Machine data analysis is an urgent business need
Half of Fortune 500 companies experience more than 80 hours of system downtime each year. If evenly distributed throughout the year, there are approximately 13 minutes of downtime per day. Although downtime cannot happen every day, it is possible to have 1.5 hours of downtime in one week or 6 hours of downtime after one months.
As a user, the frequent and unreliable operation of online banking is very disturbing. As business owners, all processes stagnate when the system is down. Ongoing work is interrupted and cannot meet SLAs, which can lead to expensive costs, negative public image, and loss of current and potential future customers. Ultimately, failure to provide a reliable and stable system can lead to financial losses. While the failure of these systems is unavoidable, the ability to predict failures in a timely manner and intercept them prior to these failures is now essential.
A possible solution to this problem can be found in the massive diagnostic data that is generated on the hardware, firmware, middleware, applications, and storage and management layers that indicate failure or error. Machine analysis and understanding of this data is becoming an important part of debugging, performance analysis, root cause analysis and business analysis.
In addition to preventing downtime, machine data analysis provides insight into fraud detection, customer retention, and other important use cases.
The problem of machine data analysis is big data problem
Machine data is undoubtedly consistent with the characteristics of large data.
Each layer of the application architecture generates a wide variety of data outputs. A large amount of such data makes it difficult to use tools manually. Using pattern-matching tools such as grep and awk to filter and correlate data across files requires a lot of labor and time. As the world becomes more material-linked, the volume of data is expected to grow.
Figure 1. Type, quantity and speed of machine data
IBM Accelerator for Machine Data Analytics
The IBM accelerator is a software component that accelerates the development of specific solutions or use cases and implementations on large data platforms.
The IBM Accelerator for Machine data Analytics is a set of biginsights applications that accelerate the implementation of machine data use cases. These applications use the Biginsights runtime technology to support their implementation.
Figure 2. Biginsights and Accelerators
See more highlights of this column: http://www.bianceng.cnhttp://www.bianceng.cn/Programming/extra/