microsoft Exchange, Microsoft SQL Server and Microsoft Clusters can be added without the worries of modular compatibility.

This white paper describes the benefits of TapeWare in a variety of network environments. Discover for yourself, backup and recovery solutions that make data storage administration easy and flexible. Find out how TapeWare can secure mission-critical data effectively and efficiently with minimal effort.

The Changing Nature of LANs
TapeWare was developed in response to the rapidly changing requirements for securing and managing network data. Industry wide changes in implementing LAN topologies have made it necessary to rethink the basic fundamentals of data storage management and security.

Storage management software has three basic functions:

  1. To secure network data, protecting it from accidental or intentional destruction
  2. To provide ongoing data storage and data security
  3. To facilitate data restoration after catastrophic loss
Network installations are increasing in complexity and size making comprehensive backup strategies more difficult to implement and control. When planning a storage management program, there are three main areas to keep in mind:

Data Set Size
Data can now range from a few gigabytes to terabytes. A magnitude of data this size has more room for error, making a partially implemented software solution unacceptable.

Data Transfer Rates
Large networks need to transfer data quickly. Software needs to work optimally with peripherals, limit unnecessary network traffic and implement schemes for fast continuous data transfer.

Network Complexity
Heterogeneous networks are commonplace, requiring backup software that performs seamlessly across mixed platforms. The changing nature of LANs demands that your backup solution be equally scalable.

Industry Response to Changing Networks
Networks are growing more complex, forcing backup applications to also grow in complexity. As new network topologies and platforms are added, they are often dealt with by developing new applications or components. Many industry software leaders offer an array of bolt-on applications and components. The strategy currently employed by many storage management software companies is comparable to a ‘cafeteria' model of software purchasing. Users are required to pick from an array of separately developed products and options to construct a custom software package that may meet their needs. As their networks expand and change, they must the try to find additional components to integrate with their existing software.

This approach introduces numerous disadvantages:

  1. Successful implementation often requires specialized knowledge
  2. Both the software and the personnel needed to administer it consume larger resources and operating budgets
  3. As the number of components grows, the integration of these components degrades.
TapeWare Response to Changing Networks
TapeWare has responded to the diversity of network installations with a software model that breaks away from the ‘cafeteria' style of choosing separately developed products and components which may or may not work seamlessly together. Due to development of a common code base, TapeWare works efficiently on all heterogeneous networks, regardless of platform or complexity. TapeWare is a comprehensive, scalable, all-in-one solution that supports Desktop, Peer-to-Peer and Enterprise environments.

Peer-to-Peer Advantage
The key to the superior TapeWare performance in heterogeneous network installations is the ability of each and every machine installation of TapeWare to perform complete backup and restore functions necessary to protect the entire network. Other backup software disperses the capacity to perform backup tasks across the network and rely on different software components for various network workstations and file servers. TapeWare empowers each installation to perform all the functions required to backup any network machine.

Each network installation of TapeWare has a peer-to-peer relationship with every other TapeWare installation, regardless of platform. All workstations and file servers are fully functional, although that functionality may never be implemented. This functionality is regardless of operating system or role in the network topology. Each installation has the ability to:

  • Control data transfer to and from a local backup device.
  • Control data transfer to and from remote network backup devices.
  • Control the transmission of data across the network.
  • Act as the Storage Management Server.
  • Serve as an administration console.
This peer-to-peer relationship between network machines makes installation flexible, adaptable and easier to perform, because there are no separate applications or bolt-on components to install.

TapeWare Features
Distributed Devices - Backup devices can be placed on any network machine, with the proper licensing. The network platform, operating system and client/server status of the machine does not affect the placement of any backup device.

Multiple Concurrent Devices - A single backup or restore job can transfer data to and from multiple backup devices concurrently across local buses and/or network connections. Single or multiple hosts can transmit data to multiple devices simultaneously.

Distributed Database Management Server - Any supported network machine can function as the Storage Management Server, regardless of network platform, operating system or client/server status.

Single Point of Administration - Run and monitor multiple jobs from a single workstation or server console or via a web browser.

Multiple Stream Processing - Process an unlimited number of streams simultaneously, with a maximum of eight per backup device.

Parallel Data Streams - Multiple data stream processing interleaves files at the block level (as opposed to the file level). This reduces network overhead for faster, more efficient data transfer.

Expert Database Management - Uses a model built around files and file instances rather than a ‘session' model. This allows intuitive database management and better file selection.

Advanced Reports and Diagnostics - The reporting features include the ability to list media needed for jobs, and reports that analyze the relative risk to the security of files and volumes.

Consistent, Standard User Interface – The TapeWare explorer-like user interface is standard across all Windows platforms. A similar XWindow user interface exists for Linux and UNIX environments. TapeWare also employs a web browser based user interface and a character-based interface for Netware, Linux and UNIX systems.

Powerful Scheduling and Rotation Manager - Schedule backup and restore jobs up to 20 years in advance with a drag-and-drop calendar. The integrated rotation manager automatically selects the proper media for each job.

Seamless Integration with Peripherals - New devices are automatically recognized and initialized.

TapeWare Storage Management
Learning how to use TapeWare is easy because of the cross-platform continuity, intuitive features and common interface. There are three main components defining a TapeWare installation:

Storage Management Database (SMD)
The Storage Management Database provides consistent, unified, and intuitive control over the various database objects (files, directories, volumes, network machines, backup devices, backup media, users, etc.). Objects are managed using an Explorer-like user interface and can be assigned user permissions individually.

Storage Management Server (SMS)
Also known as the Master Server, TapeWare allows you to designate any server or workstation on the network as the Storage Management. The Storage Management Server houses the Storage Management Database. (It is recommended that the SMS be a different machine than the Media Server or application server so that the Storage Management Server is still operational if one of the other servers were to crash.)

Storage Management Zone (SMZ)
The Storage Management Zone determines the set of network machines and devices to be included in the backup plan. The Storage Management Zone can manage any number of servers and devices.

Configuring TapeWare
Developing a comprehensive backup plan for this network includes several key decisions:

  • How many backup devices should be purchased?
  • What should their capacity be?
  • Where should the backup devices be located?
  • Which network machine should be designated as the Storage Management Server?
The answers to these questions depend on a number of variables, such as: The amount of data and how often that data needs to be backed up, and the amount of network traffic.

TapeWare is flexible enough to work with practically any developed backup plan. The network administrator can make decisions based on the amount of data, the bandwidth and speed of the network instead of any installation requirements of the software. The question of the number of backup devices and their location becomes a function of the optimal use of resources, not of the limitations of the storage management software. The following example shows some of the advantages of TapeWare. Imagine a small Ethernet network with two file servers (one NetWare server and one Windows NT server) and several workstations of various operating systems.

Here are some of the key advantages of using TapeWare:

  • TapeWare can work with a single backup device placed on one server or with a separate backup device for each server.
  • A single CD-ROM and license activation keycodes are used to install TapeWare on every machine. No distinctions between client/server machines are necessary and there are no separate applications or components to install.
  • TapeWare jobs can be created and run from any machine on the network on which TapeWare is installed. In fact, various users can be creating and managing various jobs from different machines simultaneously.
  • Efficiency of data transmission may mean that multiple backup devices are preferred on local buses (for example, one on each server.) The amount of data to be transferred and the speed of the network may be such that a single backup device (perhaps a faster one, with greater storage) is the best choice. Whatever decision is made, TapeWare adapts to either strategy.
  • When new machines, either file servers or workstations, are added to the network, TapeWare need only be installed on that machine in order for that machine to be capable of being backed up or restored.
  • As new backup devices are added, TapeWare automatically recognizes these devices and creates database objects for them.
  • TapeWare in a peer-to-peer configuration is powerful and flexible, and will work with almost any arrangement of backup devices.
TapeWare Data Routing
One of the most important features of TapeWare is how it automatically routes data as it reads volumes. Although you can specify precisely which backup device to use when running a job, TapeWare has built-in protocols that optimally route data in the most efficient manner.

Typically, local buses transmit data faster than network connections. TapeWare will initially attempt to transmit data to the backup device on the same machine as the volume is being read. When backing up multiple file servers to multiple devices, TapeWare attempts to route the data from each machine to its own backup device first before sending the data across the network. If the local backup device is unavailable, TapeWare scans the network for available devices, guaranteeing that the job completes successfully.

The Storage Management Database tracks and records what physical media stores a particular version of a file. When it comes time to restore a file, you choose the instance of the file to be restored and TapeWare automatically selects the proper media.

Even though TapeWare routes data to the optimal device automatically, devices can be manually controlled. If you specify a network container, TapeWare will use whatever devices are available on the network, preferring local devices first. If you specify a machine container, only backup devices associated with that machine will be used. If you specify a certain device or controller, TapeWare will only use that device for that job. Additionally, you can specify multiple devices or machines in this field, allowing maximum flexibility and control.

Disaster Recovery
The ability to restore lost data after a disaster is an important function of storage management software. TapeWare allows file servers to have attached devices and yet are not required to act as their own Storage Management Servers. This flexibility can make a big difference in the ease of recovering data and rebuilding an operational system after loss.

If the system acting as the Storage Management Server crashes, the Storage Management Database is also lost. The Storage Management Database is crucial to restoring the file server and must be restored before any machine data can be restored. The restore process can take even longer if the only backup device available is attached to the machine that has crashed.

With the TapeWare model, it is not necessary that the Storage Management Database be stored on the same machine as the devices they work with. With TapeWare, a separate network machine can be designated to serve solely as the Storage Management Server. This machine would serve no other role than to hold the TapeWare database. (This is not a requirement, although it may be prudent.) This machine doesn't require a high performance SCSI controller or network card, since no backup data is routed through it. It can have a medium sized hard drive attached (e.g., 9 gigabyte) and a fast microprocessor, but otherwise it is a relatively inexpensive machine. Indeed its role as a ‘server' is only to house the Storage Management Database. When a file server crashes, the Storage Management Database would still be in tact and its information is now available to be used for restoring the data lost on the file server.

How TapeWare Performs Disaster Recovery
With TapeWare, the disastrous loss of data could occur in one of two locations: data can be lost either on one of the network file servers or on the dedicated Storage Management Server. Here is a look at both scenarios:

  1. If one of the network file servers crashes, the Storage Management Server is still operational. The data can be quickly restored after the machine is repaired or can be temporarily restored to other network file servers until the repairs can be made.
  2. If the dedicated Storage Management Server crashes, the network file servers are still operational and important network operations are not impeded. The TapeWare database can be restored in a variety of ways, depending on the severity of the problem.

Consider the worst-case scenario in which the dedicated Storage Management Server is seriously damaged and cannot be repaired quickly. In this case, the best strategy would simply be to designate another network machine as the new Storage Management Server. This can be accomplished by editing the TapeWare configuration on the designated machine, or by installing TapeWare on a new machine and creating a new Storage Management Database with the same name as the old database.

Then, by using the Restore Database command to restore the lost database. After completing these steps, TapeWare will be completely operational again and you will be able to resume normal operations. TapeWare lets a remote network machine function as the Storage Management Server. By breaking the link between the physical device and the Storage Management Server, TapeWare is a powerful disaster recovery tool.

Competitive Scenarios
The following scenarios compare how TapeWare would be implemented versus its competitors.

Scenario 1
Assume a local area network with five file servers: two NetWare servers and three Windows NT servers. (Note: For the purposes of this scenario, we will focus only on the file servers and ignore the workstations.) In this scenario, we will look at three different configurations:

  • Typical Competitor Configuration - Single Backup Device
  • Typical Competitor Configuration - Multiple Backup Devices
  • Typical TapeWare Configuration
  • Typical Competitor Configuration - Single Backup Device
One way to set up a network backup plan is with a single backup device, setting up one machine as the Storage Management Server. In this case, either a Windows NT server or NetWare server can be designated as the Storage Management Server or Backup Server.

Competitor Disadvantages

  1. Whatever machine is chosen to be the Backup Server must have the backup device attached to that machine as well.
  2. The base software package to create the Backup Server must be purchased, along with an additional add-on module to support the other operating system.
  3. The data from each file server must be transmitted across the network to the one machine with a backup device. This limits the speed at which data can be sent to the backup device. If all of the file servers were backed up in a single job, the limitations of the network to transmit data would create a huge bottleneck.
  4. As data size on remote servers increases, storage capacity of the backup device must increase. It may not be possible to store all of the data on a single tape. Although autoloaders can be used, the limits of these devices will be quickly reached.
  5. The time it takes to complete backup jobs could be very lengthy. These jobs need to be run when there is no other network traffic. Scheduling problems would be difficult if these servers were used after normal business hours. And, if a job fails to run when scheduled, it might not be possible for it to run until the next available time slot.
  6. Expanding this network by adding additional servers would create more bottlenecks. The addition of new servers might require redesigning the network backup plan.
Typical Competitor Configuration - Multiple Backup Devices
An alternative way to implement a network backup plan is with multiple backup devices, attaching a device to each file server. In this case, each Windows NT server and NetWare Server has a backup device attached to it, and thus each one is also designated as a Storage Management Server.

Competitor Disadvantages

  1. Since there are five different Backup Servers, a software license agreement must be purchased for each file server.
  2. Administration is difficult. There are now five separate Backup Servers. Each one handles its own jobs and files, requiring five different job schedules. After the jobs are run, there are five different job logs that must be checked to ensure that each job has run correctly.
  3. It is not possible to manage all five servers simultaneously. In order to manage each server, you have to log on to that server's storage management software separately and individually. This process is inefficient and increases the likelihood for human error and risk to the network data.
  4. Adding backup devices does not provide redundancy or protection during device failure. Some software prohibits servers from transmitting data across the network to other servers and devices.
  5. The software necessary to implement this configuration is expensive. A license is required for each file server.
  6. This configuration does not lend itself to easy expansion. When new file servers are added, they must either be attached to existing Backup Servers or added as standalone Backup Servers with their own separate device.
  7. There is no easy way to transfer backed up data from one file server to another. File servers do not recognize backup devices on other file servers. Similarly, physically moving the tape from one backup device to another backup device will not work. The new Backup Server does not recognize this media without first "importing" the media.
Typical TapeWare Configuration
In this configuration, each machine has its own backup device. One Windows NT server has been designated as the Storage Management Server. (Note: Any machine on the network, including any NetWare server, can be designated as the Storage Management Server.)

TapeWare Advantages

  1. With the Universal Product Edition of TapeWare installed on each machine, there is no need to buy add-on modules for the Linux or NetWare servers.
  2. Administration is easy. There is a central point of administration for all file servers. It is possible to create a single job that backs up the data from each server. This means only one log to check and only one schedule to maintain. Of course, jobs that work with only a single server can also be set up. Administering these jobs is also easy since it can be done simultaneously for every machine.
  3. These jobs can be administered from any machine on the network. Multiple users can be creating and running jobs from various network workstations simultaneously.
  4. The rate of data transmission is maximized. Because each server has a backup device located on a local bus, TapeWare will by default route data from each server to its local backup device.
  5. Each server can have a backup device with a capacity that is matched to that server's particular requirements.
  6. Additional devices on the network function as a redundant, protective layer. When a local device fails, TapeWare will automatically reroute the data across the network to other available devices. This greatly increases the prospects that server data will be safely backed up and that jobs will be successfully completed.
  7. Expansion of the network is simpler. Adding new file servers does not require the purchase of additional software. The same single point of administration can now work with this new file server. You can purchase a dedicated backup device for the new server as needed or, if appropriate, use the existing devices on the network.
  8. TapeWare will automatically recognize any new machines or devices that appear on the network. It will also flag any devices or machines it was expecting to find but didn't, allowing the administrator to check at a glance that all of the machines and devices are on-line.
  9. Installation and implementation is faster and easier. License activation keycodes are used to install software on each machine, using the same CD-ROM for Windows, NetWare, Linux and UNIX platforms. No separate applications are purchased. Additionally, jobs and reports can be run for multiple servers, cutting the time it takes to implement and manage a comprehensive backup program.
Optimized Interface with Peripherals
TapeWare employs a number of features to maximize the interface between network peripherals, such as volumes, machines, and backup devices. Combined, these innovations exploit all of the capacities of current network and small computer technology. There are several goals here: speed is a primary concern, but so is ease of use and installation. Some of the features used to meet these goals include:
  • Multiple Data Streams
  • Block Interleaving
  • Flexible Data Stream Format
  • Automatic Recognition of Network Peripherals
  • Multiple Data Streams
TapeWare is capable of handling an unlimited number of data streams simultaneously, with a maximum of eight streams per backup device. Each stream represents a logically connected set of data.

For example, most volumes are treated as a single stream. This feature allows multiple drives to be simultaneously sending data to the backup device, resulting in faster backup. The number and assignment of data streams is flexible and can be controlled by the administrator. For example, TapeWare allows you to assign multiple data streams to a single volume. This is useful in the case of large RAID devices that are capable of generating and sustaining multiple data streams simultaneously. In this case, administrators can assign new data streams to various directories on the RAID device. Similarly, greater data transfer speed can be realized by assigning large files (i.e. five gigabytes or larger) their own data stream. On the other hand, sometimes logical volumes (as opposed to physical volumes) should not be assigned a separate data stream. In either case, the administrator can control whether or not a volume, directory, or file has a separate data stream.

Block Interleaving
When multiple streams are being processed at once, the most efficient way to write data to the tape is to interleave that data on the tape. In the same fashion as disk drives function, logically contiguous data need not be physically contiguous. The ability to interleave data is important because it means that data can be written to the tape as it becomes available. When there is data in the queue ready to be written, there is no need to wait for a file to be completely written to the tape (when that data has yet to be read from the volume) before writing the data that is currently available from other files to the tape.

TapeWare employs a block interleaving strategy, as opposed to a file interleaving strategy. With file interleaving, files are written to the tape in physically contiguous sets of blocks. Files from different volumes may be interleaved, as that data becomes available, but each file is written in physically contiguous blocks.

While a file interleaving strategy increases efficiency, the TapeWare block interleaving strategy is even more efficient. By interleaving blocks from different files, data can be written to the tape as it becomes available. For example, multiple data blocks from multiple files can be interleaved, meaning that the data can be written as it becomes available. Less time is wasted waiting for network transmissions and instructions to be completed, since data from other streams can be written while waiting for additional data blocks from a particular file to be transmitted.

Second, the size of TapeWare data blocks is more efficient. Network backup software programs that interleave data blocks on tape use 32K blocks. There are three problems with using blocks of this size. First, quite often the size of the data to be written to the tape is smaller than 32K. This results in wasted space on the tape because part of the 32K block will be empty.

Second, because part of the tape is empty, no data is being written to that part of the tape as it passes the write head—effectively meaning that the rate of data transmission for that time period is zero. Overall, the rate at which data is written to media is degraded.

Third, a 32K block is many times larger than the size of a data packet being sent over the network, requiring multiple network transmissions before a block can be written.

TapeWare, on the other hand, writes data in 32 byte blocks. These blocks are a more efficient use of the capacity of the tape, resulting in fewer ‘dead' spots on the tape and increasing the rate at which data is written to the tape. The exact amount of efficiency gained by the TapeWare strategy depends on particular network installations. In general, it results in more efficient use of the tape capacity and faster backups when there are numerous files smaller than 100K in size. While the exact performance enhancements vary, this is only one example of how TapeWare has been optimized to achieve maximum performance.

Flexible Data Stream Format
TapeWare supports full backup and restore functionality for mixed environments running a combination of Windows 98, Me, NT 4, 2000, XP, Server 2003, NetWare 4.2, 5.1, 6.x, Linux and UNIX operating systems. These products, however, transmit data across the network in different formats. The default format that each platform uses to transmit data is called the native data stream format.

TapeWare can write and read data in either the native data stream format or in a general, cross-platform format. Each format has its own advantages. For example, using native data streams results in faster backup jobs. The data need not be reformatted into a generic format, which reduces processor overhead.

Many times this option is the most efficient and best way to write and store data. However, what if you want to transfer data from one platform to another, for example from a NetWare server to a Windows server? In this case, the TapeWare generic stream format can be utilized. When the files are backed up, TapeWare can translate them into a generic format. After they have been backed up, they can be restored to the new server.

The necessary software to transfer files back and forth from various platforms and operating systems is incorporated into the TapeWare software installed on every machine. These are not separate applications or add-on packages, nor are they installed on some machines, but not on others. Each installation of TapeWare on every network machine is given full functionality in order to create true a peer-to-peer relationship.

Automatic Recognition of Network Peripherals
As long as TapeWare is installed on a network machine, every other network machine will recognize that machine and its peripherals. TapeWare checks to see which database the machine belongs to and only those machines which are being managed by the same Storage Management Server, are recognized by each other.

This process is automatic. For example, to add a new machine to the database managed by a Storage Management Server, you need only install TapeWare on that machine and indicate that you want it to be the member of the pre-existing database. When TapeWare is later run, the Storage Management Server recognizes that a new machine running TapeWare is on the network and creates a database object for that machine.

The recognition of peripherals is similar. TapeWare automatically scans for attached peripherals such as volumes and backup devices. If it finds these new objects attached to the machine, it creates new database objects for these peripherals. TapeWare does not use, nor does it require any special proprietary drivers or controllers in order to work with backup devices. Users need only install the backup device on the workstation or file server using the manufacturers required installation procedures. If the machine recognizes the backup device, it is virtually certain that TapeWare will recognize that peripheral as well.

Optimized Management Tools
This section concentrates on three of the most important features of the TapeWare user interface:

  • Reliable and Flexible Scheduling
  • Reports and Diagnostics
  • Familiar Consistent User Interface
Reliable and Flexible Scheduling
In order for data to be secured, regular backup jobs must run, using a proven rotation schedule. TapeWare provides a number of built-in schedules that insure the ability to reconstruct data after a disaster. The schedules vary according to the number of tapes they use, the length of the data recovery period, and access to historical data. Each schedule provides assurance that sensitive data has been securely backed up.

Even though a regular backup routine is necessary, there must be flexibility to change the routine without compromising integrity. TapeWare accomplishes this by utilizing both a built-in rotation manager to handles routine jobs and a Schedule feature that gives the user unprecedented control over customizing how and when jobs are run.

After a job is assigned one of the built-in rotation schedules, the TapeWare rotation manager automatically schedules the necessary backup jobs. Jobs are automatically run at the date and time scheduled. The rotation manager controls whether the jobs are full, incremental, or differential, and identifies the proper media to be used with the job. New files, directories, and volumes are automatically selected on machines that are specified for backup. Working seamlessly with autoloaders, it manages the rotation of media to ensure that only the correct (e.g., oldest) media is overwritten.

Built-in schedules can be easily modified and customized when needed. Changing when jobs are run or what type of job is run is as easy as drag-and-drop or bringing up a shortcut menu. This allows the user to customize the schedule to account for holidays and other unusual circumstances, such as special business cycles, while not damaging the basic integrity of the automatic rotation schedule.

Reports and Diagnostics
TapeWare includes a number of useful reports that make it easier to manage a complex network installation. For example, users can generate an "Objects at Risk" report. This report analyzes the level of risk for each file in the database by determining if sufficient backups have been made for these files.

This information can be invaluable when working with sensitive, vital data. Another report prints the contents and history of media. Job logs for audit and tracking purposes are also available.

There are a whole series of built-in diagnostic tests that can be used to test the interaction between network peripherals. For example, you can test the communication between machines or the speed at which network data is being sent. Other diagnostic tests assess whether autoloaders are working correctly and if media is in usable condition. These diagnostic tests help the TapeWare administrator rule out possible causes of problems that have developed. This makes it easier to pinpoint with accuracy the location and source of any equipment malfunction or breakdown.

Common User Interface
TapeWare provides a consistent, well-designed user interface across all supported platforms. The graphical user interface is based on Windows Explorer.

The user interface is built around an object/properties model, arranged in a hierarchy. Database objects are contained in container objects, which are arranged in a hierarchical order. Each object, including container objects, has a property sheet that is used to control its properties and functionality. Manipulating the properties of database objects is easy and consistent for all objects.