Tom Arbuckle

Measure software - and its evolution - using information content

To be able to examine software evolution - variation in software over a sequence of releases - or to compare differing versions of software with each other, we need to be able to measure artefacts representative of the software or its creation process. One can find in the literature a multitude of approaches to both measuring software - by defining and applying software metrics - and to examining software evolution in terms of these metrics. In this position paper, we claim that information content, specifically the (relative) Kolmogorov complexity, is the correct and fundamental tool for the measurement of software artefacts. Experimental results obtained from an analysis of the project udev demonstrate utility: future work should explore the breadth of applicability and determine the full scope of the approach.

Visually Summarising Software Change

Many authors have noted the problem of excessive information when attempting to create useful visualisations of software. The problem of visualising change over multiple versions of software is more complex still. We present a means of visualising changes in software, founded on information-theoretic arguments, that easily and automatically summarises difference between software versions with respect to their code, their structure or their behaviour. Further, we show, by creating visualisations in experiments on real-world data, that the method is of utility to practitioners and has implications beyond the field of software visualisation.

Learning predictors for flash memory endurance: a comparative study of alternative classification methods

Flash memorys ability to be programmed multiple times is called its endurance. Beyond being able to give more accurate chip specifications, more precise knowledge of endurance would permit manufacturers to use flash chips more effectively. Rather than physical testing to determine chip endurance, which is impractical because it takes days and destroys an area of the chip under test, this research seeks to predict whether chips will meet chosen endurance criteria. Timing data relating to erasure and programming operations is gathered as the basis for modelling. The purpose of this paper is to determine which methods can be used on this data to accurately and efficiently predict endurance. Traditional statistical classification methods, support vector machines and genetic programming are compared. Cross-validating on common datasets, the classification methods are evaluated for applicability, accuracy and efficiency and their respective advantages and disadvantages are quantified.

Optimising Flash non-volatile memory using machine learning: a project overview

While near ubiquitous, the physical principles of Flash memory mean that its performance degrades with use. During fabrication and operation, its ability to be repeatedly programmed/erased (endurance) needs to be balanced with its ability to store information over months/years (retention). This project overview describes how our modelling of data we obtain experimentally from Flash chips uniquely allows us to optimise the settings of their internal configuration registers, thereby mitigating these problems.

Estimating MLC NAND flash endurance: a genetic programming based symbolic regression application

NAND Flash memory is a multi-billion dollar industry which is projected to continue to show significant growth until at least 2017. Devices such as smart-phones, tablets and Solid State Drives use NAND Flash since it has numerous advantages over Hard Disk Drives including better performance, lower power consumption, and lower weight. However, storage locations within Flash devices have a limited working lifetime, as they slowly degrade through use, eventually becoming unreliable and failing. The number of times a location can be programmed is termed its endurance, and can vary significantly, even between locations within the same device. There is currently no technique available to predict endurance, resulting in manufacturers placing extremely conservative specifications on their Flash devices. We perform symbolic regression using Genetic Programming to estimate the endurance of storage locations, based only on the duration of program and erase operations recorded from them. We show that the quality of estimations for a device can be refined and improved as the device continues to be used, and investigate a number of different approaches to deal with the significant variations in the endurance of storage locations. Results show this techniques huge potential for real-world application.

Optimising Flash Memory for Differing Usage Scenarios: Goals and Approach

Non-volatile memories, particularly Flash memory, are becoming increasingly important commercially. In contrast with hard disk drives, they possess valuable advantages such as quieter operation, lower access latency, lower power consumption and the production of less heat. At the same time, however, the electronic device on which Flash memory is based, the floating gate transistor, has a limited operating lifetime. As a result, the ability of Flash memory chips to retain information when powered off needs to be balanced against their ability to be repeatedly programmed. This provides an opportunity to optimise the chips’ control parameters to adjust the chips’ operation to fit their desired operating scenarios, thereby resulting in a saving for both manufacturers and users.

Evolving a storage block endurance classifier for Flash memory: A trial implementation

Solid State Drives (SSDs) have a number of significant advantages over traditional Hard Disk Drives (HDDs) but are currently far more expensive and have smaller capacities. These drives are based on NAND Flash memory devices, which have limited working lives. The number of times locations in such devices can be successfully programmed before they become unreliable is termed their endurance. There is currently no way to estimate accurately when a location within a Flash device will fail, so manufacturers give extremely conservative guarantees about the number of program operations their chips can endure. This paper describes a trial implementation of Genetic Programming (GP) used to evolve a Binary Classifier that predicts whether storage blocks within Flash memory devices will still be functioning correctly beyond some predefined number of cycles. The classifier is supplied with only the measured program and erase times from a relatively early point in the lifetime of a block. Using the relationships between these times, the system can accurately predict whether the block will continue to function satisfactorily up to a required number of cycles. Experiments on test sets comprised of unseen data show that our classifier obtains up to an average of 95% accuracy across 30 runs.

Interpreting Shared Information Content in Software Engineering: What Does It Mean to Say Two Artefacts Are Similar and Why?

Recent work has examined software evolution in terms of shared information content between artefacts representative of releases. These studies employ measurements based on the non-trivial concept of Kolmogorov complexity to calculate the quantity of shared information and have shown some of the promise of this characterisation of software evolution.

Implementation of laser Doppler vibrometer employing holographic optical elements

We describe an implementation of a laser Doppler vibrometer system. Two main subsystems work in tandem to produce a novel and effective sensing combination. In the first subsystem, responsible for signal generation, many of the complex optical components commonly employed in more traditional vibrometry systems are replaced with holographic optical elements enabling an impressive reduction in the complexity of the system configuration and deployment. A visible light laser source is employed with obvious safety advantages over non-visible sources. Laser intensity fluctuation signals created in the sensing optics are then captured using a photodiode. In the second subsystem, responsible for the signal processing, the signal is low-pass filtered and the d.c. component of the signal is removed before it is digitised. The data is passed to a host computer where an implementation of synthetic-heterodyne demodulation is employed to detect vibration signals of at least 2kHz. The output from this signal processing provides a measurement of the magnitude and frequency of the vibration. A simple graphical user interface controls the systems operation and displays the vibration results.