It's a recurring question: which storage solution is best suited to my different workloads and data? Block storage and object storage each have their advantages and some limitations. It's up to each company to find the right compromise.
Storing data for security or analysis is essential. It's even becoming "mandatory" for two main reasons. Firstly, to comply in particular with the Regulation on the Protection of Personal Data(RGPD) by implementing encryption and anonymization.
Secondly, to outperform and out-innovate its competitors by leveraging Big Data and Artificial Intelligence technologies to predict trends in particular.
Which solutions to choose?
Storage is therefore a necessity. Yes, but which solutions to choose? In reality, there's not just one, but several. The purpose of this article is not to discuss aging storage arrays or storage automation. The point is to compare object storage with block storage.
Block storage (or File Object) is commonly used for most applications. Block storage devices (SAN, iSCSI, local disks) are generally formatted with a file system such as FAT32, NTFS, EXT3 and EXT4.
Also known as object-based storage, object storage is a general term referring to the way we organize and work with storage units, called objects.
Each object contains three elements:
- The data itself, in all its diversity;
- Metadata, which contains contextual information about what the data is, its purpose, confidentiality or any other information relevant to the processing of the information;
- A globally unique identifier. This is an address given to the object so that it can be found on a distributed system.
Advantages of block storage
These devices offer raw storage capacity of a fixed size. Each volume can be treated as an independent disk drive and controlled by an external server operating system. This block device can be mounted by the guest operating system as if it were a physical disk.
Limits of block storage
Volumes are only accessible when connected to an operating system. In addition, files are divided into data blocks of equal size. Each has its own address, but no additional information (metadata) to provide further context.
Benefits of object storage
Object storage is ideal for storing photos, videos, log files and other unstructured data that is mainly read and rarely modified. Data and metadata can be accessed via APIs or an https browser (more secure than an http connection...).
What's more, there are no limits on the type or quantity of metadata. This makes object storage powerful and customizable. Very practical, metadata can include anything from the security classification of the file in the object to the importance of the application associated with the information.
Data can be replicated across different data centers. A further guarantee of security in the event of a node failure.
Limits of object storage
This solution does not generally give you the option of incrementally editing part of a file (as block storage does). Objects must be manipulated as a whole unit. This means accessing the whole object, updating it and then rewriting it in its entirety. Furthermore, it does not divide files into blocks of raw data.
Each solution has its advantages and disadvantages. Contrary to popular belief, object storage is not the answer to all your problems. Sometimes, block storage is a much better solution.
Every company needs to set the cursor at the right level to reconcile scalability, resilience and performance. Without forgetting the essential: prioritize your data to focus your efforts and resources (financial, human and technical) on those that are vital to your business.
