The high-resolution Earth observation system is one of the major special programs in the National Medium- and Long-Term Program for Science and Technology Development (2006–2020). It focuses on developing a high-resolution Earth observation system based on satellites, aircraft, and stratospheric airships, and, combined with other medium- and low-resolution observation means, forming a spatially and temporally coordinated, all-weather, all-time Earth observation system.

Zhejiang University is responsible for developing the “data-service grid platform” in the high-resolution Earth observation major special program. The platform requires integration and development of multiple technologies, including security authentication across multiple security domains, highly reliable concurrent distributed data distribution, automatic registration and intelligent discovery of massive data and resources, and high-timeliness resource monitoring and integrated resource management, as well as key technical work on grid portal services, grid security authentication, grid data integration, grid data-distribution scheduling, grid resource monitoring, and grid statistics and sharing.

Customer Requirements

The spatial big-data storage technology validation system is an important part of data-grid technology validation in the high-resolution grid R&D process, and carries data sources for the other systems of the high-resolution grid. The main task of spatial big-data storage validation is to determine, in a multi-node, massive-data environment, whether the system can well meet response and throughput needs and well support the service grid and upper-layer applications.

Solution

Sugon provided the customer with two ParaStor300 storage systems, meeting rapid data-growth requirements. Capacity can be expanded online as business needs grow, while performance grows linearly. A B/S management interface offers graphical operation that is easy to learn and can connect with the customer’s related business-management systems for unified management. It supports L2 cache acceleration, data tiering, data archiving, and remote backup, enabling full-lifecycle management of data in one storage system from ingest, screening, cleansing, and distribution through processing, transformation, classified application, archiving, and backup. ParaStor supports different combinations of multi-replica and erasure coding, and supports fault tolerance of up to 4 disks and data-node faults on a single node, meeting safety and reliability requirements of different data under multiple businesses. ParaStor’s small-file aggregation supports fast recovery management of massive small files and lowers the risk of file loss. Built-in archiving and remote backup can complete data archiving and disaster-recovery backup at low cost, greatly reducing system investment compared with the expensive remote disaster-recovery backup of traditional SAN architectures. Automatic power control makes it convenient to save energy and lower cost as the cluster scales in the future.

Solution Value

With strong professional customization capability, Sugon built a ParaStor cloud-storage solution that effectively supported validation of Zhejiang University’s data-service grid platform. Deep performance tuning fully met the customer’s requirements for system performance and stability.