Le stockage dans le cloud permet aux utilisateurs de stocker des données sur des serveurs distants à l'aide d'une connexion Internet.
Cet article s'articule autour des points suivants :
Copier le lien de l'article
Cloud storage is a method of data storage and organization that takes place in the cloud, a network of remote servers that can be accessed over an Internet connection. With cloud-based storage, users and businesses can store, access, and maintain their data from any location that offers an Internet connection, rather than confining their files to a single location or device.
Before cloud computing was common, users maintained their data on a personal hard drive or on-premise data center, limiting their use of that data to one location. Think of it like choosing to invest in a landline rather than a mobile phone. As long as the user is inside their house, they’re able to get to the phone whenever it rings, but as soon as they leave their property, they become unreachable. With a mobile phone, however, they can answer a call from any location.
It works the same way with cloud storage. Users no longer have to stay home (or inside their office building) in order to access their data 24/7; instead, they can take their files on the go without lugging around expensive and limited hardware.
Cloud storage vendors allow users to rent storage space on networks of remote data servers. Users pay cloud vendors for the ability to run applications, host services, and back up data to the cloud and usually adhere to usage limits that can be expanded for an additional fee, enabling them to scale their storage needs as necessary.
Cloud storage vendors need to ensure that files aren’t compromised, lost, or destroyed in the cloud. Since there’s no way to completely safeguard against hardware failure, cloud vendors use redundant storage methods, duplicating essential hardware, data paths, and power supplies to keep their cloud storage services up and running at all times.
For example, say a user wants to back up their entire photo library from their smartphone to the cloud. Multiple copies of their photos will be sent to multiple data servers, so in the unlikely event that their phone gets irreparably damaged or one of the servers spontaneously combusts, they’ll still be able to retrieve those files from another server.
Cloud storage has several deployment models, including personal storage, private storage, public storage, and hybrid storage.
Personal storage allows individual users to store data in the cloud and access it from multiple Internet-enabled devices, such as a mobile phone, tablet, or laptop. This is the main method consumers use to back up their files to the cloud. Using cloud computing services like Google Cloud, users can sync their data across multiple devices and retrieve it wherever they have a secure Internet connection.
Public storage refers to a cloud storage vendor that manages the data of multiple users in a public cloud. Each user rents storage space in the cloud while releasing the burden and expense of maintaining the underlying infrastructure to the cloud provider. Although users share cloud storage (a practice called ‘multitenancy’), they do not have access to each other’s data, in the same way that different residents might share an apartment building without being allowed to enter each other’s apartment units.
Private storage is a cloud storage service that is privately arranged for a single business. A standard private storage solution can be managed in an on-premise data center (internal private cloud) or remotely by a third-party vendor (hosted private cloud). Because it is only accessible to one business, private storage offers a higher level of security than public storage. However, businesses that choose to deploy internal private clouds may find them a less reliable option, as they can suffer critical data loss if their on-premise facilities are compromised.
Hybrid storage refers to a combination of public and private storage. Businesses that cannot afford expensive private storage options may elect to host some of their data in a private cloud while taking advantage of the scalability and affordability of a public cloud.
Cloud services are compatible with three primary types of data storage: file storage, block storage, and object storage.
File storage is a hierarchical system of data organization that nests folders of data within directories and subdirectories. Inside each folder, files are tagged with metadata* for easy identification and retrieval, though this process becomes vastly more complicated when applied to a large amount of data. Although file storage is one of the most affordable cloud storage solutions, it doesn't scale up easily.
When comparing the benefits and drawbacks of various storage systems, it may be helpful to think of each system as a way of organizing clothing. File storage is like putting clothes in separate, labeled dresser drawers. One drawer might be for shirts, with smaller sections apportioned within the drawer for dress shirts, t-shirts, and tank tops. It’s a straightforward system that’s easy to keep track of, but one that may become unwieldy if used to organize hundreds of individual clothing items.
Block storage organizes data into storage volumes called ‘blocks.’ Each block is comprised of files that have been arranged into same-sized chunks of data and tagged with a unique identifier. The controlling operating system treats each of these volumes as a separate hard drive and allows them to be stored in whatever environment is most convenient. While block storage makes data organization and retrieval easier than file storage, particularly for businesses with massive amounts of data, it faces metadata constraints and is often a less cost-effective option.
Block storage is like trading in a dresser for plastic tubs. Each tub holds the same amount of clothing and can be labeled with the kind of clothing it contains: shirts, socks, pants, etc. Instead of trying to find room for a large dresser (or building a separate room to contain multiple dressers), the tubs can be placed wherever is convenient for easy access and retrieval.
Object storage refers to the creation of distinct units called “objects,” each of which contain unstructured data, metadata, and a unique ID. Because objects are stored in a flat address space (rather than a file directory) and enable more complex metadata customization, object storage makes data retrieval even easier than file and block storage. It also combines some of the benefits of file and block storage, as it’s both cost-effective (users pay only for the storage space they require) and scalable (users can use object storage systems to manage massive amounts of data).
Object storage is like creating compact, custom containers for each type of clothing and labeling every container with a Sharpie. One container might be for blue dress shirts, another for pink biking shorts, a third for animal-print socks. Although the containers cannot be modified, they can be more easily stored and identified than tubs or dressers.
*Metadata helps organize files by assigning descriptors to pieces of data, such as titles, tags, categories, prior modifications, and permissions. The more detailed the metadata is, the easier it is to find and retrieve a file within a data storage system.
Cloudflare acts as an independent third party in any cloud storage deployment, using a global Anycast network to accelerate traffic, improve security, and boost performance for businesses utilizing cloud storage solutions. Cloudflare also helps minimize vendor lock-in by supporting multicloud implementation and reducing the number of operational changes that need to be made at the network level.
Glossaire sur le cloud
Navigation dans le centre d’apprentissage