Every organization uses disk as a backup target. That is a given in organizations of all sizes. However, using a “dumb” or “slow” disk array as a backup target no longer aligns well with current organizational requirements. Ransomware places new demands on disk backup targets for analytics, instant recoveries, and data immutability. In response, current disk backup targets offer more scalability, security, density, and performance features than ever before.
Ransomware Driving Changes to Disk Backup Targets
Ransomware’s threat represents the single biggest force changes to disk backup targets. When first released, high-capacity disk backup targets met most organizations’ initial expectations. (See sidebar No. 1) However, these disk backup targets may fall short today when it comes to meeting new demands driven by ransomware.
First and foremost, more organizations seek to secure their backups. Sophisticated strains of ransomware now specifically target backup data stores. As part of their attack, they compromise existing backups stored on disk by either deleting or encrypting them. Storing backups on disk in a secure, immutable format prevents backup deletions or encryptions from occurring.
Once safely stored, more organizations want to analyze their backup data. Performing analytics often helps detect the presence of ransomware in their production environment. Analytics also serves other purposes. It can validate the viability of backups for recovery as well as provide a means for organizations to do data mining.
However, viable backups and backup data security represent only one half of the disk backup target equation. If ransomware were to compromise production applications and data, organizations would need to recover quickly, perhaps in minutes. These requirements may dictate using disk backup targets that can host production workloads, potentially for extended periods.
These combined performance and security demands impact organizations in multiple ways. Backup budgets and staffing levels may stay the same despite these increased demands. These targets must also deliver performance and scalability while remaining cost-effective, flexible, and simple to deploy and manage.
Disk Backup Targets Experience Makeover
Recently introduced features on disk backup targets reflect this combination of new, organizational requirements. While every organization may not possess every one of these new requirements, most organizations need to implement multiple features. To accommodate these diverse organizations’ requirements, disk backup targets have experienced a makeover.
Many disk backup targets offer multiple new deployment and implementation options. These primarily surface in their abilities to store backups in an immutable format and in the cloud as well as host production workloads. These options give organizations more flexibility to obtain features they need while still being able to afford and manage them.
Immutable Storage
The introduction of an immutable storage option represents one of the most notable recent improvements to disk backup targets. Using this feature, backups become unchangeable once stored. As each backup gets stored, it receives an associated retention date. Only after the retention date expires may one delete or change the data. This prevents hackers or even malicious or compromised employees from attempting to delete or change backups. (See sidebar No. 2)
Organizations will find immutable storage features on both cloud and on-premises disk backup targets. If using cloud storage as a disk backup target, organizations should use their object storage offering. Most object storage offerings utilize simple storage services (S3) which offer an object lock option which supports data immutability. Many cloud storage offerings from general-purpose, purpose-built, and private cloud storage providers support this option.
Many backup software products also include object storage and its object lock support as part of their overall management capabilities. Backup software often provides the best way to manage object storage and its object lock feature. Organizations may simply set policies in the backup software that assign a retention date to each backup.
Production Workloads
A disk backup target’s ability to support production workloads comes into play for organizations needing instant recoveries (five minutes or less.) In these cases, organizations want or need to access the data or run the application directly on the disk backup target’s storage. Many disk backup targets now facilitate this activity. However, they use various architectures and technologies to perform instant recoveries.
Choosing the right disk backup target requires that organizations answer the following four questions to the best of their ability:
- How much data or how many applications do they anticipate instantly recovering?
- How well do they want the recovered applications, data, or both to perform?
- How long do they want or need the disk backup target to host the data?
- Do they need options to add, change, and/or delete the data once recovered?
The answers to these questions will influence their choice of a disk backup target in the following three ways.
1. Capacity
Many organizations initially focus on the answers to the first two questions about performance and capacity. If organizations only expect to access backup data with minimal or no performance expectations, that simplifies the choice. They simply need to ensure a disk backup target scales to host the amount of data they need to recover.
The type of disk backup target organizations should choose will change depending on the amount of data they need to backup and recover. Organizations may use the following guidelines in selecting a disk backup target:
- If under 100TB of data to backup, recover, and host, give preference to a disk backup target with a single or dual-controller architecture.
- If under 1 PB, give preference to a single or dual-controller architecture. However, if an organization anticipates its data growing to more than 1 PB, give preference to a hyperconverged or scale-out architecture.
- If more than 1PB, give preference to a hyperconverged or scale-out architecture.
2. Performance
If organizations expect their disk backup target to provide production levels of performance during recoveries, then the decision becomes more complicated. They still need to account for the disk backup target’s capacity. However, they must also more carefully examine the type of storage housed in the disk backup target.
More disk backup targets support solid state disk (SSDs) in addition to hard disk drives (HDDs). While more expensive (SSDs cost about 2-3x more than HDDs), they deliver better read and write performance. SSDs can improve write backup throughput by up to 20x versus HDDs and deliver up to 10X better read performance.
Organizations must also validate how much SSD capacity the disk backup target supports if they plan to recover lots of data or many applications. Some disk backup targets only support a few TBs of SSD. Other providers offer all-SSD models. Organizations may optionally obtain SSDs in most disk backup target architecture – single-controller, dual-controller, hyperconverged, and scale-out.
3. Length of Time Hosting Recoveries
The length of time that organizations plan to use a disk backup target to host a recovery represents one more consideration. Some only plan to use the disk backup target for reading the recovered data.
This holds true in at least two use cases. For instance, some restore backups to perform analytics or scan the data for malware. Alternatively, they start application or data restores by only reading data from the disk backup target. They only resume full operations once they restore the data to production disk. In both these cases, they only need to read the data residing on the disk backup target.
However, if organizations need to do more than just read the data, this adds another wrinkle to the decision. Making additions, changes, or deletions to data may not be possible when hosting data on some disk backup targets. This especially holds true if storing data in an immutable format on disk backup targets. In these circumstances, organizations will need a disk backup target that supports making additions, changes, or deletions to recovered data.
Cloud Integrations
Storing backups off-premises is a must-have for most organizations with many using cloud providers to accomplish that purpose. The challenge often becomes identifying the best means to move backups stored on disk backup target to the cloud.
To help, most disk backup targets natively integrate with cloud providers at some level. In this scenario, the disk backup target essentially functions as a disk cache. The backup software sends backups to the local or on-premises disk backup target to complete backups more quickly. Once the backup completes, the disk backup target copies the backup to the cloud provider of the organization’s choice. This activity occurs in the background.
Organizations should validate that the disk backup target integrates with the organization’s backup software. This integration helps ensure organizations can track the backup data’s location once stored in the cloud and its retention period.
Disk Backup Targets Get Modern Makeover
Organizations can still easily obtain and use offerings that function as “dumb” disk backup targets that perhaps deduplicate and compress backups. However, organizational backup needs have changed, driven in large part by the threat of ransomware. Organizations must take steps to protect their backups from ransomware and to position themselves to recover from an attack.
These ransomware realities have resulted in a modern makeover of disk backup targets. Organizations can use these disk backup targets to store backup data in an immutable format to secure it from ransomware attacks. They can simultaneously utilize these targets to perform instant recoveries and even potentially run production workloads on them.
However, disk backup targets offer differing architectures, capacities, performance metrics, and feature sets. Organizations must first answer questions around their specific capacity, performance, and recovery requirements to choose the best disk backup target for their environment.
