“Smart Data” Cloud Warehouse
● Solution Overview
"Smart Data" cloud warehouse is a high-performance completely-hosted PB-level cloud data warehouse service that enables users to analyze massive data with ease.
The cloud data storage service provided by "Smart Data" is for carriers and enterprises to empower them the big data storage and analysis capability and enable them to store, analyze and compute massive data with ease. With this solution, various data generated during operation can be monetized.
"Smart Data" cloud data warehouse is integrated with the query capability of the MPP architecture, flexibility of the big data platform, and elastic resources of cloud computing, which resolves the insufficiencies in traditional data warehouses and the big data platform.
● Technical Architecture
● Product Values
The high-performance completely-hosted PB-level cloud data warehouse service enables users to analyze massive data with ease.
When using "Smart Data" cloud data warehouse products deployed in the cloud resource pool, organizations can start a data warehouse cluster that consists of dozens even hundreds of nodes in several minutes, and begin data analysis tasks as soon as data is loaded. Because this is a completely-hosted cloud service, "Smart Data" cloud data warehouse is responsible for all work, including cluster resource configuration, data backup, monitoring & audit, error correction, HA, upgrade, and complex and error-prone O&M, allowing users to focus on service analysis.
● Performance Advantages
● Deployment Case
Project name: Internet access log retention project of a branch of China Mobile
Background: A branch of China Mobile uses "Smart Data" cloud data warehouse solution to build an Internet access log query system that enables the branch to implement Internet access log retention, query, reporting, and account monitoring functions throughout the network.
User quantity: 6 million
NAT log record: 100 GB/d
ADIUS log record: 500 GB/d
DPI log record: 50 TB/d
CDN log record: 5 TB/d
Effect: The hardware cost has been reduced significantly, and the number of devices has been reduced by 60%.
● Other Cases
Big Earth Data Engineering Data Platform Service of Chinese Academy of Sciences
With the efficient concurrent executing engines embedded in the "Smart Data" cloud warehouse, rich spatio-temporal data processing functions （including vector data and grid data）, and object storage-based data persistence policies, the big data cloud platform service can perform PB-level data analysis of geographic information with ease to support the entire scientific platform of the digital earth.
Great Amount of IoT Data Processing Service of XX
XX has deployed a great number of sensors in factories in China. These sensors produce a great amount of data. Regarding how to use the IoT data efficiently, XX faces the following challenges:
1.Computing resources and storage resources are insufficient.
2.To process structured data and a great amount of unstructured data （JSON and XML）, multiple sets of systems must be deployed.
3.The density of the unit value of data generated by the sensors is not high, but the cost of storage based on the existing solution and the analysis and processing cost are high, especially if real-time analysis is required.
4.With the great amount of data, existing solution poses the performance problem.
After "Smart Data" cloud data warehouse provides a new solution for XX, the following achievements are made:
1."Smart Data" is deployed as an under-layer platform of the entire data analysis system to provide elastic computing and storage resources to process the rapidly-growing data.
2.With the high-performance distributed executing engines of "Smart Data"， the speed of complex query is increased by 10 times.
3.By using object storage as the data persistence layer, "Smart Data" not only the response time requirement of user analysis is met, but also the use cost is reduced significantly.