NAS, tape, HDDs, SSDs - or rather in the public cloud?
Even if archive data is no longer actively used, companies must store it for the long term for legal reasons. Data storage must therefore be inexpensive, secure, and easy to use.
Which storage medium is best suited for long-term archiving? In this article, we compare the different media and explain the advantages and disadvantages of each.
Two of today's most common storage solutions are NAS (network attached storage) and public cloud storage.
A NAS consists of several hard disks which are connected directly to the company network via a LAN connection. The storage capacities of the hard disks are combined and managed centrally. NAS is often used in companies to archive data in their own data center.
In the public cloud, providers make storage resources available to companies via the Internet. The hardware is located in the cloud provider's data center and is managed by the provider.
How do public cloud and NAS compare? Four criteria play a role:
The biggest difference between NAS storage in your own data center and public cloud storage is the cost of maintenance.
With NAS storage, you yourself have to ensure that...
But even with local archiving on a NAS system, the effort can be reduced. With managed services, you can use your own servers for archiving and outsource maintenance to a service provider. This way, you have your own NAS systems and can use services like those in the public cloud.
The effort required for public cloud storage is significantly lower than for classic NAS storage. Ordering, implementation, and provisioning are usually handled at the push of a button without any special effort. However, management is your own responsibility, which means that you also need your own experts for public cloud storage and the time/management effort should not be underestimated.
How many TB or PB do you need in 3 years? In practice, this can only be roughly planned. Archive data often has to be stored for several decades. If the data in the archive grows, the storage space quickly fills up.
NAS storage can be easily scaled by connecting more hard disks or another NAS system. The individual components then form a central storage pool.
This is even easier in the public cloud. If the storage space is not sufficient, a larger storage quota can be ordered with a click. Since you don't need more space in the data center for this, you can simply scale storage capacities up or down.
Storage systems can be scaled up (vertical scaling) or scaled out (horizontal scaling).
Scale-up: You replace your existing storage system with a larger one.
Scale-out: You expand your existing storage system with additional components (e.g. hard disks) or cluster nodes.
NAS storage works according to the scale-out principle, in which additional hard disks are connected and the storage capacities are combined to form a storage pool.
In the public cloud, the provider takes care of the security of the infrastructure. The provider has IT security specialists who secure the systems. Companies often do not have enough IT staff to do this.
Nevertheless, most public cloud providers give no guarantee that no data will be lost. So how do you recognize a secure cloud provider? A trustworthy provider can show security certificates such as ISO 27001. A good indicator is if the data centers are operated in Germany or the EU. The provider is then obliged to comply with EU or German law.
With public cloud storage, the user is responsible for configuring the storage setup and applying the security guidelines. This requires internal knowledge and regular training. If policies are not applied consistently, this leads to exposed buckets and data.
When you store your data locally on a NAS, you know how and where your data is stored and have full control over it. The data is not shared with third parties, such as a cloud provider.
But with full control comes full responsibility. You must ensure that your data is stored in compliance with the law, is not lost, and is protected from cyber-attacks.
The costs for public cloud or NAS storage are calculated from different models. So, which is cheaper?
With NAS storage, you bear the cost of the hardware, power, and cooling, and the IT staff time to maintain the storage. A NAS is more expensive to purchase compared to the public cloud. However, once the hardware is purchased, there are no major costs.
For public cloud storage, you pay a flat rate which includes the storage space and all services. The monthly operating costs of the public cloud grow with your booked storage volume. This is especially true for data which you need to access frequently, because with public cloud providers you pay for each read access.
The lifespan of NAS systems is estimated at three to five years. If you compare the cost of five years of public cloud storage to the total cost of NAS storage, NAS can be an affordable alternative. The analysts at Enterprise Strategy Group (ESG) have made exactly this total cost comparison.
There are various storage media to choose from for data archiving in your own data center. We will discuss which storage media make sense for which application.
Data can be stored on tapes (tape cartridges). This technology is older, but inexpensive. In addition, the data can no longer be changed after writing - actually ideal for long-term archiving of data.
But it is not possible to access a single file; the entire tape must be restored. The effort involved is great and the process lengthy, which is why tapes are now only used in special cases. As a rule, archive data does not have to be accessed in a matter of seconds, but in recent years the demands on performance and access frequency have increased significantly.
When are tapes useful? Tapes are a suitable storage medium if you want to store internal knowledge securely for the future or store hardware with sensitive data in a safe.
HDDs are suitable for NAS systems and offer large storage space for a low price. Data from the TB to the PB range can be stored on an HDD. The storage is not as performant as SSDs, but individual archive data can be accessed quickly if required.
But HDDs are susceptible to external influences. Shocks, heat, and humidity can cause the disks to be damaged and data to be lost. The magnetic field can also lose its power after some time. Then the data stored on it is also no longer available. This is why you should always make 2-3 copies of archives on HDD. If one fails, you can fall back on the others.
SSDs are very performant - but also expensive. This is not necessary for archiving.
If you have free storage capacity, data can certainly be stored here on a transitional basis. In the long term, however, it is worth setting up a cost-effective system with HDDs, because the data volumes in archive storage are also growing rapidly.
There is no such thing as THE best storage medium. Depending on the application, a different solution or combination of different solutions is suitable.
Sensitive data is archived in the company's own data center. Tapes can be used to store data which needs to be kept for a very long time but is rarely accessed.
Archive data from data sources which are growing rapidly are written to the public cloud. In this way, IT can easily make new capacities available quickly if required. However, the costs and efforts must be kept a close eye on here, as these can quickly and significantly exceed expectations.
Many organizations do not have the ability to store data in the public cloud for privacy and compliance reasons. Hospitals, for example, are not allowed to archive patient data in the public cloud.
A good alternative to the public cloud is offered by on-premises solutions with managed services - for example, set up as a scale-out cluster. This allows you to implement the user experience of the public cloud in your own data center and retain full control over your archive data.
