Everyone is familiar with "storage".
Especially in recent years when electronic products are rapidly popularized, various storages can be seen everywhere.
Have this?
This one?
and this?
Haha, yes, we are in the information age, and data files such as videos, music, pictures, text, and tables are completely inseparable.
For many small partners, the value of storage is to allow us to save our important information so that we can find, observe, and "learn" when we need it .
But for enterprise users, especially in the era of big data, the value of data has been brought to an unprecedented position. Data is the lifeblood of an enterprise.
▉ Block storage, file storage, and object storage are unclear
When it comes to enterprise storage, the stack length has been troubled by three storage terms for a long time. Block storage, file storage, and object storage are really stupid.
Recently, the stack leader has not forgotten sleep and food, and has studied the next block storage, file storage and object storage.
Today we will report to you the results of our learning: what is the difference between block storage, file storage, and object storage?
In fact, the purpose of storage is to provide space for data.
Hard disk/solid state disk is the ultimate storage carrier. The reason why there are different types of storage devices such as block storage, file storage and object storage are mainly due to the different means or methods of using media to store data.
▉ Block storage
Block storage provides raw disks without a file system and needs to be initialized before use. Our more common form of block storage is our Windows hard disk or mobile phone storage space, and data is stored and accessed in bytes.
When using block storage, the file system is on the client side, such as a Windows system. We can view the content we store on the Windows desktop, but we don’t know anything about the content and format of the hard disk, and we can only rely on the client’s file system. View.
In layman's terms, as shown in the figure above, the data is like every melon seed, stacked in the storage bin. Guazi is each data block, and this storage compartment is the disk. Block storage only cares about the in and out of melon seeds, not the relationship and use of melon seeds.
Since block storage is only responsible for data reading and writing, it has the advantages of high bandwidth and low latency, but has limited scalability and is suitable for systems that require high response time. For example, the storage of core enterprise applications such as databases and ERP. The DAS and SAN storage devices belong to the block storage type.
▉ File storage
The storage end of file storage has a file system, and our common NAS storage is file storage device. These file storage devices also have file systems in addition to disks, and users can call storage resources directly through the file system on the storage side.
File storage generally manifests in the form of directories and files (such as C:\User\Program Files\Common Files). Data is stored and accessed in the form of files, organized according to the directory structure. File storage has its own file system, which is a unified file system for each client, which can be used by multiple clients at the same time and is easy to share.
For example, in the figure above, the data forms sunflowers together like melon seeds, and then correspond to different sunflower stalks. To find a sunflower seed, first find the corresponding sunflower stalk, then find the sunflower, and then according to the corresponding sunflower Find this seed at the location.
Compared with block storage, file storage has its own file system, which can achieve more advanced management and can be easily shared, so it has a wide range of uses. For example, commonly used NFS, CIFS, ftp, etc. are all based on file storage. But compared to block storage, file storage read and write speeds are slower than block storage.
▉ Object storage
Block storage has excellent performance but cannot be shared. File storage can be shared but the speed is always unsatisfactory; as an adult who does not know how to do multiple-choice questions, he wants performance and sharing, and at the same time meets the needs of large-scale expansion. All have object storage.
The general manifestation of object storage is a UUID (such as our common website https://new.qq.com/omn/20201119/20201119A0CEI01210.html). This UUID is unique, and only this UUID can be passed even in the world.
Data and metadata are packaged together as a whole object and stored in a large pool. The user wants to access, only need to pass its UUID to find it.
For example, in the above figure, the sunflower seeds of the data are made into packaging bags, each packaging bag has a unique factory barcode, but when looking for the corresponding corresponding melon seed bag, you can only find the corresponding melon seed bag through the unique barcode, but every time It can only be a bag.
The file system on the object storage side uses this hash table-key value (which can be understood as a dictionary search, up to two levels of directories) to improve the speed of reading and writing. Object storage can be easily expanded to ultra-large scale, so it is very suitable for videos, images, etc. with large data volume and rapid growth, such as Baidu network disk and big data storage;
▉ To be continued...
With the rapid development of the Internet, the amount of data is continuously increasing, and traditional storage systems can no longer meet the demand.
Especially after the rise of cloud computing, will there be any brand new storage solutions?
Please see the next issue - "The Deep Secret of Distributed Storage"