Adaptive algorithms for managing a distributed data processing workload
IBM Systems Journal
Discovering models of software processes from event-based data
ACM Transactions on Software Engineering and Methodology (TOSEM)
Mining Process Models from Workflow Logs
EDBT '98 Proceedings of the 6th International Conference on Extending Database Technology: Advances in Database Technology
Pinpoint: Problem Determination in Large, Dynamic Internet Services
DSN '02 Proceedings of the 2002 International Conference on Dependable Systems and Networks
Workflow mining: a survey of issues and approaches
Data & Knowledge Engineering
Mining logs files for data-driven system management
ACM SIGKDD Explorations Newsletter - Natural language processing and text mining
E2EProf: Automated End-to-End Performance Management for Enterprise Systems
DSN '07 Proceedings of the 37th Annual IEEE/IFIP International Conference on Dependable Systems and Networks
Tracking in a spaghetti bowl: monitoring transactions using footprints
SIGMETRICS '08 Proceedings of the 2008 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
Mining activity data for dynamic dependency discovery in e-business systems
IEEE Transactions on Network and Service Management
A universal method for composing business transaction models using logs
IM'09 Proceedings of the 11th IFIP/IEEE international conference on Symposium on Integrated Network Management
Hi-index | 0.00 |
Existing transaction monitoring solutions are either platform-specific or rely on instrumentation techniques, which limit their applicability. Consequently, transaction monitoring in enterprise environments often involves the manual collation of information spread across a variety of infrastructure elements and applications, and is a time-consuming and labor-intensive task. To address this problem, we have developed an online, non-intrusive and platform-agnostic solution for transaction monitoring. The solution includes a transaction model discovery component that leverages historical system log files, containing transaction footprints and generates a model of the transaction in terms of valid sequence of steps that a transaction instance may execute and the expected footprint patterns at each step. The online monitoring system, in turn, takes in only (a) online system log files and (b) the transaction model, as inputs and generates a dynamic execution profile of ongoing transaction instances that allows their status to be tracked at individual and aggregate levels, even when transaction footprints do not necessarily carry correlating identifiers as those injected through instrumentation. In this paper, we describe the transaction model discovery and monitoring system including the architecture and algorithms, followed by results from an empirical study, ongoing work on run-time model validation and directions for future research.