With object storage, there is an opportunity to improve on your data protection plan.
First, let's consider why we are deploying object storage. It is usually deployed for massive scale and for its durability characteristics. Massive scale tells us that the traditional methods of data protection used with primary storage, such as running periodic backups, are impractical.
Object storage systems currently available on the market offer two primary methods to protect data. The first method disperses segments of data using information dispersal algorithms so multiple data access failures can be tolerated. In this scenario, forward error correction is used with correcting codes (such as erasure codes in some implementations) to recreate the data that could not be accessed. The other commonly used method is to replicate data to another system that may be at a remote location and to make a new copy whenever data is changed. Many object systems allow the administrator to control how many versions to maintain. Two copies of data are maintained for each instance of data, providing effective data protection.
Objects are containers for data that are similar to files, and may even encapsulate files in some implementations. Object containers are handled differently from files. Objects do not exist in a hierarchical structure; instead they exist in what is termed a flat namespace, where access to object information depends upon knowing the object ID or having an index of the objects where the object ID can be acquired. An object also has metadata associated with it.
This is the point at which it is helpful to think about data protection. Object metadata includes system information that is similar to file system metadata. It can be established by applications, users or administrators. In an object storage system, the metadata is stored with the data and may be used to control access to the data and manage retention. For example, the metadata can store data protection controls regarding the number of copies maintained, as well as data dispersal requirements. Most object storage systems will write an object once and an update to the object, resulting in the new version being stored. Thus, the term versioning.
It's important to note that information stored on object storage systems does not have the same requirements as primary storage does when it comes to changes and frequency of access. So, a data protection plan can exploit those differences. Automation of the process and the new capabilities introduced with versioning are valuable data protection tools contained within object storage systems.
Why object storage and file storage data protection is different
Using erasure coding and mirroring for cloud data protection
Dig Deeper on Data center storage
Related Q&A from Randy Kerns
What is the one hidden gotcha that you'd advise users about if they were shopping for an all-flash storage array? Continue Reading
How much control do you have with all-flash storage arrays? How much control do you have over how arrays handle your data? Do you control the caching? Continue Reading
Vendors often publish numbers for 'usable' capacity versus 'effective' capacity. Can you explain this and how can you plan flash capacity needs with ... Continue Reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.