I've always associated the term "primary storage" with main memory, so memory which can be accessed directly by the CPU and its of the volatile type. The past few days I've been reading some technical articles/publications where the term "primary storage" seems like it's used as synonym for secondary storage and in the Wikipedia page for Computer data storage it says:

Recently, primary storage and secondary storage in some uses refer to what was historically called, respectively, secondary storage and tertiary storage

So, I'm guessing that primary storage ≠ main memory and nowadays it's synonym with secondary storage which persist once the system is shut down. Is that correct ?

Also, one of the articles I was reading was: Deduplication for primary block storage

Another one: iDedup: Latency-aware, inline data deduplication for primary storage


What is the contemporary meaning of primary storage?

I believe it now depends on the context that term is used in.

In the context of computer architecture, your understanding of "primary storage" is correct and still used. There is a hierarchy of storage systems in every computer and it still makes sense to use these well established names when speaking of the entire computer's storage systems.

There is also the context of persistent storage systems, like a data center or a NAS. It also makes sense to classify the drives into a hierarchy based on how they are used and what data they contain. The drives that are used more often (aka the "hot" storage) can be referred to as "primary storage", while the drives containing backups would be considered "secondary storage".

Think of it this way, a computer may have:

  • Primary Storage (RAM)
  • Secondary Storage (Persistent)
    • Primary "secondary" storage (actively used data)
    • Secondary "secondary" storage (backups of the "primary" drives)
  • and so on...

You have to admit that the WikiPedia Online Library has its ways and methods to confuse the readers instead of making their minds clearer. This idea is true for the paragraph that you have mentioned:

"Recently, primary storage and secondary storage in some uses refer to what was historically called, respectively, secondary storage and tertiary storage."

The paragraph is not backed up by anything else than a reference. However, the primary storage is the main memory as you have understood it correctly. Their definition for the primary storage appears in the beginning of the section that is ended with that confusing paragraph.

"Primary storage (also known as main memory, internal memory or prime memory), often referred to simply as memory, is the only one directly accessible to the CPU."

So basically the Central Processing Unit has direct access to the primary storage. Its most common form is the Random Access Memory.

The non-primary storage could be seen as the type of storage that is indirectly accessed by the Central Processing Unit through Peripheral Devices. In WikiPedia's view, it has been categorised as secondary storage, tertiary storage and off-line storage.

  • 3
    I didn't notice the reference for that paragraph. The reference mentions "Primary storage devices refer to storage hardware for data in active use, in contrast to storage used for backup purposes" So, I guess in some context it refers to data in active use in Storage are Network or Network attached storage, etc. – Papaya-Automaton Nov 19 '19 at 18:09
  • It should be irrelevant whether it is related to networks. As they said, it should be actively accessed and not saved in archives during BackUp-and-Restore Operations. Whether local or remote, it should be easily modified also. However, instead of confusing the readers by switching the ordering of the expressions, they could have used other types of expressions and let the readers stick to the expression of primary storage as the most used type of storage, the way that you understand it as the main memory. – user1018743 Nov 19 '19 at 18:18
  • Ok I think I understand now. So, it's primary storage because even though a NAS and flash storage which might be non-volatile they provide a way to access/modify data easily? and have access to the CPU\RAM memory as well? Sorry, for the questions I'm really trying to understand this clearly. – Papaya-Automaton Nov 19 '19 at 18:44
  • 1
    The Central Processing Unit is accessing everything. The Random Access Memory is a type of primary storage because it is both easy to access and it is also used directly. Other types of memory might also be considered as primary storage if they are used directly and not for archiving during BackUp-and-Restore Operations. WikiPedia must have thought that accessing a .ZIP archive on a remote tape machine is even beyond the term of secondary storage and this is why they must have come up with the term of tertiary storage. – user1018743 Nov 20 '19 at 11:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.