Monday, 2017-07-24

Sponsors

 
 

 
 

  

ICPE 2016 Tutorials

The tutorials will take place on March 13, 2016, with two simultaneous tracks of tutorials.

 

Track One Track Two
Incorporating Software Performance Engineering Methods Practice into the Software Development Life Cycle
Andre Bondi
(3 hours + break, 09:00-12:30) 
Building Custom, Efficient and Accurate Memory Monitoring Tools for Java Applications
Verena Bitto, Philipp Lengauer
(3 hours + break, 09:00-12:30) 

Application Performance Management in Virtualized Datacenters
Rean Griffith, Anne Holler, Xiaoyun Zhu
(3 hours + break, 14:00-17:30) 

Automated Parameterization of Performance Models from Measurements
Giuliano Casale, Simon Spinner, Weikun Wang
(3 hours + break, 14:00-17:30) 

 

Incorporating Software Performance Engineering Methods Practice into the Software Development Life Cycle
Author: Andre Bondi (Software Performance and Scalability Consulting LLC)
Tutorial date/time: Sunday, March 13, 2016 (3 hours + break, 09:00-12:30)

Abstract: Too often, attention is only paid to performance concerns
after functional testing, when it usually too late to remedy disabling
performance problems. We describe how early attention to performance
concerns and early planning of performance requirements and performance
testing can prevent debacles like the early rollout of healthcare.gov
while enhancing reliability and scalability, and while addressing other
cross-cutting concerns. We show how performance engineering methods may
be integrated into all phases of the software lifecycle, from the
conception of a system to requirements specification, architecture,
testing, and finally to deployment. By reviewing the architecture of a
system before design and implementation take place, we reduce the risk
of designing and developing a system that contains inherent performance
vice. Modeling can be used to justify architectural and scheduling
decisions such as the use of scheduling rules. The outputs of
performance tests planned with reference to performance models enable
us able to identify concurrent programming issues and other issues that
would not be apparent in unit testing.



Building Custom, Efficient and Accurate Memory Monitoring Tools for Java Applications
Authors: Verena Bitto (Johannes Kepler University, Linz, Austria), Philipp Lengauer (Johannes Kepler University, Linz, Austria)
Tutorial date/time: Sunday, March 13, 2016 (3 hours + break, 09:00-12:30)
Download tutorial slides

Abstract: Traditional monitoring techniques can distort application
behavior significantly. In this tutorial, we will provide an overview
about state-of-the-art monitoring techniques and their impact on memory
behavior. It will show lightweight techniques that can be used to build
a custom but efficient monitoring tool. We will include topics ranging
from capturing interesting events, serializing and processing the data
offline, dealing with large amounts of data, to visualizing it.
AntTracks is a custom memory monitoring tool built into the Hotspot
Java Virtual Machine. It achieves very low run-time overhead (4.68%)
although capturing all allocations as well as object movements. We will
thus use it exemplary throughout the tutorial to demonstrate the
discussed techniques.



Application Performance Management in Virtualized Datacenters
Authors: Rean Griffith (VMware, USA), Anne Holler (Independent Researcher) and Xiaoyun Zhu (Futurewei Technologies, Inc., USA).
Tutorial date/time: Sunday, March 13, 2016 (3 hours + break, 14:00-17:30)
Download tutorial slides

Abstract: Virtualized private or public cloud datacenters provide
flexible access to computing resources, but their use can present
challenges to dynamically meeting application performance goals
efficiently. In this tutorial, we will first examine the kinds of
resource schedulers currently available for the datacenter, along with
their use cases. We will then present techniques for the automatic
scaling of applications horizontally and vertically to maintain their
Service Level Objectives with right-sized encapsulation. And finally,
we will discuss analytics pipelines for the telemetry data of workloads
running in the datacenter.



Automated Parameterization of Performance Models from Measurements
Authors: Giuliano Casale (Imperial College London, UK); Simon Spinner
(University of Würzburg, Germany); Weikun Wang (Imperial College London, UK)
Tutorial date/time: Sunday, March 13, 2016 (3 hours + break, 14:00-17:30)
Download tutorial slides

Abstract: The goal of this tutorial is to present the problem of
estimating parameters of performance models from measurements of real
systems and discuss algorithms that can support researchers and
practitioners in this task. The focus will be on performance models
based on queueing systems, where the estimation of request arrival
rates and service demands is a required input to the model. The
tutorial will review existing estimation methods for service demands,
ranging from regression-based methods to maximum likelihood techniques,
and present models to characterize time-varying arrival processes. The
tutorial will also demonstrate the use of relevant tools that automate
demand estimation, such as LibRede, FG, and M3A.