Alibaba Cloud Object Storage Service (OSS) is a secure, cost-effective, and highly durable cloud service for storing massive amounts of data. It is designed for at least 99.9999999999% (12 nines) of data durability and 99.995% of service availability. OSS provides easy-to-use APIs to store and access data from any application, anywhere, at any time.
Storage classes
OSS provides multiple storage classes, including Standard, Infrequent Access (IA), Archive, Cold Archive, and Deep Cold Archive, to comprehensively cover various data storage scenarios ranging from hot to cold data.
|
Storage class |
Description |
|
Standard |
Provides highly reliable, highly available, and high-performance object storage services that support frequent data access. Suitable for business scenarios such as images and audio and video files for social networking and content sharing, large websites, and big data analytics. |
|
Infrequent Access (IA) |
Provides highly durable object storage services at lower storage costs. Has a minimum billable size (64 KB) and a minimum storage duration (30 days). Supports real-time data access, and a data retrieval fee is charged when you access data. Suitable for business scenarios in which data is accessed infrequently (an average of one to two times per month). |
|
Archive |
Provides highly durable object storage services at extremely low storage costs. Has a minimum billable size (64 KB) and a minimum storage duration (60 days). You must restore data (which takes about 1 minute) before you can access it, and a restoration fee is charged. If you enable real-time access of Archive objects, you can access data without restoration, and a real-time retrieval fee is charged. Suitable for scenarios in which data is stored for a long period of time, such as archival data, medical images, scientific materials, and video and film footage. |
|
Cold Archive |
Provides highly durable object storage services at lower storage costs than the Archive storage class. Has a minimum billable size (64 KB) and a minimum storage duration (180 days). You must restore data before you can access it. The restoration time depends on the size of the data and the selected restoration mode, and a data retrieval fee is charged during restoration. Suitable for cold data that must be stored for an ultra-long period of time, such as data that must be retained for a long period of time for compliance requirements, raw data accumulated over a long period of time in the big data and AI fields, media resources retained for a long period of time in the film and television industry, and archived videos in the online education industry. |
|
Deep Cold Archive |
Provides highly durable object storage services at lower storage costs than the Cold Archive storage class. Has a minimum billable size (64 KB) and a minimum storage duration (180 days). You must restore data before you can access it. The restoration time depends on the size of the data and the selected restoration mode, and a data retrieval fee is charged during restoration. Suitable for extremely cold data that must be stored for an ultra-long period of time, such as raw data accumulated over a long period of time in the big data and AI fields, media data retained for a long period of time, regulatory and compliance archives, and tape replacement. |
For more information about the preceding storage classes, see Overview of storage classes.
Use cases
OSS is ideal for the following use cases:
Store and distribute static website content, audio, and video
Each object stored in OSS has a unique HTTP URL for content distribution. You can also use an OSS bucket as an origin for a content delivery network (CDN). Because OSS does not require partitioning, it is ideal for hosting data-intensive websites with user-generated content, such as image and video sharing sites. Various client devices, web applications, and mobile apps can read from or write to OSS directly. OSS supports both streaming and file uploads.
Host static websites
OSS provides a low-cost, highly available, and highly scalable solution for storing static content such as HTML files, images, videos, and client-side scripts like JavaScript.
Data storage for computing and analytics
The horizontal scalability of OSS lets you access data from multiple compute nodes simultaneously, unconstrained by a single node.
Data backup and archiving
OSS provides a highly available, scalable, and secure solution for backing up and archiving critical data. You can configure lifecycle rules to automatically transition cold data to infrequent access or archive storage classes to reduce storage costs. You can also use cross-region replication (CRR) to asynchronously replicate data between buckets in different regions in near-real-time to support cross-region disaster recovery.
Performance
For the fastest access to data in OSS, your Elastic Compute Service (ECS) instance should be in the same region as your bucket. By design, the server-side latency of OSS is negligible compared to network latency. The scalability of OSS in terms of storage capacity, request rate, and number of users makes it well-suited to support large-scale web applications. If you use multiple threads, applications, or clients to access OSS at the same time, the aggregate OSS throughput typically scales far beyond the rate that any single server can generate or consume.
To improve upload performance for large objects (over 5 GB), OSS provides the multipart upload feature. This allows you to split a large object into multiple parts and upload them independently. After all parts are uploaded, OSS combines them into a single object to enable resumable uploads. Multipart upload is ideal for poor network conditions because if an upload fails, you can retry only the failed part instead of re-uploading the entire object.
To speed up data access, many developers use OSS with search services such as OpenSearch or databases such as Tablestore and RDS. OSS stores the actual data, while the search service or database stores metadata, such as keys, sizes, and keywords. Metadata stored in a database is easy to index and query. This combination helps you precisely locate and retrieve objects in OSS.
To accelerate large file uploads and downloads over long distances, and to speed up dynamic file updates and non-hot file downloads, OSS also provides transfer acceleration. It delivers an end-to-end acceleration solution through an intelligent scheduling system, optimized transfer links, a tuned protocol stack and transfer algorithms, and deep integration with OSS server-side policies.
To accelerate downloads of static, hot files, you can integrate OSS with Alibaba Cloud CDN. Alibaba Cloud CDN uses an OSS bucket as an origin and distributes content to its points of presence (POPs). With a precise scheduling system, Alibaba Cloud CDN routes user requests to the most suitable POPs, enabling end-users to retrieve content as quickly as possible. This helps mitigate internet congestion and improves response times.
Data durability and service availability
OSS uses a data redundancy mechanism that stores your data across multiple availability zones (AZs) within the same region. If one availability zone becomes unavailable, your data remains accessible. For buckets that use this multi-AZ mechanism, OSS provides 99.9999999999% (12 nines) of data durability and 99.995% of service availability.
You can also enable cross-region replication (CRR) for your buckets. After you enable CRR, data is automatically and asynchronously replicated in near-real-time between buckets in different regions to meet cross-region disaster recovery requirements. Both the source and destination buckets provide 99.9999999999% (12 nines) of data durability and 99.995% of service availability.
Scalability and elasticity
OSS has no upper limit on the total storage capacity or the capacity of a single bucket. You can store an unlimited number of objects in a bucket. OSS automatically stores replicas of your data on different servers in the same region, and all data replicas benefit from the high-performance infrastructure of Alibaba Cloud.
Security
Alibaba Cloud Object Storage Service (OSS) provides comprehensive security capabilities, including server-side encryption, client-side encryption, hotlink protection using Referer whitelists, fine-grained access control, log auditing, and WORM retention policies. In the Chinese mainland, OSS is the only cloud service audited by Cohasset Associates to meet strict electronic record-keeping requirements, such as SEC Rule 17a-4(f), FINRA 4511, and CFTC 1.31. For more information about OSS security capabilities, see the OSS Security and Compliance White Paper.
API
OSS provides an API to store objects in buckets. A bucket is a top-level container for your objects, and its name must be globally unique. Each object has a key (filename) that serves as its unique identifier within the bucket. OSS does not have a native concept of folders; all items are stored as objects. However, you can simulate a folder structure by using a forward slash (/) as a delimiter in your keys, such as folder1/folder2/file.
Developers typically use tools or SDKs that encapsulate the APIs to build applications. OSS currently provides SDKs for more than 10 programming languages, including Java, Python, PHP, Go, Android, and iOS. The ossutil command-line tool provides a rich set of commands to manage buckets and objects, including ls, cp, cat, and config. It delivers high performance, supports concurrent uploads, and runs on Windows, Linux, and macOS. You can also use graphical tools like ossbrowser 1.0 to browse files, upload and download files and folders (directories) with a simple, visual interface, and manage your OSS resources with ease.
You can also use the event notification feature to stay informed about operations on your OSS resources. For more information, see Event notifications.
Billing model
OSS fees consist of charges for storage, data transfer, API requests, and data processing. By default, OSS uses a pay-as-you-go billing model, where you are charged hourly for the resources you actually consume. For some billable items, you can also choose a subscription model to further reduce costs. A subscription involves purchasing a resource plan in advance, and your usage is then covered by the plan's quota. For more information, see Billing overview.