The Pros and Cons of All Backup Storage Targets Revealed

The days of only having tape solutions are over. Other media have caught up with it in terms of cost, capacity, ease of use, and reliability. Now you have various storage options. Backup applications that work only with tape have very little value in modern business continuity plans.

15 Dec 2023 SintelSedat Akfidan 17 min read
The Pros and Cons of All Backup Storage Targets Revealed

Make sure your software is matched with your hardware.

Use the trial version of the software or carefully read the manufacturer's documentation to determine what types of environments it works with and how you can use them. Backup software targets the following media types:

  • Magnetic tape;
  • Optical disk;
  • Directly connected fixed disks and mass communication devices;
  • Media-independent network targets;
  • Cloud storage accounts.

Magnetic Tape in Backup Solutions

IT departments have relied on tape for backup since the inception of the concept. This technology has matured well and kept pace with the speed of innovation in IT systems. However, the physical properties of magnetic tape impose a serious speed limit on backup and restore operations.

Advantages of Magnetic Tape

  • Most backup software can target it;
  • The cost per gigabyte of tape media is relatively low;
  • Reliable for long-term storage;
  • Light media, easy to transport outside the premises;
  • Easily reusable.

Disadvantages of Magnetic Tape

  • Extremely slow;
  • The cost of tape drives is relatively high;
  • Sensitive to magnetic fields, heat, and sunlight.

For most organizations, the slow speed of tape is its greatest disadvantage. You can find backup applications that support optional features such as running directly from the backup medium. This won't be from a cassette.

That said, tape has a good reliability record. Tapes stored away from magnetic fields and in cool places can easily last for ten years or more. Sometimes, the biggest problem with retrieving data from an old tape is finding a working tape drive. I've seen many technical approaches for tape management over the years.

One of the worst cases involved an employee who carefully took the previous night's tape out of the premises and left it in the control panel of their car. Every evening, they would bake it for several hours under the sun. Therefore, despite the company and its personnel being well-intentioned and following the recommendation to keep backups outside the workplace with a sense of duty, they encountered multiple dead-end cassettes.

On the other end, a customer used a padded, magnetically shielded carrying case to move tapes to an alternate location. There, they placed the tapes in a fireproof safe in a concrete room.

 

Once, I was asked to try to recover data from a ten-year-old cassette. It took nearly a week to find a working tape drive that could read it. That was the only complication. The cassette was still readable.

Optical Media in Backup Solutions

For a short time, developments in optical technology made it appealing. Optical equipment is inexpensive and forms a good interface with operating systems. It even supports drag-and-drop interaction with Windows Explorer. They were the most popular in the domestic market. Some optical systems found their way into data centers. However, as capacities of optical media surpassed those of magnetic media, magnetic media quickly regained the advantage.

Advantages of Optical Media

  • Very durable medium;
  • Shelf life up to ten years;
  • Inexpensive, easily replaceable equipment;
  • Drag-and-drop target in most operating systems;
  • Light media, easy to transport outside.

Disadvantages of Optical Media

  • Extremely limited storage capacity;
  • Very slow;
  • Few corporate backup applications target optical drives;
  • Weak reusability;
  • Wide variation in data integrity after a few years.

When recordable optical media first appeared on the market, people found its reliability attractive. CDs and DVDs are not affected by magnetic fields and have higher tolerance to heat and sunlight. Also, since there is no mechanism in the media itself, they are better protected from rough usage than tapes.

However, they have a few more advantages over other media types. When recordable CDs first appeared, their ability to hold 700 megabytes on a plastic disc was impressive. However, optical media capacities couldn't keep up with magnetic storage.

When recordable DVDs appeared with approximately five gigabytes of capacity, hard disks and tapes were already far beyond that limit.

Moreover, people discovered that the data storage material had a much shorter lifespan than the observable structural materials of optical disks. A disk may look good, but its content might have become unreadable long ago.

Recordable optical media has a wide range of data lifespans from several years to several decades. It has been proven difficult to predict media lifespan.

Because of its low speed, low capacity, and need for frequent testing, you should avoid optical media in your disaster recovery solution.

Directly Connected Storage and Mass Communication Devices in Backup Solutions

You should not limit your backup solutions to systems that distinguish between devices and media. You can also use external hard drives and multi-bay drive enclosures. Some are temporarily connected, usually via USB. Others, especially larger units, use more permanent connections like Fibre Channel.

As the cost of magnetic disks has decreased, these types of systems have become more popular. They have a somewhat limited application scope in disaster recovery, but some organizations can use them extensively.

Advantages of Directly Connected External Devices

  • Fast;
  • Reliable for long-term storage;
  • Inexpensive when using mechanical drives;
  • Easily expandable;
  • High compatibility;
  • Use as a standard file system target.

Disadvantages of Directly Connected External Devices

  • Difficult to transport;
  • Additional concerns when the connection is lost;
  • Many mechanical drives have multiple points of failure;
  • Expensive when using semiconductor drives;
  • Not a valid target in every backup application.

Portability represents the greatest concern when using directly connected external devices for backup. Unlike tapes and disks, the environment is not easily removed after backup completes.

For USB devices, you must inform the operating system that you expect to be removed to complete any write operations including metadata processing and automatic maintenance.

Directly connected Fibre Channel devices usually do not have any fast removal mechanism. In an emergency situation, people should be more concerned with evacuation rather than spending time going through a lengthy removal process. Under normal circumstances, people tend to find excuses to avoid tedious procedures. Expect these systems to remain fixed and in place.

Once upon a time, these limitations made these solutions unsuitable for an operational continuity solution. However, as you will see in the following sections, other developments have made them quite applicable. Nevertheless, you should not use a directly connected device alone. Such equipment should be part of a larger solution.

You may encounter some issues when using external devices with certain backup applications. Fortunately, you won't come across any modern program that definitely cannot back up to a disk. However, some may only allow you to use the disk for short-term storage.

Others may not work properly with removable disks. If you purchase your devices before the software, make sure to thoroughly test their compatibility.

Even though mechanical hard drives have significantly improved in reliability, they still contain a large number of moving parts. Also, the designers of typical 3.5-inch drives did not build them for portability. They can travel, but they are not as good as tapes or disks. Even if you don't transport them, they have more potential points of failure than tapes. Don't exaggerate this risk, but don't ignore it either.

Network Attached Storage in Backup Solutions

Network-based solutions share several features with directly connected storage. Where you find the differences between them, you'll also find their trade-offs. You can use the same pros and cons list for network-based solutions as you did for directly connected systems. Still, we highlight some different points.

In the "Pros" column, network attached storage scores even higher for scalability. Almost every storage unit designed for the network provides multiple partitions. You can start with a few drives and add more when needed. Some even allow you to connect multiple chassis physically or logically. In short, with these types of solutions, you can expand your backup storage space indefinitely.

Network components cause a higher cost per gigabyte for network attached storage. However, the infrastructure required to connect a storage device to a network tends to be an added feature: more features.

Almost all these systems provide some level of security filtering. Cheaper devices, usually marketed simply as "Network Attached Storage" (NAS), may not provide much more than that.

High-end equipment, usually referred to as "Storage Area Network" (SAN), offers even more features. SAN storage can appear to your computer as if it were directly attached storage. In the end, the more you pay, the more you get. Unfortunately, costs increase faster than features.

The advantage you gain in capacity and features is offset by a loss in portability. Many NAS and SAN systems are mounted on racks, so you cannot move them out of the workplace without significant effort.

However, because these devices are network entities, you can place them in remote locations. Using remote storage requires a type of site-to-site network connection, which brings additional costs, complexity, security concerns, potential speed reductions, and more points of failure.

Although placing network attached storage outside the workplace brings additional risks, it also offers opportunities. Most NAS and SAN devices have replication technology. You can back up to a local device and configure it to be automatically replicated to one or more remote sites.

If your device cannot perform replication or if you have different devices that cannot replicate with each other, your backup software may have its own replication methods.

In the worst case, you can use the built-in scheduler of your operating system with free tools like XCOPY and RSYNC.

Using Commodity Computing Equipment as Backup Storage

So far, I've only talked about network devices from a specialized device perspective. SANs have earned a reputation for price tags that exceed their feature sets. At best, this reduces your purchasing power. More commonly, an organization cannot afford to fully utilize a SAN's potential—if they can even afford one.

In the end, you now have options in software-based solutions running on standard server-class computing equipment. Some backup applications can target anything that offers a standard network file protocol such as NFS or SMB.

Software vendors and open-source developers provide applications that enable network storage features on general-purpose operating systems. These solutions fill the price and feature gap between NAS and SAN devices. However, they require more management effort than specialized devices to deploy and maintain.

When I first created my backup solution targeting a specific device, I quickly learned that hardware vendors emphasized their systems' performance features. Since I only needed large capacity, I priced a low-quality rack with many drive bays filled with large SATA drives. I saved a bit by choosing from the device options.

Using Commodity Computing Equipment as Backup Storage

The Role of Hyper-Converged Infrastructure in Backup

A relatively new type of system, commonly known as "hyper-converged infrastructure" (HCI), has increasingly taken on a significant role in data center infrastructure. In the traditional scale-out model, server-class computers handle processing tasks, SAN or NAS devices store data, and physical switches and routers connect everything.

In HCI, server-class computers handle all roles, including much of the network communication.

Very few organizations will design an HCI solely for backup. Instead, they will deploy HCI as the core data center solution.

Initially, data centers used purpose-built hosts for specific roles such as domain controllers and SQL servers. As technologies matured, vendors and administrators increased resilience by clustering the hosts.

These clusters remained purpose-built hosts. In the second generation, server virtualization started to break down the model of single-use physical hosts. However, most administrators continued to host and store based on themes for the sake of organization and permissions.

HCI replaces this paradigm by enabling true "cloud" concepts. With HCI, we can still define logical boundaries for processing, storage, and networking, but the barriers are only logical. We may not even know which physical resource hosts a particular server or database file.

Even if we do learn, we can react to environmental events.

Files can be distributed across the storage layer bits throughout the data center, or even between well-connected data centers. In short, HCI administrators only need to be concerned with the organization's overall capacity.

If some resources decrease, they purchase more equipment and scale out the HCI footprint. Well-designed, hardware purchases and allocations occur in different cycles and levels than server provisioning and storage allocation.

All of this provides two things for backup with HCI:

  • You can place your internal backup infrastructure in HCI just like other server roles for on-premises backups and general cloud transitions
  • You may have concerns about mixing what you're backing up with the backup itself

The first perspective has the strongest supported argument. In any case, you need multiple independent backup copies, so sending data to off-site locations reduces dependency on HCI.

Additionally, many managers (and the non-technical people in the reporting chain) may not understand that living together doesn't automatically imply a common viewpoint.

You can design your HCI so that backup has no effective visibility into production components. It works exactly as we've always set up data center backups, but the partitions are in software rather than hardware. However, how much of anyone's fear is justified doesn't matter.

If you face significant resistance to bundling backup with your HCI deployment, go for the traditional architecture. You sacrifice a bit of efficiency, but not to the point of being crippling.

Portability represents the greatest concern when using directly connected external devices for backup. Unlike tapes and disks, the environment is not easily removed after backup completes.

For USB devices, you must inform the operating system that you expect to be removed to complete any write operations including metadata processing and automatic maintenance.

Directly connected Fibre Channel devices usually do not have any fast removal mechanism. In an emergency situation, people should be more concerned with evacuation rather than spending time going through a lengthy removal process. Under normal circumstances, people tend to find excuses to avoid tedious procedures. Expect these systems to remain fixed and in place.

Once upon a time, these limitations made these solutions unsuitable for an operational continuity solution. However, as you will see in the following sections, other developments have made them quite applicable. Nevertheless, you should not use a directly connected device alone. Such equipment should be part of a larger solution.

You may encounter some issues when using external devices with certain backup applications. Fortunately, you won't come across any modern program that definitely cannot back up to a disk. However, some may only allow you to use the disk for short-term storage.

Others may not work properly with removable disks. If you purchase your devices before the software, make sure to thoroughly test their compatibility.

Even though mechanical hard drives have significantly improved in reliability, they still contain a large number of moving parts. Also, the designers of typical 3.5-inch drives did not build them for portability. They can travel, but they are not as good as tapes or disks. Even if you don't transport them, they have more potential points of failure than tapes. Don't exaggerate this risk, but don't ignore it either.

Cloud Storage in Backup Solutions

Various technological advances over the last few years have made Internet-based storage feasible. Now most organizations can access reliable, high-speed Internet connections at a low cost. You can use this to solve one of the most challenging problems in backup: keeping your backup data safe in a location that's protected from local disasters. Of course, these rewards don't come without risk and cost.

Advantages of Cloud Backup

  • Future-proof;
  • Inherently off-site from the beginning;
  • Wide geographic diversity;
  • Extremely reliable;
  • Effectively infinite scalability;
  • Accessible from anywhere;
  • Security.

Disadvantages of Cloud Backup

  • Dependencies outside your control;
  • Switching to another vendor is expensive;
  • Risk of unrecoverable outages;
  • Speed.

Cloud providers replicate storage across geographic regions to ensure they keep their promises to customers (cheaper plans may not provide this protection).

Therefore, although you need to worry about failures in your connection chain and outages at the cloud provider, you can be sure that you will eventually regain access to your data. This provides an inherently unmatched level of reliability for cloud backup.

All major cloud providers invest significant effort to ensure their customers' security. They boast compliance with accepted, standardized security practices. Each has a large team of security experts dedicated solely to ensuring customer data safety.

This means you don't have to be overly concerned about violations at the provider level.

Nevertheless, you will still need to ensure the security of your account and access points. Like other Internet-based resources, the provider should somehow make your data available to you.

Malicious attackers may target your access path rather than the provider itself. Therefore, you still have some responsibility for securing your cloud-based data.

When using cloud storage for backup, two things are most likely to cause a failure. Your Internet provider is the first.

If you can't establish a reliable connection with your provider, your backup operations may frequently fail. Even with a solid connection, you may need more bandwidth to support your backup requirements.

For the second issue, you can choose a backup solution like Hornetsecurity's VM Backup that offers features such as compression and data deduplication to reduce network load.

The second major concern is temporary providers. You can trust that your cloud provider will continue to take security precautions, but many third-party providers follow less stringent practices.

If your backup system sends encrypted data directly to a cloud account that you control, there's not much for you to worry about. Confirm that your software uses encryption and keep track of updates, so there will be very little for you to worry about beyond your organization's firewall.

However, some providers send your data to an account under their control and resell it to customers. If they are insufficient in security measures, they put your data at risk. Be very careful when examining these types of providers.

"Cost" does not appear on either the pros or cons list. Cost will always be a source of concern, but how it compares to on-premises storage will vary between organizations. Using cloud storage allows you to eliminate so-called "capital expenditures": payments made in advance for physical assets, which are usually significant.

If you have an Internet connection, you won't need to purchase any additional equipment. Additionally, it removes some "operational expenses": recurring costs for maintaining goods and services.

You will still need to pay for your software licenses, and your cloud provider will bill you regularly for storage and possibly network usage.

However, you won't need to purchase storage hardware, and your employees won't need to spend time maintaining it. You transfer all the difficulties and costs of hardware ownership to your provider for a lower total cost.

Unfortunately, you shouldn't transfer all your backup load to a single cloud provider. Due to the risks and speed limits of relying on an Internet connection, it's still prudent to keep at least part of your solution on-site. Therefore, you should still expect some capital expenditures and local maintenance activities.

Taking Action

The previous section helped you work on your software options. If you've made a final choice, it gives you at least some control over your hardware purchases. If not, you can explore hardware options and work backward to choose your software.

In particular, the full deployment style you use for the internal part of your solution matters only to the extent that it ensures your backups operate smoothly.

Prioritize fitting your needs with any paradigm. You need space to store your backups, software to capture them, and network and transport infrastructure to move them from live systems.

Four Steps to Selecting Hardware

The truth is, your budget plays the biggest limiting role in hardware options.