Effects of software aging and rejuvenation on performability of layered distributed systems

When a fault-tolerant layered distributed system continues its operation despite the presence of component failures, its performance is usually degraded. Its performance can also be degraded if it is executing continuously for a long period of time due to a phenomenon known as software aging. To prevent unexpected or unplanned outages due to aging, a pro-active technique called software rejuvenation can be employed. This technique involves gracefully terminating an application and immediately restarting it with a refreshed internal state. For proper modeling of these systems, their performance and dependiability characteristics need to be considered in a unified way, called performability. This thesis proposes a new model called "Rejuvenated-FTLQN", to evaluate the effects of software aging and rejuvenation on performability of these layered systems. Specifically a Layered Queueing Network (LQN) is used for performance analysis and a Multi State Fault Tree (MSFT) is used for dependability analysis. The model is also used to study the impact of performing rejuvenation, time to perform rejuvenation and rejuvenation frequencey on performability of a system. A software tool called "Rejuvenated-FTLQNS" has been developed to automate the model solution.

Effects of software aging and rejuvenation on performability of layered distributed systems

When a fault-tolerant layered distributed system continues its operation despite the presence of component failures, its performance is usually degraded. Its performance can also be degraded if it is executing continuously for a long period of time due to a phenomenon known as software aging. To prevent unexpected or unplanned outages due to aging, a pro-active technique called software rejuvenation can be employed. This technique involves gracefully terminating an application and immediately restarting it with a refreshed internal state. For proper modeling of these systems, their performance and dependiability characteristics need to be considered in a unified way, called performability. This thesis proposes a new model called "Rejuvenated-FTLQN", to evaluate the effects of software aging and rejuvenation on performability of these layered systems. Specifically a Layered Queueing Network (LQN) is used for performance analysis and a Multi State Fault Tree (MSFT) is used for dependability analysis. The model is also used to study the impact of performing rejuvenation, time to perform rejuvenation and rejuvenation frequencey on performability of a system. A software tool called "Rejuvenated-FTLQNS" has been developed to automate the model solution.

Software aging prediction and rejuvenation in cloud computing environment: a new approach

<span>Service availability is one of the major requirements for user satisfaction. Several researches were conducted in recent years to find suitable infrastructure to enhance the availability. Even though both hardware and software are to be in good condition, in recent years, software faults are the major concern for service availability. Software aging is a type of software fault. Software aging occurs as a result of errors accumulation in the internal environment of the system leading to performance degradation. To manage software aging, technique used is software rejuvenation. There exist two kinds of approaches for studying software aging and deriving optimal software rejuvenation schedules. The two approaches are measurement based and model based. In model based approach, analytic models are built for capturing system degradation and rejuvenation process. In measurement based approach, attributes are periodically monitored and that may indicate signs of software aging. In this work, a prototype of measurement based model has been developed. The model captures the aging indicator metrics from cloud environment and rejuvenates once the system reaches aged status. The proposed model uses platform independent, non-intrusive technique for capturing metrics. The rejuvenation carried out after analysing the captured metrics, increases the availability of the service.</span>

RESEARCH OF SOFTWARE AGING IN ANDROID SYSTEM PROCESSES AND USER APPLICATIONS

Android operating system is vulnerable to the aging-related effects such as performance degradation and increased of aging-related failures rate due prolonged usage of a mobile device without rebooting. This paper considers software aging phenomenon in system processes and user applications of the Android operating system and means for counteracting this phenomenon. Experimental research was performed using a methodology that consists in performing stress tests on mobile applications, collecting system data on running processes, converting the collected data into time series for the relevant metrics and analyzing these data using statistical methods. Thus, the analysis of oom_adj_score for determining processes priorities in the context of software aging allowed to identify two groups of processes, namely system processes and user applications. It is also pointed out the possibility of using oom_adj_score to determine the state of system usage in the tasks of software aging predicting and performing software rejuvenation. The results of the system processes analysis showed that the indicators of aging are system_server and surfaceflinger processes, as well as com.android.phone, cameraserver in the case of active use of contacts and camera applications. The considered processes can be used to implement software rejuvenation. Research has shown that user applications are also vulnerable to aging-related effects, but the rejuvenation procedure cannot be applied to them at the system level. It is important to take steps to prevent aging-related errors, such as using appropriate data structures and algorithms for efficient memory management, minimizing the load on the main UI stream, and using effective graphics techniques to reduce the number of delayed frames. In future works it is important to investigate the considered system processes and services in tasks of software aging forecasting and performing of rejuvenation procedure. It is important for user applications to develop tools that provide developers with information about the state of software aging in the system, which would allow to decide on the feasibility of performing important and resource-intensive tasks in conditions when the system is already in a state with a high probability of aging-related failure.

Availability Analysis of Software Systems with Rejuvenation and Checkpointing

In software reliability engineering, software-rejuvenation and -checkpointing techniques are widely used for enhancing system reliability and strengthening data protection. In this paper, a stochastic framework composed of a composite stochastic Petri reward net and its resulting non-Markovian availability model is presented to capture the dynamic behavior of an operational software system in which time-based software rejuvenation and checkpointing are both aperiodically conducted. In particular, apart from the software-aging problem that may cause the system to fail, human-error factors (i.e., a system operator’s misoperations) during checkpointing are also considered. To solve the stationary solution of the non-Markovian availability model, which is derived on the basis of the reachability graph of stochastic Petri reward nets and is actually not one of the trivial stochastic models such as the semi-Markov process and the Markov regenerative process, the phase-expansion approach is considered. In numerical experiments, we illustrate steady-state system availability and find optimal software-rejuvenation policies that maximize steady-state system availability. The effects of human-error factors on both steady-state system availability and the optimal software-rejuvenation trigger timing are also evaluated. Numerical results showed that human errors during checkpointing both decreased system availability and brought a significant effect on the optimal rejuvenation-trigger timing, so that it should not be overlooked during system modeling.