From intelligent operation and maintenance to smart operation, Qingchuang Technology gives enterprises the ability to detective Sherlock

[51CTO.com original article] Operation and maintenance work has evolved from early manual operation and maintenance to automated operation and maintenance, and now to intelligent operation and maintenance. Based on the existing operation and maintenance data, enterprises use machine learning to solve problems that automated operation and maintenance cannot solve, and AIOps was born.

With the rapid development of technologies such as cloud computing and big data, under the dual-state IT architecture of "stable and sensitive integration", enterprise operation and maintenance still has a long way to go on the road to intelligentization. In an interview with 51CTO, Yang Chen, founder and CEO of Qingchuang Technology, said that the essence of intelligent operation and maintenance is actually to improve the cognitive ability of operation and maintenance data. Sherlock AIOps developed by Qingchuang Technology will help enterprises improve their cognitive ability of operation and maintenance data from the two levels of intelligent operation and maintenance "technique" and operation "way", thus moving towards the road of intelligent operation and maintenance.

[[342542]]

Yang Chen, founder and CEO of Qingchuang Technology

The vision of intelligent operation and maintenance is beautiful, but the reality is bleak

Intelligent operation and maintenance can help enterprises quickly discover anomalies, effectively diagnose the root causes of problems, conduct business-oriented operational analysis and decision-making, and continuously and effectively improve the quality of operation and maintenance data. Yang Chen said that intelligent operation and maintenance is a new digital operation and maintenance capability, and will also be an essential capability for digital transformation.

Since intelligent operation and maintenance is an essential capability for the digital transformation of enterprises, can it really achieve intelligent operation and maintenance without human involvement? Yang Chen believes that the vision of unmanned operation and maintenance is beautiful, but the reality is very bleak. From the perspective of vision, intelligent operation and maintenance must be self-consistent and independent of people. In a cloud-native environment, the components of the infrastructure and system will be based on standardization, and the operation and maintenance work can be completed entirely based on autonomous, automatic, and intelligent decision-making business systems, thus forming a closed loop. However, today, in the process of transformation, the infrastructure of enterprises is diversified. In addition to traditional architectures and applications, distributed evolution and the introduction of cloud native will also be carried out. The operation and maintenance work during the transformation process is more complex and more challenging.

Yang Chen said that under a diversified architecture, enterprises can improve their operation and maintenance level and data cognition ability through human-machine integration. When some scenarios are not effectively improved, iterative feedback is needed to establish a closed-loop mechanism at the product level, allowing the algorithm to absorb human feedback, thereby turning the personal capabilities of operation and maintenance personnel into organizational capabilities for algorithm and platform processing. This process is gradually evolving and advancing indefinitely, eventually forming a closed loop and reaching a highly self-consistent state.

If enterprises want to achieve truly intelligent operation and maintenance, they still need algorithms, industry experience, and superb engineering standards. Enterprise-level intelligent operation and maintenance products need to facilitate customers to adjust algorithms and models based on usage scenarios to match business needs; in addition, they must have platform capabilities that support efficient operation of algorithms, including big data processing capabilities, machine learning platform capabilities, and streaming data processing capabilities. Yang Chen said,
"There is no god-level algorithm that can solve all problems. Enterprises need to combine industry experience, platform engineering and appropriate algorithms to continuously break through different scenario applications and achieve true intelligent operation and maintenance."

Two major difficulties in implementing intelligent operation and maintenance

AIOps is an algorithm-based IT operation (Algorithmic IT Operations), a new category defined by Gartner in 2016 and adjusted to Artificial Intelligence for Operations in 2017. It is a platform (technology) that empowers traditional IT operation and maintenance management with big data, artificial intelligence or machine learning technology. Some people think that the AIOps market is full of rumors but little action. In Yang Chen's opinion, it is normal for new technologies to have such voices in their early development. However, after four years of exploration and development, many customers have begun to practice AIOps. However, technological development requires a stable mentality, and both supply and demand sides need to patiently find the right landing point in terms of scenarios to generate greater value.

Yang Chen told 51CTO that the most difficult part of implementing intelligent operation and maintenance is modeling, that is, the combination of algorithms and scenarios. In different scenarios, the algorithm model needs to be adjusted according to environmental requirements to improve accuracy and predictability. Yang Chen pointed out that the difficulty of implementing AIOps is not a breakthrough in algorithms, but the combination of environmental factors and industry characteristics, which are fed back into the algorithm. This requires that algorithms and operation and maintenance experts can work together effectively during the R&D stage.

In addition, the second difficulty in implementing AIOps lies in data governance, as data conditions limit the effectiveness of algorithms and their models. Yang Chen said that when helping companies build intelligent operations, Qingchuang Technology tends to first make an overall plan, sort out the customer's existing operation and maintenance conditions and operation and maintenance data status, and then help customers develop a sustainable intelligent operation and maintenance roadmap, rather than forcefully recommending algorithms or related application scenarios to customers.

There are many players in the AIOps field, which is the core competitiveness of Qingchuang Technology

Founded in 2016, Qingchuang Technology is the first domestic supplier of intelligent operation and maintenance AIOps solutions. The company's core product, Sherlock AIOps, has been implemented in industry benchmark companies such as China UnionPay, Bank of Communications, Xiamen International Bank, Founder Securities, Zhengzhou Commodity Exchange, and China Eastern Airlines Group, covering multiple industries such as banking, insurance, securities, manufacturing, and transportation. In the AIOps track, there are many competitors, including old IT manufacturers and emerging manufacturers. So how did Qingchuang Technology become the leader in the intelligent operation and maintenance track, and what are its core competitiveness?

Yang Chen told 51CTO that compared with its competitors, Qingchuang Technology pays more attention to text-like data, such as alarm data and log data, in processing intelligent operation and maintenance scenarios. Qingchuang Technology will optimize the algorithms for text-like data. Alarms are the eyes of operation and maintenance personnel. Historical alarms reflect historical phenomena. It is difficult for operation and maintenance personnel to find associations and patterns from a group of alarms. Qingchuang Technology will associate alarm data with log data to help operation and maintenance personnel find problems early, improve the efficiency of problem solving, and improve the effectiveness of root cause judgment.

Why does Qingchuang Technology attach importance to text-like data? Yang Chen explained that if an enterprise wants to find the root cause of a problem, it can only use the indicator data of the business and infrastructure to preliminarily determine the possible scope of the root cause. If you want to know the root cause of the problem, you still need to rely on text-like data to understand the actual situation. Therefore, Qingchuang Technology will use text-like data (alarms or logs) and the fluctuations of indicator data for comprehensive investigation, so as to help enterprises quickly find the root cause of the problem.

In addition, Qingchuang Technology has a relatively strong operation and maintenance background. The operation and maintenance management experience of Qingchuang Technology team members is basically around 15 years. They have a strong and in-depth understanding of the industry and scenarios, and can use advisory consulting methods to help companies find the right development path and establish a thinking foundation for long-term development.

Sherlock AIOps gives enterprises the ability to act like detectives

Sherlock Holmes is a talented detective created by British detective novelist Conan Doyle, who is good at solving problems through observation, deductive reasoning and legal knowledge. Operation and maintenance work is like a detective consultant, unraveling a vast amount of clues, gaining insights from complex and vast clues, and improving data cognition capabilities. This is a kind of "detective" ability. Therefore, Qingchuang Technology's product is named Sherlock, giving enterprises "detective" capabilities.

At the 4th Dual-State IT Wuzhen User Conference in 2020, Qingchuang Technology released the new Sherlock AIOps Intelligent Operation Platform. It is understood that the Sherlock AIOps Intelligent Operation Platform can integrate existing diversified monitoring tools, monitor alarm events, performance indicators, logs, capacity and other multi-dimensional data, and cover the entire life cycle of intelligent operation and maintenance management in a three-dimensional manner, including alarm analysis center, indicator analysis center, log analysis center, daily intelligent analysis experts, operation decision center and operation and maintenance digital middle platform.

Sherlock AIOps Intelligent Operation Platform Product Architecture Diagram

Yang Chen emphasized that in the past Sherlock focused on the operation and maintenance side, including anomaly discovery and root cause location, but after the product upgrade, it was positioned as an intelligent operation center, using operation and maintenance management products to help companies enhance their value from the perspectives of technology and business operations, such as conducting health assessments of business systems, and correlation analysis of the ratios between business data and IT infrastructure data and resource data.

In addition, the new version of Sherlock AIOps Intelligent Operation Center achieves decoupling at the three levels of algorithms, models, and scenarios, allowing users to debug algorithms, generate different models, and perform model orchestration according to different scenarios, thereby realizing different scenarios and models for different people, and effectively implementing them in different industries.

Conclusion

Qingchuang Technology's years of practical experience in the field of AIOps has won the trust of customers from different industries, including many leading customers in many industries. According to industry reports, Qingchuang Technology's customer subscription repurchase rate can reach 100%. To this end, Yang Chen explained that, first of all, Qingchuang Technology ensures that the customer's usage scenarios and production operations are integrated. Secondly, customers not only use Qingchuang Technology's products, but also adjust algorithm model parameters through the sharing and accumulation of product implementation experience, thereby achieving enterprise best practices. "Qingchuang Technology has deepened the industry scenarios in the field of AIOps, allowing customers to generate real income. This is why customers continue to subscribe."

The future of O&M is bound to move from O&M to operations. Qingchuang Technology will continue to enhance its ability to recognize O&M data, create the most practical tools in intelligent O&M products, and build the smartest products among practical tools to help enterprises embark on the path of intelligent O&M.

[51CTO original article, please indicate the original author and source as 51CTO.com when reprinting on partner sites]