ExplorViz: Visual Runtime Behavior Analysis of Enterprise Application Landscapes

Transcription

1 ExplorViz: Visual Runtime Behavior Analysis of Enterprise Application Landscapes Florian Fittkau, Sascha Roth, and Wilhelm Hasselbring Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

2 Motivation Introduction Enterprise application landscapes are complex systems Enterprise models (EMs) seek to capture relevant information Our observations 1) EMs often get outdated 2) Current approaches lack application-level detail Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

3 Approach Introduction ExplorViz [FWWH13] provides live visualization of enterprise application landscapes Used to ensure consistency between an enterprise model and the actual information systems Design Science as research methodology and its two main steps Build and Evaluate Exemplify our approach via modeling of an application landscape Exemplary drill-down to our EPrints operational installation Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

4 Related Work Related Work Research community around Matthes [Mat08] Only relationships between applications and no details Research community around Leymann [BBKL13] Do not provide details on the actual root cause City metaphor, e.g., Wettel et al. [WL07] Only static analysis Application Performance Management (APM) tools, e.g., AppDynamics or dynatrace No abstractions from server and applications Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

5 Enterprise Application Landscape Meta-Model ExplorViz Approach Figure 1 : Excerpt of our enterprise application landscape meta-model Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

6 Landscape Perspective ExplorViz Approach Figure 2 : Modeling the Kiel Data Management Infrastructure for ocean science Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

7 Application Perspective ExplorViz Approach (a) EPrints with closed Screen Figure 3 : Application perspective visualizing the Perl-based EPrints Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

8 Application Perspective ExplorViz Approach (a) EPrints with closed Screen (b) EPrints with opened Screen Figure 3 : Application perspective visualizing the Perl-based EPrints Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

9 Performance Analysis of EPrints Applying ExplorViz Figure 4 : Visualizing the highest 1% of the product of average response time and the call count Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

10 Performance Analysis of EPrints (cont d) Applying ExplorViz Figure 5 : Visualizing the communications of the DataObj.User class Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

11 ExplorViz Approach ExplorViz Implementation Figure 6 : Data processing in our ExplorViz approach [FWWH13] Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

12 Monitoring ExplorViz Implementation Application-level monitoring through AspectJ Adapter to read logs from other tools, e.g., Kieker [vhwh12] Monitoring of remote procedure calls Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

13 Elastic Trace Analysis ExplorViz Implementation Live processing Millions of monitoring records per second in enterprise landscapes Cost-efficient through cloud scaling Figure 7 : Multiple levels of analysis workers Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

14 Visualization ExplorViz Implementation Web browser as front end WebGL as rendering technology Generating JavaScript through Google Webkit Tool (GWT) Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

15 Summary and Outlook Conclusions Live trace visualization of enterprise application landscapes Application monitoring to ensure consistency between the EMs and the real systems Open source web-based tool 2 Future Work: Improving the 3D layout algorithm More advanced filters increasing the analysis capability 2 Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

16 Tobias Binz, Uwe Breitenbucher, Oliver Kopp, and Frank Leymann. Automated discovery and maintenance of enterprise topology graphs. In Proc. IEEE 6th International Conference on Service-Oriented Computing and Applications (SOCA), pages IEEE, December Florian Fittkau, Jan Waller, Christian Wulf, and Wilhelm Hasselbring. Live trace visualization for comprehending large software landscapes: The ExplorViz approach. In Proc. 1st IEEE International Working Conference on Software Visualization (VISSOFT). IEEE, September Florian Matthes. Softwarekartographie. Informatik Spektrum, 31(6): , André van Hoorn, Jan Waller, and Wilhelm Hasselbring. Kieker: A framework for application performance monitoring and dynamic software analysis. In Proc. 3rd ACM/SPEC International Conference on Performance Engineering (ICPE), pages ACM, April R. Wettel and M. Lanza. Program comprehension through software habitability. Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14

17 In Proceedings of the 15th IEEE International Conference on Program Comprehension (ICPC 2007), pages , June Fittkau, Roth, Hasselbring ExplorViz: Visual Runtime Behavior Analysis / 14