Se hai bisogno di aiuto, fai clic qui:

The Challenge of Mass Unstructured Data.

According to Huawei's Intelligent World 2030 report, 1 YB — yottabyte — of data will be generated annually worldwide by 2030, representing a 20-fold increase over 2022. That's equivalent to the capacity of four trillion 256 GB mobile phones.

Undoubtedly, we are entering a new era, one where more than 80% of data is unstructured. Here, evolving big data and Artificial Intelligence (AI) technologies have paved the way for extracting value from this unstructured data and it's time for enterprises to consider how to properly back up and archive it, including long-term retention in line with regulatory compliance.

  • Insufficient capacity for mass data

    • Traditional scale-up storage arrays cannot keep up with exponential data growth. • Unstructured data requires long-term retention in accordance with regulatory compliance. • Mass data storage often leads to soaring Information Technology (IT) costs.

  • Data Silos for Different Backup and Archiving Applications

    • Different applications use different access protocols, leading to siloed deployment that restricts data mobility. • Applications are characterized by different read-write performance models, causing mutual interference that negatively impacts the user experience.

  • Reliability for Valuable Unstructured Data

    • Unstructured data is now a key digital asset that requires higher levels of protection and reliability. • While nearby data access is required to ensure the user experience, cross-site data access also needs to be more reliable.

Object Storage Is a New Way to Store Mass Data

  • object en 1

    Easier Access

    Object storage allows different devices to access data, from anywhere and at any time. Unlike Network-Attached Storage (NAS), which involves complex protocols and network interactions, object storage reads and writes data via RESTful Application Programming Interfaces (APIs), eliminating cyber security risks and allowing for data transmission over public networks. Thanks to the development of the Internet, a wider range of devices support S3.

  • object en 2

    Faster Read

    With a flat data structure, object storage delivers higher read performance when handling mass data. In traditional file storage, data is stored as files in folders and organized in a tree structure, similar to putting items in different drawers of different cabinets in different rooms. In contrast, object storage directly reads files through URLs, akin to putting all items in one large drawer and simply opening it to directly select items. Tests show that, with 100 billion files, object storage delivers 10 ms read speeds, compared to the seconds that file storage needs.

  • object en 3

    Quicker Search

    Unlike traditional file storage , user-defined metadata is a part of the object itself, meaning that users can customize tags for each object. This makes it easier to search mass data for objects with the same tag. To implement a similar function, traditional file storage needs an additional, separate tagging and indexing system.

Huawei OceanStor Pacific Backup and Archive Storage

Running on a distributed object storage foundation, Huawei OceanStor Pacific provides End to End (E2E) data infrastructure services for mass unstructured data, such as videos and images, covering production, backup, and archiving. It supports hybrid workloads and offers unique features like multi-protocol interworking, multiple active data centers, deduplication, and compression.

Cutomer Benefits

  • Large-Scale Deployment at Lower Costs

    • High-ratio Erasure Code (EC) with a maximum utilization rate of 91.6% for optimized Return On Investment (ROI).

    • An ultra-high-density capacity design, with 120 x 3.5-inch disks in a 5U space, unlocks the value of every single rack unit in a data center.

    • Industry-leading deduplication and compression increases effective capacity.

  • Multiple Active Data Centers for Higher Reliability

    • Up to 12 active data centers, tolerating the simultaneous failure of two data centers.

    • Write Once Read Many (WORM), encryption, and enterprise-class protection technologies guard against accidental deletion.

    • Failure of up to four nodes tolerated in a single data center, with failover falling below 10s.

  • Elastic Expansion for Better Agility

    • Exabyte-level capacity expansion, up to 4096 nodes.

    • Elastic expansion of unlimited objects, with support for 100 billion objects/buckets.

    • Consistent, high performance maintained even with 100 billion objects.

Related Products

  • 01

    OceanStor Pacific 9950

    High-density performance model

    5 U, 8 nodes

    80 NVMe SSDs

    160 GB/s bandwidth, 2 million IOPS

  • 02

    OceanStor Pacific 9920

    Performance model

    1 node per 2 U chassis

    All-flash distributed storage

    Superior performance and flexible component configurations

  • 03

    OceanStor Pacific 9550

    High-density capacity model

    5 U, 120 x 3.5-inch SATA disks

    1.68 PB of raw capacity

    Elastic EC

  • 04

    OceanStor Pacific 9520

    Capacity model

    1 node per 2 U chassis

    Distributed storage

    Multiple device types and flexible component configurations

  • 05

    OceanStor Pacific 9540

    Capacity model

    1 node per 4 U chassis

    Distributed storage

    High capacity density and flexible component configurations