Applied OSS Reliability Assessment Modeling, AI and Tools

In recent years, many computer system failures have been caused by software faults which were introduced during the software development process.

Many OSS’s have been embedded in the software development paradigm. For example, the software development in several application software, mobile software, and cloud computing use the OSS, because of the cost reduction, quick delivery, and work saving. In particular, the cloud OSS has a unique characteristic such as the provisioning processes, the network-based operating, and the big data. Therefore, it is important for the cloud managers to assess the reliability in terms of the big data factor during the operation phase of cloud computing. At present, the amount of the data registered in the bug tracking system become large by using the cloud data operation.

An OSS has been mainly used under the network service [1]. At present, the cloud computing service is changing to the edge computing [2‐6]. The edge computing is useful for the continuity and responsibility of network service. Generally, it is well known that the fault-detection phenomenon depends on the maintenance effort, because the number of software faults is influenced by the effort expenditures. Our research group has proposed many effort maintenance models in the past [7‐14]. All of these models have used the effort data sets of OSS. On the other hand, the OSS reliability assessment methods of this chapter are used the OSS fault count data. In the past, the software reliability has been generally assessed by using the SRGM’s and the fault-detection count data. Then, the methods of software reliability assessment based on the SRGM’s have been proposed by several researchers [15, 16].

Many open source software (OSS) projects continues to increase around the world. Many OSS are operated under the irregular status of component-collision. In particular, several components of OSS will be made a change according to the progress of OSS development as time passes. Also, OSS include several versions such as bug-fix version, minor version, and major version, etc. Considering the characteristics of OSS projects, the operation performance of OSS development will take an irregular fluctuation during the operation, because several OSS components are updated by the OSS version-upgrade. In particular, the characteristics of OSS changes with the replacement of several components according to the numbers of several version upgrade. Several research works in terms of the reliability of OSS have been proposed by several researchers [1‐3]. In the past, the methods of software reliability assessment based on the SRGM’s have been proposed by several researchers [4, 5].

At present, the cloud computing service has been mainly used as the network service [1]. The network service is changing from cloud computing to edge computing[2‐6]. The edge computing is useful for the continuity and responsibility of network service. Also, OSS is embedded to the software system of edge computing. In the future, the edge computing will someday supersede cloud computing. Figure 5.1 shows the structure of edge computing and cloud one. Then, we define the wireless communication environment as white noises. Thereby, we can simultaneously consider the characteristics of edge and cloud, OSS, ubiquitous communication environment, and IoT.

Many OSS projects are existing around the world. OSS are operated under the irregular status of component-collision. In particular, several components of OSS will be made a change according to the progress of OSS development. Also, OSS include several versions such as bug-fix version, minor version, and major version. Considering the characteristics of OSS projects, the operation performance of OSS development will take an irregular fluctuation during the operation, because several OSS components are updated with the OSS version-upgrade. In particular, the characteristics of OSS changes with the switch in several components according to several version upgrades. Several research papers in terms of the reliability of OSS have been proposed by several researchers [1‐3]. In the past, the methods of software reliability assessment based on the SRGM’s have been proposed by several researchers [4, 5]. Also, our research group has been proposed the method of reliability assessment for the OSS [6].

Our research group has proposed several reliability assessment tools. In particular, we have developed a three-dimensional tool for OSS reliability assessment. It is useful to easily understand the trend of reliability from various points of view by using three-dimensional modeling. We show the cumulative number of detected faults \(M_{*}(t)\) at time t of our three-dimensional model proposed in the past as follows [1‐3]:

Many OSS are useful for many software engineers and general users around the world. However, there is no established standard method of quality/reliability assessment for OSS. The bug tracking system is well known as the useful system for quality improvement of OSS. The bug tracking system is implemented in various open source projects in recent years. Also, various fault data sets are registered on the database of bug tracking system.

The OSS have been developed under the initiative of many corporate organizations. Also, many OSS’s have been maintained by many corporate organizations. Especially, many OSS have been developed by using the bug tracking systems. The specified bug tracking systems have been used by several OSS projects. Also, the fault big data sets recorded on the bug tracking system will be very useful to assess the reliability of OSS, because the cumulative number of detected software faults is only used in order to assess the typical software reliability in the past. On the other hand, we can use various fault big data sets obtained from the bug tracking system in case of OSS system.

We focus on the OpenStack Project [1] which includes several edge components. In this chapter, we show numerical examples by using data sets on the assumption of the edge OSS service. The data used in this chapter are collected from the bug tracking system OpenStack Project [1]. The demonstration of our prototype tool is available from “DEMO APPLICATION” at the following URL; however, the function of calculation cannot execute considering the security: http://​www.​tam.​eee.​yamaguchi-u.​ac.​jp/​, accessed on 23 December 2023. Our prototype tool has been released as the OSS based on GNU General Public License (GPL) in December 2023. The source code of our tool is available from “SOFTWARE” at the following URL: http://​www.​tam.​eee.​yamaguchi-u.​ac.​jp/​, accessed on 23 December 2023.

This chapter shows several exercises for understanding the reliability assessment measures for OSS reliability measurement and assessment. The following problems will be useful for software managers to evaluate OSS quality/reliability. Figure 12.1 shows the estimated expected number of remaining faults. Discuss the estimated expected number of remaining faults at 8,000 days in Fig. 12.1.