1. Backup and Restore Backup and Restore Introduction
In this lesson you will be introduced to the concepts of backup and restore.
Importance
Backup and restore are two of the applications of Storage Area Networking (SAN) that allow businesses to run mission critical applications. If the Systems Engineer (SE) is to effectively sell to storage customers, it is important that they understand when a customer’s business applications can benefit from backup and restore solutions.

2. Lesson Objective
Upon completion of this lesson you will describe the basics of enterprise backup and restore solutions
Performance Objective
Upon completion of this lesson you will describe the basics of backup and restore solutions.
Enabling Objectives
Describe the need for backup and restore solutions
Describe the concept of backup and restore
Explain the challenges to data backup
Explain different strategies used to provide backup solutions
Compare tape and disk as backup media
Explain different methods used to achieve backup solutions
Explain different levels used to achieve backup solutions
Compare the recovery time for the different backup levels

3. Lesson Outline The Need to Backup Backup and Restore Backup Challenges
Backup Architectures/Strategies
Local Backups of individual servers
Network (LAN Based) Backup
LAN-free Backup
Backup Media: Tape vs. Disk
Backup Methods
Off-line Backup
On-line Backup
Raw Device Backup
Backup Levels
Backup Levels and Recovery Times
Prerequisites
Completion of the Unit 1 prerequisites

4. The Need to Backup Human errors Disasters
70%+ of restore requests result from user file errors
Disasters
Natural
Earthquakes, fires, floods, hurricanes, tornadoes
Man-made:
Computer viruses, angry employees, industrial accidents
Electromagnetic disruptions, criminal/terrorist activities
Equipment failures and malfunctions
Product failures, transmission errors
System Maintenance
Upgrades, Data restructuring, system cloning
The Need to Backup
Objective
Describe the need for backup and restore solutions.
Introduction
Businesses are becoming increasingly more dependant on their data. Organizations that lose their data may lose millions of dollars or, worse yet, lose their business completely. The ability to backup and restore information in many cases may determine a business’s ability to survive following a disaster.
Facts
Several factors motivate the use of backup and restore technologies. The first is protection from failures. Losing the physical media used to store the information can have disastrous consequences, but at least these failures are noticed rather quickly. Protection from logical failures is a little more tricky. It may take hours or even days before this type of failure is even noticed. At that point, data must be restored to some point in time prior to the error taking place. During the period between the error’s occurrence and it being detected, valid data may be created. The restoration process must be able to restore older information and accommodate the newly created valid data as well.
Problems with data protection and integrity are common, not an exception. Problems may be created by human error, nature or failed equipment and devices. When (not if) they occur, valuable information may be corrupted or lost.
A good backup strategy is designed to minimize the impact of those problems and return systems to an operational state.
Continued…

5. Secondary Tape Storage
Backup and Restore
Secondary Tape Storage
Primary Disk Storage
Server System
2. Read from tape
3. Write to disk (through the file system)
Backup software
Backup
Database
1. Data’s tape location found in database
Backup and Restore
Objective
Describe the concept of backup and restore.
Introduction
Backup is a commonly used method to protect against disaster. Backup is used to copy information (data) from a primary storage device to a secondary storage device.
Restoration of data involves the copying of data from secondary to primary storage.
Definition
Backup is the creation of a duplicate copy of data on a secondary storage device.
Restore is the ability to retrieve data from a secondary source and use it to bring the primary system(s) back on line.
Example
Backups would not be done if there were no requirements to restore data.
The most common type of restore request is to restore a single file for a user. Sometimes directories or volumes are requested. The backup database is practically a necessity for locating a tape with the desired file version.
In a disaster recovery scenario, all the system’s data needs to be restored, including a lot of system data having very little to do with “user space” data.
Restored data has to be processed by the filing system in the host machine, otherwise there is no way of knowing that the data exists in the system.
Continued…
Basic Restore Process

6. Backup Challenges Historically, tape drives were not shared
Net backups solve that problem, but impact LAN performance
Shrinking backup windows
Backup Challenges
Objective
Explain the challenges to data backup.
Introduction
One of the primary enterprise information systems’ responsibilities is to ensure that company data can be recovered in the event of data loss or corruption. This section explains some of the several challenges they face.
Facts
Historically, tape devices attached to individual servers were the primary method used for data backup and recovery. This was usually considered an unappealing solution as tape devices are expensive and often sit idle when dedicated to a single system. In addition, significant management resources are required to organize and store tapes generated on multiple servers. 
Network backup enables the sharing of tape devices among many servers. However, such a strategy can negatively impact network performance.   
Another challenge faced by backup administrators is finding sufficient time, referred to as the “backup window”, when servers can be taken offline to perform a backup while data is not being modified. As enterprise computer systems often need to provide 24x7 service, identifying a sufficient back-up window can be difficult, if not impossible, using traditional backup methods.
Practice Items
What is a backup system? (software, hardware and media working together)
What are some of the problems facing backup today?

7. Backup Architectures/Strategies
Local Backups of Individual Servers
Devices located on each system to be backed up
Network (LAN based) Backup
Centralized backup devices accessed via LAN
LAN-Free Backup
Backup traffic moved off the LAN
Backup Architectures/Strategies
Objective
Explain different strategies used to provide backup solutions.
Introduction
There are several architectures or strategies that can be used.
Facts
The various architectures include:
Local application Server Backup
Network (LAN based) Backup
LAN-Free Backup
Serverless Backup

8. Backup Architectures/Strategies (cont.)
Application Servers
Backup Architectures/Strategies (Cont.)
Facts
In local backup the devices are located on each system to be backed up. The most popular choice is tape. Depending on the size of the tape drive and the amount of information requiring backup, an auto-changer may be used. Local backup has several disadvantages including the administrative nightmare associated with multiple devices, organization, categorizing and deployment of tapes as well as the removal of said tapes to an off-site location. Furthermore, there tends to exist a higher quantity of lower quality devices used in these implementations. This may have a dramatic effect if the information needs to be restored. The necessity actually to perform the backups with little to no automation is a rather undesirable quality as well. Local backup devices are usually lower quality, providing about 3MB/s. At this rate, 50GB of data would take 4.6 hours to backup. Finally, because the entire process interferes with user access, local backups are used on a very limited basis.
Local Backups of Individual Servers

9. Backup Architectures/Strategies (cont.)
Server with Backup Software
Application Servers with Backup Agents
LAN (slow)
SCSI Backup
Devices
Backup Architectures/Strategies (Cont.)
Facts
In a network (LAN based) backup, the backup devices are centralized and accessed via the LAN. This provides better utilization of backup devices and, because fewer higher quality devices are used, the backups tend to be more reliable and provide greater automation capabilities. Disadvantages include:
The increase in network traffic as backup information is sent across the LAN
The process still interferes with user access
A network backup involves the use of many entities, including a backup server (includes a media server) and a data source (Client)
Network backup involves parallelism. This is the ability to run multiple save streams to a backup server from different clients. Often the central backup device will combine data from multiple sessions with multiple data sources to a single tape—known as data interleaving. With data interleaving, it is necessary to keep the tape streaming because, if it stops, the tape rewinds and when it tape begins to write again it has to scan forward.
The ability to gather information from multiple sources greatly improves device utilization but is still constrained by bandwidth issues on the performance front. The LAN may be able to provide 2-3MB/s, if not used for other activities (during an off peak period). At this rate, 1 terabyte would take 92 hours to backup. While compression could reduce the overall time, the current computing environment simply cannot provide that level of service window. To reduce the LAN traffic problem, a dedicated backup LAN can be built (still maintaining the same bandwidth bottleneck) or a separate network can be utilized.
Network (LAN Based) Backup

10. Backup Architectures/Strategies (cont.)
Application Host
Backup Server
Tape Library
Corporate LAN
Backup data flow
Backup Network
Backup Strategies/Architectures (Cont.)
Facts
Many companies do not want to impact their Corporate LAN and build out a Backup/Restore LAN. LAN-Free Backup has been made possible through advances in storage networking technology. The overall time required for full backups is improved moderately by Fibre Channel, although scheduling itself is no longer dependent on or interferes with LAN traffic patterns. The greatest limitation in network based backup is bandwidth. Typical LAN speeds are 100Mb/s. When each of the servers is provisioned with a Fibre Channel HBA, there are options for reducing backup times.
Practice Items
What is an advantage of shared drives over dedicated drives?
LAN-Free Backup

11. Backup Media: Tape vs. Disk
~5-20 MB/second streaming performance
Multi-terabyte capacity libraries
Removable off-site protection
Managed by backup software
Disk
Faster access and transfer times
Random access
Non-removable local or remote sites
Managed by file systems or databases
Backup Media—Tape vs. Disk
Objective
Compare tape and disk as backup media.
Introduction
This section compares the use of tape and the use of disk as backup media.
Facts
Historically, backups have been made to tape storage devices. Tapes provide good performance for reading and writing, and the ability to accommodate a large amount of data. The main advantages of current tapes is their removability, and ability to hold multiple versions and copies of the same data. Managing tapes is an expensive proposition as backup administration can consume considerable resources.
Tapes must be inventoried, scheduled and used in the proper sequence to make sure backup data is preserved properly.
Tapes are stored at both onsite and offsite locations, which creates ongoing operational problems.
Tapes have a shelf life, which is determined by the manufacturer, and must be refreshed after a certain time interval. For heavily used tapes this may be months or years.
Tapes store data by creating magnetic images on small metal particles spread across the surface of the tape. This metallic material corrodes and fractures.
Tapes do not survive long in high humidity, high heat, in environments heavy with airborne contaminants or in environments with heavy electromagnetic radiation.
All tapes shed their magnetic layers, which turns to dust and coats the heads of the tape drives. Therefore, tape drives need to be cleaned on a periodic basis.
The majority of new tape technologies are fairly advanced and work well if they are well-maintained, but the reliability of tape should never be taken for granted.

12. Backup Media: Tape vs. Disk (cont.)
Disk Drives for Backup
Challenges:
Backup software needed
Disk drives not removable
Keep only current versions of data
Uses:
Store current snapshots
Backup to tape without affecting on-line systems processing
Backup Media—Tape vs. Disk (cont.)
Disk drives are growing in popularity as a backup medium. Disk capacity now exceeds tape capacity, and disk access times (seeks) are faster than tape.
Disk transfer rates are ~ 2x the fastest tape drives
Disk drive costs are quickly approaching tape costs
Disk backup is being promoted by some as an important new technology, but first software has to be written that can integrate disk effectively with 100% accuracy. The challenge of succeeding with this is considerably more difficult than it appears. Backup software is written for tape. Backup to disk is seeing heavy development. Disk drives would be used universally for backup, if not for two important limitations:
Generally, they are not removable
They are not used to store multiple versions of files
Restores demand multiple versions of data, but on-line disk storage keeps only the current version. Therefore, backup to disk is uncommon. Disks are being used as part of backup as a way to store current snapshots of data where they can be backed up without affecting on-line systems processing.
Enterprise backup can easily involve thousands of tapes. It is not hard for an enterprise backup environment to include in excess of 10,000 or even 25,000 tapes. Tape automation significantly reduces the overhead of managing tapes and eliminates the inevitable administrator errors.

13. Backup Media: Tape vs. Disk (cont.)
Facts
There are 2 classes of tape automation:
Autoloaders: Typically single job devices that often need attention on a daily or every other day basis
Libraries: Can be fairly large, holding hundreds of tapes and 4 to 8 tape drives
Practice Items
What are the main disadvantages of data tape as a backup solution?
What are the main advantages of data tape as a backup solution?
What is the main challenge facing disk as a backup medium?
Tape Automation
Robotic tape handling saves time
Automated tape selection reduces errors
Tape autoloaders:
6 to 12 slot tape magazines or up to 20 tape slots
1 or 2 drives
Tape Libraries (also called tape subsystems)
20+ tape slots, 2 or more drives
Bar code readers
Input/output doors and slots

14. Backup Methods Off-Line (cold) backup
Must be done within the backup window
On-Line (hot) backup
Ignores the backup window, but creates heavy system resource conflicts
Difficult to maintain data integrity
Raw device (image) backup
As cold as it gets
Difficult to restore individual files
Backup Methods
Objective
Explain different methods used to achieve backup solutions.
Introduction
There are several methods used in backup solutions. Each method carries unique characteristics, advantages and disadvantages.
The real problem in enterprise backup is getting the job done in a small amount of time. In general, the time available to accomplish the job is continually decreasing while the amount of data is continually increasing.
If backup is not finished on time, systems managers have difficult choices to make: Either stop backup processing prematurely, or delay starting production systems until the backup is complete—potentially costing revenue.
Facts
The list of different methods of backup include:
Off-line
Online
Raw device (image) backup.

15. Off-Line (Cold) Backup
Backup Methods (cont.)
Off-Line (Cold) Backup
System I/O is stopped
File system or database activities are quiesced
Guaranteed integrity of copied data
Fastest data copies—less I/O contention
Not-applicable to 24x7 system availability
Backup Methods (cont.)
Facts
Off line backup is the traditional method. Systems are quiesced, and without any other I/O activity there are no concerns about data integrity or system resource contention.
In off-line backup the server goes off-line and is unavailable to the users. This period is referred to as the “service window”. This method is a good choice for large databases that can be taken off-line for a short period of time. As the resources of the involved systems are not being used by the application, this backup can outperform on-line backups.
Traditionally, backup consists of shutting down the application, performing the backup operation and restarting the application. However, for applications that cannot be taken off-line then some other method must be used.

16. Backup Methods (cont.) On-Line (Hot) Backup For 24x7 environments
System I/O continues during backup copies
Copy on write process maintains integrity
Updates are spoofed to new locations
Intercepts kernel/file system interaction
Hot backup slows system performance
Data copies, COW and backup database processing can saturate system processors
Backup Methods (cont.)
Facts
On line backup is a method developed primarily for database backup, but is being used more frequently for other systems with shrinking backup windows. It has to maintain dual block locations for file updates that occur while the file is being backed up. In general, this means making changes to low level system functions. System performance can suffer significantly during on-line backups
Online or Hot Backup is the least intrusive strategy, and is a popular solution for databases that must be available 24 hours a day.
There are two integrity issues in hot backups:
Each file or database entity needs to be backed up as a complete consistent version.
Database entities that have correlated data versions must be backed up as a linked group. Write access to this information must be altered during the time the backup is taking place on that particular entity. This could be accomplished by simply not allowing access (causing a brief moment of inaccessibility) or with a patch file (in the case of an Exchange Server).
The patch file serves as a temporary storage for changing information. When the backup is complete, the patch file changes are applied to the database.

17. Raw Device (or Image) Backup
Backup Methods (cont.)
Raw Device (or Image) Backup
Bypasses the file system or database
Like cold backup, requires no I/O activity
Directly copies disk blocks to tape blocks, sequentially
No knowledge of data content
Cannot restore individual files or database tables
Backup Methods (cont.)
Facts
Raw device, also called image backup, can only be done when the system is cold. In addition, the ability to restore individual files is much more difficult. Usually, image backup is used for bare-metal restores. Raw device backups are the simplest way to do database backups as they directly copy the raw disk devices to tape.
They do however require the database to be in a quiescent state so they require a service window (similar to off-line backups). These backups are fast because the reads are sequential (file by file) and no database involvement is required. This bypasses the file system and backs up the storage device directly. It is equivalent to the offline backup of a file system which has its applications quiesced.
Note: Only databases which use “Raw partitions” should use this method, as the backup server cannot restore parts of the partitions (as in the case of file system backup).
Practice
Name three methods of backup
Describe the three methods of backup
Provide an advantage and a disadvantage of each of the three backup methods

18. Backup Levels Full: Incremental: Differential:
Backup whole volume or file system
Requires the most backup time
Requires the fewest amount of tapes for restoration
Limited to weekends
Incremental:
Backup only the changes since the last backup
Differential:
Backup only the changes since the last full backup
Backup Levels
Objective
Explain different levels used to achieve backup solutions.
Introduction
There are three levels of backup. Organizations typically use a full backup with either incremental or differential backups in between. This section explains the levels of backup.
Facts
Backup levels determine the amount of information that is backed up and have a lot to do with backup and restore performance. The most popular levels include full, incremental and differential backups. The level determines the amount of backup information that will be processed.
The most common is a full backup with either differential or incremental backups in between. In the event of some problem that requires a restoration, the choice of backup greatly determines the speed at which systems can come back online.
A full backup will backup the whole volume or file system. This level requires the most amount of backup time but would require the least amount of tapes for restoration. Unfortunately, due to shrinking backup windows this level is not usually possible for most organizations to implement everyday. Organizations typically use a full backup with either incremental or differential backups in between.

19. Backup Levels (cont.) Sat Sun Mon Tue Wed Thu Fri
Fact
The size of an incremental backup depends on the files that have changed since last incremental or full backup.
Incremental backups only copy the newly changed data that has been created or updated since the last backup operation (previous incremental) or the full backup operation. The main advantage is to perform the shortest possible backup. However, the restore process requires the full backup and all the incremental backups since the last full backup, which takes a lot longer time to restore.
Saturday Full Backup
Daily Incremental
Backup
Weekly Incremental Backup Schedule

20. Daily Differential Backups
Backup Levels (cont.)
Mon
Tue
Wed
Thu
Fri
Sun
Sat
Daily Differential Backups
Weekend Full Backup
Backup Levels (cont.)
Fact
The size of a differential backup depends on the files that have changed since last full backup, rather than just the last backup.
Differential backup is similar to incremental backup. However, differential backups combine all changed data since the last full backup, as opposed to the last backup.
Daily backup operations take longer to complete every day.
The restore process is less complicated and faster than incremental backup because fewer tapes are needed. To restore, simply use all the tapes of the full backup and the tapes of the most recent differential backup (not all the differential backups).
Practice Items
What are the different backup levels?
What is the difference between incremental and differential backups?
Weekly Differential Backup Schedule

21. Backup Levels and Recovery Times
Full Backup
Longest time to backup, shortest time to restore
Full with Incremental
Shortest time to backup, longest time to restore
Full with Differential
Best balance of backup vs. restore times, depending on full backup interval
Backup Levels and Recovery Times
Objective
Compare the recovery time for the different backup levels.
Introduction
Restore times depend heavily on the schedule or level used. This section compares the recovery times for various backup schedules.
Facts
The backup schedule affects restore times. A full restore from an image requires that the full range of disk blocks are restored. At the file level, this means restoring the entire baseline volume. While the full backup takes the longest time to backup, it provides the shortest time to restore, which may be of significant benefit to organizations faced with rising costs of downtime.
As small backup windows often prevent daily full backups, an organization often chooses a full with incremental backup. This requires the restoration of the full backup and all subsequent incremental tapes. While this provides the shortest time for daily backups, it provides the longest time to restore.
Restoration of the full with differential requires the restoration of the full backup and the most recent differential. This probably provides the best balance of backup versus restore times, depending on the full backup interval.

22. Lesson Practice Develop a “Cadillac” backup and restore solution.
Backup solutions must consider:
Media
Method
Level
Strategy
How does your solution address:
Time pressures
Data integrity
Practice - Backup and Restore
Describe the basics of enterprise backup and restore solutions.
Your client wants to have the most effective enterprise backup solution. The enterprise operates 24/7, and customer service is key. As a result, cost is not an issue. Develop a “Cadillac” backup and restore solution, explaining your choices. Consider all the items listed.
Additional Practice Questions
1. Describe challenges associated with the following three backup architectures or strategies: Local Backups of Individual Servers Network (LAN based) Backup LAN-free backup
2. A company performs a full backup of its primary server on Sunday, and incremental backups at the end of each business week day. Friday at noon a drive fails. How many tapes need to be read to restore the disk?
3. A company performs a full backup of its primary server on Sunday, and differential backups at the end of each business week day. Friday at noon a drive fails. How many tapes need to be read to restore the disk?

23. Summary This lesson presented these key points:
Enterprises must backup to survive disasters
Backup solutions must consider:
Media
Method
Level
Strategy
Time pressures and data integrity are key to backup strategies
Server-free Backup provides the best overall performance
Summary: Backup and Restore
Review
Backup is necessary to protect an organization data from many potential sources of destruction
There are several levels, methods and strategies for implementing backup and restore solutions.