Shulou(Shulou.com)06/03 Report--

[this article is excerpted from Li Fudong "big data Operation" 3.6. for more information, please follow Wechat official account: Li Fudong Channel)

Compiling according to

Big data's service process includes: service catalog management, capacity management, availability management, continuity management, service level management, information security management, supplier management and so on.

Text

In terms of design methods, there are both differences and connections between big data's service and the service that supports enterprise operation. The difference is that the design of big data's service mainly takes "data" as the reference point, and the more types, richer and fresher the "data" is, the more helpful it is to design a good service; what the two have in common is that big data's service, in the final analysis, serves the operation of the enterprise, in order to enhance the ability of the enterprise in construction, marketing, product sales, customer service, enterprise management, and so on.

The process of big data service in the design stage includes: service catalog management, capacity management, availability management, service level management, information security management, supplier management and so on.

1.1.1 big data service directory management

The service catalog is equivalent to the ordering menu in the restaurant. Users can see what services are available through the service catalog, and managers can also view the resources on which the service depends, and then calculate the cost-effectiveness of the service.

With the increase of the number of big data services, it is necessary to carry out hierarchical and classified management in order to quickly retrieve and locate big data services. Similarly, big data service will continue to be optimized and improved, so it is necessary to distinguish the way big data service adds version tags.

Big data service catalogue can be classified and organized according to the business applications supported by big data services. For example, first-level big data services can be divided into investment and construction, market operation, resource operation, administrative synthesis and enterprise management. Can be further subdivided on the basis of the first level, such as market operation category can be subdivided into marketing, sales, customer service and billing.

According to this classification, the location of the business application supported by big data service can be defined, and users can find big data service more efficiently. For example, if the goal of a big data service is to support the network planning and design of an enterprise, then the big data service that meets this requirement should be found in the investment and construction category.

1.1.2 big data service capacity management

Capacity is the throughput of an organization's IT resources to provide services. The capacity metrics provided by IT resources include the maximum number of concurrent users supported, the maximum number of online users, the maximum computing power of the server, the maximum storage space, the maximum network egress bandwidth, and so on.

Capacity management is important not only for IT service design, but also for big data service design. Big data services are usually provided in three forms, and different types of big data services have different requirements for capacity.

The first category is the big data service that supports operational decision-making. Such big data service needs to be embedded in the production process, and users will also invoke big data service in the process of using production applications. Therefore, big data service is required to provide the ability to ensure business continuity. This type of big data service is similar to the capacity requirements of operation-oriented transactional applications, so when designing the capacity of this kind of big data service, it can be regarded as a transactional service.

The second type of big data service belongs to statistical analysis type, this kind of big data service is more to satisfy the middle managers of enterprises to count the data of a certain period of time, such as product sales data in a certain quarter, cash flow in a certain year, and so on. Assist managers to find problems in production and operation. Because such big data services are not embedded in production applications, the real-time requirements are not so high. When designing the capacity of this kind of big data services, we mainly consider the capacity requirements of a specific period of time, such as the beginning and the end of the month.

The third type of big data service is mainly aimed at the senior strategic managers of the enterprise, such as the general manager and strategic planner of the enterprise, who usually pay attention to the medium-and long-term planning for more than half a year. We need to grasp the market situation and the gap between big data and our competitors with the help of big data service. Such big data services usually do not require high response time, pay more attention to the hidden rules behind the data, and focus on decision-making models. Because this kind of big data services often need to be analyzed based on years of historical data, we can consider adopting a cloud-based infrastructure to flexibly adapt to the growing demand for supporting capacity.

1.1.3 big data service provider management

In the era of big data, data has become the core asset of enterprises, and because of social specialization, data is bound to be scattered in different enterprises. Big data service is the same as other raw materials in the enterprise, if it can not be guaranteed to provide timely and accurate, it will reduce big data's service capacity. It can be seen that the effective management of big data service suppliers is also very important.

The introduction of big data service by enterprises is the same as the introduction of production services by enterprises, which needs to be managed effectively. For example, big data service provider access and exit management, service quality management, service performance management and so on. The purpose of supplier management is that enterprises can timely and effectively obtain big data services that meet the requirements, including the timeliness of data provision, data quality and so on.

The supplier access management mainly requires the supplier big data's service provision ability to reduce the risk of big data service provision. the enterprise can sign the big data service supply contract with the supplier, legally guarantee the loss caused by the supplier's failure to provide services in accordance with the requirements, and reduce the production and operation risk of the enterprise. Regular assessment of the big data service provided by the supplier is used as the basis for big data's service withdrawal.

1.1.4 big data service security management

While we regard big data as the core asset of the enterprise, it also means that the data has extraordinary value and role in the enterprise. In addition, big data is different from other assets of the enterprise, for example, the customer data of the enterprise will involve the privacy of the individual or the enterprise, and may involve the trade secret of the enterprise.

In order to ensure the security of big data's service, we need to start from the following three aspects.

The first is to ensure that the data will not be obtained illegally, and enterprises can achieve authentication and authorization through the access control mechanism.

The second is that when enterprises or individuals use data, they should record the use of data and keep "traces" to prepare for the audit work.

The third is to provide data anonymously or by means of statistical data to ensure that the users of the data will not see the real individual data. If necessary, the data can be protected in the legal system by means of examination and approval and contracts. Violators should be severely punished.

1.1.5 big data service level management

The service level is the common agreement between the users of big data service and the service provider of big data. The service provider of big data needs to provide services according to the agreed service level.

When the big data service provider does not provide services according to the agreed service level, it needs to enhance the service capacity to ensure that the service is provided in accordance with the agreed service level.

For example, it is stipulated in the service level that the time for a user to submit a big data service request to the service response is less than 3 seconds. If the user does not meet such a service level requirement during actual use, big data service provider needs to confirm whether there is a problem with the capacity design of the information system.

If there are problems, the performance requirements can be met by expanding the capacity of big data's service infrastructure. Of course, when big data service provider fails to provide users with the corresponding level of service, big data service provider should give users a certain amount of financial compensation.

In addition to the requirements of system response performance, the service level is mainly whether the data quality provided by big data service can meet the requirements. For example, data integrity and data accuracy should be guaranteed to exceed the percentage agreed in the contract. A data quality verification method that can be accepted by both parties should be established in advance.

1.1.6 big data service availability management

The availability of the service is directly related to the user's experience. If the user experience is good, it will improve the efficiency of users, otherwise, it may lead to the loss of users and reduce enterprise revenue, it can be seen that usability management is very important.

Big data services are divided into three categories: services embedded in the production process, services that provide decision-making reference and services that provide trend prediction. The above three types of big data services have different requirements for availability.

For the big data service embedded in the production process, it is necessary to ensure high availability, otherwise the production and operation of the enterprise will be affected because the decision can not be made in time, for example, the credit evaluation service is integrated into the loan business process of a bank. The credit evaluation service is a big data service. Only after the credit evaluation big data service exports the customer's risk exposure can it be determined whether it can provide loans and loan lines to customers. If the credit evaluation big data service is not available, it will prolong the time for users to obtain loans, thus reducing the efficiency of bank loan business and even leading to the loss of customers.

Comparatively speaking, big data service, which provides decision reference, and big data service, which provides trend prediction, have relatively low requirements for real-time response, so the availability requirements for big data services are relatively low. Of course, their requirements for availability also need to be judged according to the specific situation. If the enterprise deals with emergencies and emergencies, the availability requirements for the above two types of big data services are also very high, and if big data services are not available, it will bring great losses to the enterprise, because the unavailability of big data services affects the decision-making efficiency of the organization and misses a good opportunity to adjust business strategy.

It can be seen that the availability of big data services is very important for the production and operation of enterprises, and it is necessary to ensure the high availability of big data services through the methods and means of availability management.

There are two types of methods to achieve high availability of big data services: passive and active.

The passive method requires the system to monitor the operation of big data service in real time, measure and analyze according to the monitoring results, and show the analysis results in the form of report forms. according to the analysis results to locate and solve the failure points that affect the availability of big data services.

The active method is to collect user usage and system operation data for active analysis, predict problems that may affect big data's service availability, optimize and improve in advance, and prevent problems before they occur.

1.1.7 big data service continuity management

As the name implies, service continuity management is to ensure uninterrupted service. For operational transactional applications, service continuity is an important consideration of service quality. When a service failure occurs, the problem should be found and solved as soon as possible. The length of service recovery time reflects the level of service continuity management.

For big data service, the service continuity of big data service which is closely integrated with the production process is the type that needs to be considered. For other types of big data services, we should focus on ensuring the continuity of data collection services, because if data collection fails, it means that big data services rely on fewer data samples, which in turn affects the results of data analysis.

A brief introduction to the author

Li Fudong, senior big data and information expert, trainer, author of big data Operation, is now a CEO of a high-tech company in Beijing. He has 20 years of cross-industry work experience in telecommunications, finance, Internet, etc., and has long been committed to innovation and practice in enterprise architecture, big data, artificial intelligence, blockchain, virtual reality, digital transformation and so on.

For more up-to-date practical information, please scan the code to join the author's knowledge Planet (Mathematical Firefly), and you will at least get:

1. Answer the question. Answer questions related to digital technology within 72 hours.

2. high-quality articles. Share at least one original and edited article every week.

3. Essence PPT. The latest training, open courses PPT practical information sharing.

4. exquisite gifts. Small gifts, red envelopes and author's signature books are given.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.