Part 3: Disaster Recovery Backup and Restore Media
Part 3: Backup and Restore Media
In the last post I touched on a couple popular schedules for backing up data, but I didn’t mention what the data is getting backed up to. I also introduced the 3-2-1 rule for backups, here is a quick reminder on that rule:
- There should be three copies of all data
- Production copy
- Local backup copy
- Remote backup copy
- At least two types of media should be used
- Tape / VTL
- One copy should be remote
- To protect against catastrophic failures at the production site. This can play a role in disaster recovery.
So, let’s take a look at the different types of backup and restore media that are used for the 3-2-1 rule, where they are typically used, and their pros and cons. I will start by providing a high level table with this information, and then go into additional detail.
To meet the 3-2-1 rule you can use any combination of backup and restore media, as long as two copies of the data are local, and one is remote. However, there are practical considerations, such as not using tape for production storage and taking steps to protect your backup and restore media from malware attacks. Typical implementations use relatively fast storage for the production data, such as tier 1 or tier 3 disk since it provides the greatest performance for production workloads. It is impractical to use high speed disk for the local backup or remote backup copies due to cost. Any of the other storage media can be used for these copies. Typical configurations are:
- Disk-to-disk-to-tape where the tape copy is made locally but taken off site.
- Disk-to-disk-to-disk where the third disk copy is offsite and populated across a WAN connection. Both the second and third disk copies are to lower speed/less expensive disk.
- Disk-to-disk-to-cloud where the second disk copy is to a lower speed/less expensive disk.
- Disk-to-VTL-to-VTL where the second VTL is offsite and populated across a WAN connection.
- Disk-to-disk-to-VTL/cloud where the VTL is acting as a cloud gateway.
Backing Up to Disk
Due to the large amount of data that typically needs to be backed up, and ever shrinking backup windows, many organizations are moving to some sort of disk-to-disk-to-X backup. Instead of backing up from the servers directly to tape, the servers are backed up to a pool of disks connected to the backup server, which is much faster than traditional tape drives. The disk pool for backup is sized so that it can hold multiple backups depending on the organizations retention policies. Backing up to disk has several benefits including:
- Reduced time required to complete backup by going to faster storage media (disk)
- Allows multiple simultaneous backups to occur
- Allows post processing of the backup data
- Fast synthetic full backup builds
- Fast restores from disk
The drawback of a pure disk-to-disk backup is that you don’t have removable media that can be taken off site. That is why most backup and restore systems give you the option of creating a third copy (original data on servers, backup, off site backup). Many backup systems allow disk-to-disk-to-disk, where you replicate the data that has been backed up to disk to another disk pool which can be located at a remote site. Another option is disk-to-disk-to-cloud, which is a version of disk-to-disk-to-disk, where one of the disk pools is provided by a cloud provider as a monthly service.
Disk-to-disk-to-cloud solutions have many benefits over solutions that utilize tape including:
- Tapes are vulnerable to damage from poor environmental conditions, overuse and age, or being dropped. Cloud storage services provide 99.999999999% data durability (eleven 9s).
- There is no cost other than internet bandwidth associated with transporting data to the cloud. Tapes require physical transportation off site and proper storage conditions, which are typically provided by a third party for a monthly fee.
- Tapes require management and organization to make sure the proper tapes are loaded for each backup and restore.
- Tapes may need to be transported back to the production site to restore files, which incurs delays and costs. Cloud-based storage is accessible 24x7x365.
- Most tapes in a tape-based backup solution are only partially filled with data, leaving wasted space. Cloud-based storage is an operational expense billed on utilization – $/GB/month – and only bills for storage used, so there is no wasted space.
- Tapes ultimately need to be destroyed to ensure that the data they contain is destroyed. Cloud based storage is encrypted, and can be deleted/overwritten to destroy the data.
- Storage and technology upgrades are included in the price of cloud storage services. Tape systems require periodic technology upgrades including replacing tape drives and tapes and require maintenance contracts.
Total cost of ownership (TCO) calculations should be done to determine if cloud-based storage solutions are right for your organization. The TCO calculation should include:
- Tape media, including unused space on tapes vs. cloud storage costs
- Tape infrastructure and maintenance costs
- Tape handling and storage fees
- The benefits of higher data durability in the cloud
- Data destruction costs
- The value of speed of recovery from the cloud
- Data retrieval fees for cloud data
- Internet bandwidth needed for cloud storage
- Backup retention policies
Storing backups to the cloud is not the solution for everyone, but it can be an attractive solution for many organizations. Cloud-based backups can also be used to help meet many compliance regulations and form the foundation for a disaster recovery to the cloud solution.
Another option, mentioned above, is the use of a virtual tape library (VTL). A VTL is simply a pool of disks that presents itself as one or more tape libraries, which each have one or more tape drives. The disk itself is used to present virtual tapes to these tape libraries. This allows traditional tape backup solutions, which only understand how to work with tape drives, to now store data to a disk pool. VTLs typically provide the capability to de-duplicate, compress, and encrypt the data stored on them, as well as the ability to automatically replicate themselves to another VTL. Since the VTL stores data on disk, which is faster than tape, it allows for higher levels of deduplication and compression than tape, as well as encryption, all with no performance impact. Some VTLs can also act as a gateway to cloud-based storage services, providing an easy way to get traditional tape data to the cloud.
It should be noted that many people ignore the 2 in the 3-2-1 rule and only run disk-to-disk-to-disk backups. This approach has a major weakness as it leaves organizations open to malware attacks that can affect not only production data, but backups as well. In disk-to-disk-disk backup scenarios, the backup disk pools typically have a file system on them, and are mountable by either the backup server or other servers making them vulnerable to ransomware attacks. There is nothing worse than finding out your servers have been maliciously encrypted, only to find out that your backups of these servers have also been encrypted by the same malware. Steps can be taken to protect against this by using tape, VTL, or other WORM capable media for one of the three copies of backup data.
Next up, Part 4: Data Archiving.
Get the FREE eBook
This is part 3 of 10 in the From High Availability to Archive: Enhancing Disaster Recovery, Backup and Archive with the Cloud series. To read them all right now download our free eBook.