Data has been regarded as the most important asset of the modern enterprise. Based on how enterprises use their data assets in daily production and business activities, modern enterprise business intelligence (BI) maturity can be divided into two phases, and five maturity stages.

No matter what stage an enterprise is at, it is driven to move towards a higher stage by management and service demands. As such, the following trends become apparent:

Focus Changes onto Backline: Large Volume of Structured Data Needs to Be Processed

• The focus of IT enterprises changes from online transaction support to operational analysis.
• Data warehouses are becoming the new investment target of enterprises.

Data Warehouse: From Reports to BI

• Traditional data reports are yesterday's news. Today, real-time data reports are preferred.
• Custom and flexible query modes are also preferred.

Big and Fast: Keep Up with Current Demand

• Big: With sky-high data growth, larger volumes of data need to be processed. The Big Data era has come.
• Fast: Real-time monitoring and analysis on services is required.

As business and data grow, the enterprise IT infrastructure is facing the following challenges in data warehouse scenarios:

• The increasing demand on traditional IT infrastructure to expand capacity requires high costs.
• The technical architecture of traditional servers and storage devices faces an imminent performance bottleneck.
• The variety of devices and platforms requires complex and expensive O&M inputs.

The "traditional databases + midrange computers + high-end arrays" method is no longer dominant in the competition for cost-effectiveness. The scalability of traditional symmetric multiprocessing (SMP) is reaching its upper limit. New IT architecture is needed to meet the increasing demands of processing large volume of data. To ensure service continuity during the construction of new architecture, enterprises have to take measures to minimize adverse impacts on applications. However, the cost of vertical data warehouse integration is out of the range of many enterprises. Therefore, Huawei Data Warehouse Acceleration Solution, based on the Huawei FusionCube system, is developed and optimized to better support the evolution of traditional data warehouses.

For Large Data Warehouses

Challenge: Large data warehouses still have performance bottlenecks and are difficult to expand.

Solution: Use the Huawei FusionCube system as the infrastructure to build data marts and reporting server clusters outside of the existing data warehouses, thereby unloading the primary data warehouses.

For Small- and Medium-sized Data Warehouses:

Challenge: The current data scale is about 10 TB to 60 TB. To process an increasing data volume, enterprises need to create new data warehouses or take measures to improve the performance of existing data warehouses.

Solution: Use the Huawei FusionCube system as the primary data warehouse.

Based on the Huawei-developed distributed storage system and the converged architecture of hardware and software, Huawei FusionCube is optimized for data warehouse scenarios to provide greatly improved I/O performance. It is compatible with mainstream data warehouses in the industry and is easy to manage, easy to expand, and easy to configure.

• FusionCube converged hardware platform: integrates computing, storage, and switching resources to provide optimized infrastructure and application performance. It also provides high-bandwidth, short-delay, and multi-protocol-supported switching capabilities.
• FusionStorage: uses resource scale-out architecture and deeply integrates computing and storage resources to provide high performance and scalability.
• InfiniBand (IB) switching: provides high-speed interconnection, reduces latency, and increases bandwidth.
• PCI-E solid-state drive (SSD): improves random I/O performance.
• Unified management: uses unified entries and authentication. Furthermore, all physical servers and VM resources can be managed in a centralized manner.

In the case of a customer's core financial data warehouses, Data Warehouse Acceleration Solution helped the customer to accelerate its existing primary data warehouses by offloading workloads to data marts provided by FusionCube. In this case, the performance of the existing primary data warehouses was increased when large volume of data needed to be concurrently processed. FusionCube also helped the customer to reduce subsequent expansion and maintenance costs by around 10 million RMB.

In a comparison test against peer data warehouses, Data Warehouse Acceleration Solution was proven to provide a 27% lower query response latency for 1000 concurrent users and a faster report query speed by a factor of 3 to 9. Therefore, Data Warehouse Acceleration Solution better meets the requirements for enterprise development.