SLAML 2010 : Workshop on Managing Systems via Log Analysis and Machine Learning Techniques
Call For Papers
Modern large-scale systems are challenging to manage. Fortunately, as these systems generate massive amounts of performance and diagnostic data, there is an opportunity to make system administration and development simpler via automated techniques to extract actionable information from the data. This workshop addresses this problem in two thrusts: (i) the analysis of raw system data logs and (ii) the application of machine learning to systems problems. We expect the large overlap in these topics to promote a rich interchange of ideas between the areas.
Log Analysis: It is well known that raw system logs are an abundant source of information for the analysis and diagnosis of system problems and prediction of future system events. However, a lack of organization and semantic consistency between system data from various software and hardware vendors means that most of this information content is wasted. Current approaches to extracting information from the raw system data capture only a fraction of the information available and do not scale to the large systems common in business and supercomputing environments. It is thus a significant research challenge to determine how to better process and combine information from these data sources.
Machine Learning: The large scale of available data requires automated and machine-assisted analysis. Statistical machine learning techniques have recently shown great promise in meeting the challenges of scale and complexity in datacenter-scale and Internet-scale computing systems. However, applying these techniques to real systems scenarios requires careful analysis and engineering of the techniques to fit them to specific scenarios; there is sometimes also the opportunity to develop new algorithms specific to systems scenarios. This workshop thrust thus also presents a substantial research area: the exploration of new approaches to using machine learning to help us understand, measure, and diagnose complex systems.
This will be a 1.5-day workshop, beginning on the afternoon of October 2, 2010, and running through October 3, 2010.
Topics include but are not limited to:
* Reports on publicly available sources of sample system logs
* Prediction of malfunction or misuse based on system data
* Statistical analysis of system logs
* Applications of Natural-Language Processing (NLP) to system data
* Techniques for system log analysis, comparison, standardization, compression, anonymization, and visualization
* Applications of log analysis to system administration problems
* Use of machine learning techniques to address reliability, performance, power management, security, fault diagnosis, scheduling, or manageability issues
* Challenges of scale in applying machine learning to large systems
* Integration of machine learning into real-world systems and processes
* Evaluating the quality of learned models, including assessing the confidence/reliability of models and comparisons between different methods
Interested speakers should submit their full papers by June 13, 2010, via the Web submission form, which will be available here soon. All papers will be subject to peer review under conference standards. Experience reports and papers on work in progress are welcome as long as there is a clear contribution. Submissions must be in PDF format and must be no longer than eight 8.5" x 11" pages, including figures and tables, but not including references, formatted in two columns, using 10 point type on 12 point (single-spaced) leading, with the text block being no more than 6.5" wide by 9" deep.
All papers will be available online to registered attendees before the workshop. If your accepted paper should not be published prior to the event, please notify firstname.lastname@example.org. The papers will be available online to everyone beginning on the first day of the workshop, October 2, 2010.
Papers accompanied by nondisclosure agreement forms will not be considered. Accepted submissions will be treated as confidential prior to publication on the USENIX SLAML '10 Web site; rejected submissions will be permanently treated as confidential.
Simultaneous submission of the same work to multiple venues, submission of previously published work, or plagiarism constitutes dishonesty or fraud. USENIX, like other scientific and technical conferences and journals, prohibits these practices and may take action against authors who have committed them. See the USENIX Conference Submissions Policy for details.
Questions? Contact your program co-chairs, email@example.com, or the USENIX office, firstname.lastname@example.org.