Function Overview

OBS provides a bunch of tools, including OBS Browser+, obsutil and obsfs, for data migration and management in different scenarios.

You can use these tools for OBS resource management, including creating buckets, mounting parallel file systems, and uploading and downloading objects.

OBS provides REST APIs that support HTTP/HTTPS methods. You can call these APIs to create, modify, and delete buckets, as well as to upload, download, and delete objects.

OBS provides SDKs for secondary development. The SDKs are available in the following programming languages: Java, Python, C, Go, BrowserJS, .NET, Android, iOS, PHP, and Node.js.

Buckets are containers that store objects in OBS.

Each bucket has its own attributes, such as the access permission, storage class, and region. You can specify these attributes when creating buckets. You can also configure advanced attributes.

OBS provides easy bucket management. You can conveniently create, list, search for, view, and delete buckets.

You can replicate the settings of an existing bucket to the bucket you are creating. The configurations that can be replicated include bucket policies, CORS rules, lifecycle rules, back-to-source rules, image processing styles, and online decompression rules.

Data uploaded to OBS is stored as objects (fundamental units) in buckets.

An object consists of object data and metadata that describes the object's attributes. You can perform different operations on objects, including upload, download, listing, searching, resumable transfer, and multipart uploads.

Parallel File System (PFS), provided by OBS, is a high-performance file system, with access latency in milliseconds. PFS supports TB/s-level bandwidth and millions of IOPS, which is ideal for processing high-performance computing (HPC) workloads.

You can call OBS APIs to read data in a parallel file system. You can also use obsfs (an OBS tool) to mount a parallel file system to a Linux cloud server. This way, you can manage files and directories in the parallel file system just as you would operate a local file system.

Image Processing, a feature provided by OBS, is stable, secure, efficient, easy-to-use, and cost-effective. You can use it to process images stored in OBS anytime and anywhere.

A wide range of image processing operations are available, including slimming, cropping, resizing, watermarking, and format conversion.

OBS provides the Standard, Infrequent Access, and Archive storage classes for buckets and objects, which meet different requirements for storage performance and cost.

The Standard storage class features low access latency and high throughput. It is therefore ideal for storing a massive number of hot files (frequently accessed every month) or small files (less than 1 MB). Its application scenarios include big data analytics, mobile apps, hot videos, and social apps.

The Infrequent Access storage class is ideal for storing data that is accessed infrequently (less than 12 times a year) and has quick response requirements. Its application scenarios include file synchronization, file sharing, and enterprise backup. It provides the same durability, access latency, and throughput as the Standard storage class but at a lower cost. However, the Infrequent Access storage class has lower availability than the Standard storage class.

The Archive storage class is ideal for archiving data that is rarely accessed (once a year on average). Its application scenarios include data archiving and long-term data backups.The Archive storage class is secure, durable, and inexpensive, and can be used to replace tape libraries. However, it may take hours to restore data from the Archive storage class.

When creating a bucket, you can specify a storage class for it. Similarly, when uploading an object, you can specify a storage class too. You can also configure a lifecycle rule to automatically change the storage classes of objects.

OBS uses IAM permissions, bucket/object policies, and ACLs to manage permissions.

You can grant different access permissions to different accounts and users, and configure bucket/object policies or ACLs to control read and write permissions for buckets and objects.

You can bind a domain name to an OBS bucket and use the domain name to access data in the bucket.

For example, if you need to migrate files from a website to OBS while keeping the website address unchanged, you can bind the website domain name to an OBS bucket.

Cross-region replication replicates objects, automatically and asynchronously, from a source bucket to a destination bucket in another region, with the precondition that the source and destination buckets are under the same account.

This provides you with the capability for disaster recovery across regions and remote data backup.

Static websites usually contain static web pages and scripts that can run on clients, such as JavaScript and Flash.

You can upload the content files of a static website to an OBS bucket, grant anonymous users with the read permission to these files, and configure static website hosting for the bucket to host them. When the configuration takes effect, the static website can be accessed through the access domain name of the bucket.

Cross-origin resource sharing (CORS) is a browser-standard mechanism provided by the World Wide Web Consortium (W3C). It defines the interaction methods between client-side web applications in one origin and resources in another. For general web page requests, website scripts and contents in one origin cannot interact with those in another because of Same Origin Policies (SOPs). OBS supports CORS rules and allows resources in OBS to be accessed across origins.

You can use JavaScript and HTML5 to develop web applications that can directly access resources in OBS, without using proxy servers.

You can also use the dragging function of HTML5 to upload files directly to OBS (with the upload progress displayed) or update OBS content using web applications.

External web pages, style sheets, and HTML5 applications hosted in different origins can access web fonts or pictures stored in OBS, implementing resource sharing.

Some rogue websites may steal links from other websites to enrich their content without any costs. Link stealing hurts the interests of the original websites and it is also a strain on their servers. URL validation is used to resolve this problem.

URL validation protects your data in OBS from being stolen using the Referer field in HTTP requests. Such authorization is controlled using whitelists and blacklists.

You can share a file or folder stored in OBS with all users by using a temporary URL. The sharing is temporary, so all shared URLs have a validity period.

OBS Console and OBS Browser+ are available for you to temporarily share objects. To permanently share objects, you can configure a bucket policy or an object policy.

By default, objects stored in the Archive storage class need to be restored before being downloaded. After you enable direct reading for a bucket, you can directly download Archive objects from the bucket without restoring them first. Direct reading is a billable function.

When creating a bucket, you can select multi-AZ storage for the bucket to store data redundantly in multiple AZs, improving data reliability. OBS ensures data redundancy by using the Erasure Code (EC) algorithm, instead of multiple copies.

Once configured, multi-AZ storage cannot be changed after the bucket is created.

OBS provides server-side encryption with KMS-managed keys (SSE-KMS) and server-side encryption with customer-provided keys (SSE-C) to enhance data security.

Enabling server-side encryption for a bucket ensures that any objects uploaded to the bucket are encrypted into ciphertext before being stored on the server, and then are decrypted into plaintext on the server before being downloaded.

When uploading an object, you can choose whether to enable server-side encryption.

OBS supports bucket encryption. After a bucket is encrypted, all objects uploaded to the bucket are automatically encrypted.

You can enable the default encryption when creating a bucket or for an existing bucket. OBS encrypts only the objects uploaded after the default encryption function is enabled. The statuses of objects uploaded to the bucket before the default encryption function is enabled remain unchanged.

Lifecycle management allows you to configure rules to automatically delete objects or transition objects between storage classes at a scheduled time.
With lifecycle management, you can:
Regularly delete uploaded log files that are meant to be retained for only a specific period of time (one week or one month).
Transition documents that are seldom accessed to the Infrequent Access or Archive storage class or delete them.
Schedule the deletion of a large number of objects from a bucket.

OBS provides tags for you to identify and classify OBS buckets. If you add tags to a bucket, service detail records (SDRs) generated for it will be labeled with these tags. You can classify SDRs by tag for cost analysis. For example, if you have an application that uploads its running data to a bucket, you can tag the bucket with the application name. In this manner, the costs on the application can be analyzed using tags in SDRs.

Bucket inventories help you manage objects. You can configure an inventory rule to let OBS periodically scan all objects or a set of objects with the same prefix in a bucket. Then, OBS generates the metadata (including the object size, modification time, and storage class) of the scanned objects and saves it as a CSV file to a specified bucket.

You can enable logging for buckets for analysis or auditing to analyze the characteristics, types, and trends of access requests sent to your buckets. With logging enabled, OBS automatically logs access requests for the bucket, and writes the generated log files to the specified bucket (or target bucket).

OBS provides the AppendObject API for you to write additional data to an appendable object in a specified bucket.

An object created by calling the AppendObject API is an appendable object, while an object created by calling the PutObject API is a normal object (not appendable). A normal object's content can only be read. In scenarios where data is continuously generated in real time, you can upload and store data as appendable objects in a bucket, so that newly generated data can be contiguously written to the objects and can be read upon being successfully written.

Object metadata is a set of name-value pairs that describe an object's attributes, which is used for object management.

The metadata is classified into the following types: system-controlled and user-controlled. Metadata like Last-Modified is controlled by OBS and cannot be modified, but metadata such as ContentLanguage and Expires are user-controlled and can be modified.

You can add, modify, or delete user-controlled metadata for uploaded objects.

You can configure a storage quota for a bucket to limit how much data it can hold. The maximum value allowed for this quota is as follows: 2⁶³ - 1 in bytes. By default, a newly created bucket has no such limit imposed on it.

In a multipart upload, fragments will be generated when the upload fails in case of the following and other possible events: poor network conditions, disconnections between OBS server and your local ends, manual interruptions during upload, device faults, and unexpected power outage.

Fragments are stored in OBS buckets. You are advised to regularly clear fragments to free up storage space.

Enterprise projects are used for centrally managing resources and services. Resources and services in different regions can be added to the same enterprise project. For example, you can create several enterprise projects based on your company's divisions or projects, and then add your cloud resources and services to these enterprise projects accordingly.

When creating a bucket, you can specify an enterprise project to which the bucket will belong, to facilitate bucket resource and permission management.

When versioning is enabled for a bucket, OBS can keep multiple versions of an object in the same bucket. That way you can quickly retrieve and restore every object version as needed, or recover data from both accidental actions and application failures.

By default, versioning is disabled for new OBS buckets. In this case, if a newly uploaded object is using the name of the previously uploaded one, the new object will overwrite the previous one.

You can create an IAM agency to delegate a cloud service or another Huawei Cloud account to implement O&M on your resources based on the assigned permissions.

For example, to implement cross-region replication in OBS, you need to delegate OBS to help you replicate objects from a source bucket to a destination one.

When using OBS, you may send PUT and GET requests continuously, which generates upload and download data traffic, and you may receive error responses from the server. Cloud Eye can perform automatic and real-time monitoring for your buckets. It triggers alarms and notifications for operations, making it easy for you to understand your bucket access requests, data traffic, and error responses, so that you can optimize or plan resource usage.

Cloud Trace Service (CTS) is a log audit service provided by Huawei Cloud. CTS collects, stores, and queries records of operations on buckets and objects in OBS, facilitating security analysis, compliance audit, resource tracking, and fault locating.

With a back-to-source rule, if the data you requested is not found in OBS, OBS automatically pulls the data from its origin server and returns the data to you.

The following figure describes how back to source works.

With an online decompression policy, you can compress multiple files into a ZIP package and upload it to OBS for auto decompression.

You can configure write-once-read-many (WORM) policies on OBS Console to prevent objects from being deleted or tampered with during a particular period of time.