Three environments where enterprises store large data

The deployment of large data requires a combination of specific application scenarios. In fact, http://www.aliyun.com/zixun/aggregation/13861.html "> The storage of large data can be explained by the story of" three piglets building houses "(using straw, wood and bricks, respectively) This story can reflect more vividly the different levels of protection (integrity and reliability) that correspond to the delivery service (cost) in the data storage environment.

Financial data, external reporting, and compliance data need to be stored in a brick house (bricks) environment. This data requires a reliable hardware infrastructure and is consistent with its original source. It is common for multiple functional departments in an enterprise to use financial data such as product services pricing decisions, sales performance and analysis, and critical employee/Management compensation incentive calculations.

The carefully designed "wood room" (STICK) environment ensures that storage data is strong and durable. This environment is dedicated to applications and not designed for enterprise-class use and cross-functional data sharing. This data type can be used specifically for data transformations, typically including a large number of marketing data marts. Only necessary functions such as data transformation, coordination, and lineage can be used for specific business purposes. Compared with the above "Brick house", "wooden house" in essence, the cost is lower, the speed is quicker.

Finally, "thatched cottage" (HAY) is introduced. "Cottage" is actually refers to the use of data on a specific date to the data conversion, grouping and summary. In this case, the data may exist in the original source data format, and almost no data structure is required. The user can adjust the data format arbitrarily. Although "thatched cottage" design cannot be easily reproduced or vertically expanded, it is suitable for dealing with ad hoc, non repetitive business issues. The program has low demand for data coordination and replication.

The analogy of using "Three Little Pigs" is fairly straightforward, but the specific solution should refer to the data governance guidelines. If you can handle it, the business unit wants to quickly get low-cost solutions, while IT departments need to rely on reliable solutions to provide sound, reliable services. This is an inherent contradiction in most discussions in the business and IT departments.

"Cottage" solutions are a concern because of the rapid deployment, low cost and low cost of failure. Under the new economic mechanism, especially in the self-service environment, users ' recognition of data (including large data) is the reason for the rapid development of the data laboratory and the exploration environment. It is therefore not surprising that business units choose Fast, low-cost solutions.

But the cost of the IT department was shocking when it was upgraded to a "wood-house" or "brick-house" environment. "Why can't they use the solution we designed in two weeks?" They can. But in the "cottage" based on the deployment of "brick house" or even "wooden house" plan is not workable. The use of "cottage" design plans to deploy the "Wood House" and "Brick House" program, IT department will waste a large number of budget.

The main challenge is to identify data control strategies and processes that are important to data. When a "creative" solution designed in a "cottage" environment needs to be migrated to a more stable environment, the person responsible for the decision to participate in the data management (thatched cottage, wood or brick house) needs to have a comprehensive understanding of the importance of downstream data.