On July 9, 2021, Qiniu Cloud, a leading enterprise cloud service provider in China, held a new product conference in Shanghai to release its intelligent operation and maintenance product PISA.

PISA, Pandora Intelligent Service Analysis, is an Intelligent Service Analysis tool developed based on the machine data Analysis platform Pandora of Qiniu Cloud. Qiniu defines PISA as a brand new app on Pandora, which means a lot to the industry and Qiniu itself.

Pandora is a cloud-native machine data analysis platform launched by Qinu Cloud, which provides the core ability to explore the value of data through the platform and the ability of AI algorithm. PISA is the landing application of this core ability in the field of intelligent operation and maintenance. Through PISA, Pandora’s capabilities are fully leveraged and open to developers, making it easier and more direct for operations developers to flexibly use Pandora’s capabilities based on their own business scenarios.

How has the opening of analytical capabilities been achieved? From the perspective of architecture, PISA is divided into three layers: bottom, middle and upper. The bottom layer is based on the Pandora core computing engine, which outputs powerful, stable, high-performance and low-cost computing power. Mid-level relies on the Pandora open platform to provide developers with rich, open and easy-to-use support functions. The upper layer is the application layer, which relies on the rich AIOPS experience of Qiniu Cloud to implement the application of intelligent operation and maintenance services to help end users flexibly deal with different business and IT operation and maintenance analysis scenarios.

Intelligent operation and maintenance

Since Pandora’s power can be exported through PISA, we have to talk about how the combination of PISA and Pandora can help the operation and maintenance of enterprises in real world scenarios.

In a practical scenario, for example, the customer calls the bank to report the problem of paying water and electricity bills. The traditional way of dealing with IT is that the customer service submits the work order to IT for operation and maintenance investigation, and then the operation and maintenance troubleshoots the problem and feedbacks the result to the customer through the customer service. At this point, however, the breakdown has already occurred and the customer experience is beyond repair. With the help of PISA and Pandora, the entire operation system of banks can be fundamentally transformed.

First, turn post-processing into ex ante discovery. In the same case, the bank’s IT department was alerted to a possible problem with the billing service before the customer called. Then, compared with the traditional layer by layer and block by block operation and maintenance troubleshooting, PISA can quickly find problems and fix faults through the business transaction health service analyzer.

The fundamental change that PISA brings to the actual operation and maintenance business process comes from the different way of thinking of Qiniu Cloud.

1. Get through business and operation and maintenance, and post-processing is transformed into pre-discovery

In the real world, business and operations are often two separate departments. Because of career requirements, business units generally don’t care about how the underlying systems are implemented; they focus more on delivery. At the same time, the two departments generally communicate very little, and only get together to solve problems when there is a breakdown or when the business wants to compete for IT resources.

As a result, operations have always been the cost center of the company, where the money is spent. From a profit perspective, it is natural that the smaller the cost, the better, but it also increases the likelihood of problems. However, if the operation and maintenance can detect the possible problems in advance, reduce the loss for the company and transform the cost center into a profit center, it is a better operation logic.

Qiniuyun believes that business and operation and maintenance to get through, so that operation and maintenance can predict risks in advance, business can also be more stable, more secure.

2, through the operation and maintenance, only data exchange can faster repair faults

Every link of the IT system may have failures, but the existing monitoring, operation and maintenance system can only focus on a certain level, such as the Skywalking monitoring application layer, Zabbix monitoring service layer. Therefore, it is very time-consuming and cumbersome to sort through each one.

Moreover, when the operation and maintenance department of an enterprise is subdivided into multiple responsibility lines such as application operation and maintenance, host system, network and DBA, it is even more difficult to troubleshoot due to the lack of data intercommunication. It’s like the inability of departments to share clues in solving a crime is bound to delay progress.

Therefore, the most basic thinking of Qiniuyun is that the data must be gathered. Founded in 2011, Qiniuyun’s most important asset over the past decade has been building a unified, heterogeneous data lake, the foundation on which Pandora can operate.

Heterogeneous data lakes are having their value

From the overall point of view of the company, the main businesses of Qiniu Cloud include Media PaaS and Data PaaS, etc. All Data will be gathered into the heterogeneous Data lake derived from the set structured, unstructured and semi-structured Data.

Pandora is a machine data analytics platform that analyzes data generated by any machine or system, such as servers, sensors, applications, network devices, etc.

Pandora has three big advantages. First, Pandora allows for complete decoupling of computing and storage, with resources changing as needed.

Second, Pandora’s ability to support Schema on Read natively. Support collection time, index time, search time three parsing methods. Parsing while searching results in an extremely simple data access experience.

Meanwhile, Pandora uses SPL (the standard language for machine data analysis, SPL=SQL+ UNIX Pipeline) for data analysis. The SPL language has the ability to unify the requirements of retrieval, analysis, visualization and alarm, support the direct processing of raw data, and is optimized for time series data.

The three advantages of Pandora perfectly solve the problem of efficient intelligent analysis of complex data, which, combined with PISA, brings the answer to the disruptive innovation of intelligent operation and maintenance. The above mentioned active discovery and data access are actually the result of the combination of the accumulation of heterogeneous data and the analysis ability of heterogeneous data.

In the era of the industrial Internet, new sparks of intelligent operation and maintenance have erupted

On August 24, 2020, when the global online office was affected by the epidemic, ZOOM stopped working for 3 hours. While bringing inconvenience to users, it also makes its growing momentum to be questioned. For Internet and technology companies, the stability of services means the business itself, which is self-evident.

Today, in addition to the Internet itself, IT technology and intelligent technology into the deep water. In industry and manufacturing, operations are more powerful than ever before. Predictive maintenance, fault prediction, remote intelligent operation and maintenance, industrial data analysis… , these relatively mature functions have been applied in factories, power systems, rail systems and other fields.

In these industries, machines and systems are generating data all the time, and “data without algorithms” is the norm. Therefore, how to collect and analyze machine data is the key for Qiniuyun and other companies to win intelligent operation and maintenance. From this perspective, the combination of PISA and Pandora, released today, is a move by the seven giants to stir up the industry’s waves.