Building and maintaining complex distributed systems
June 19–20, 2017: Training
June 20–22, 2017: Tutorials & Conference
San Jose, CA

Predictive system performance data analysis (sponsored by Salesforce)

Jasmin Nakic (Salesforce ), Samir Pilipovic (Salesforce)
9:00am–12:30pm Tuesday, June 20, 2017
Sponsored
Location: LL20 C
Level: Beginner
Average rating: *****
(5.00, 2 ratings)

Prerequisite knowledge

  • A working knowledge of Python or a similar scripting language

Materials or downloads needed in advance

  • A laptop with Python scripting language support and the Python scikit-learn machine-learning plugin installed
  • Please check additional information by visiting https://github.com/sfperfdemo/vel2017-ml-wave. If you hit a problem installing the software, please let us know.

What you'll learn

  • Learn how to apply a linear regression predictive model on time series performance data

Description

Jasmin Nakic and Samir Pilipovic examine the application of a linear regression predictive model on time series performance data, discussing and evaluating different models to find the optimal choice for a given dataset. All steps will be supported with Python-based scripts so that you can easily implement similar models on your own data. You’ll also learn how to analyze performance data and business data from other sources to detect anomalies in important performance metrics and generate alerts with dynamic thresholds, taking into account weekly and seasonal spikes in traffic, as you explore the process from data collection to applying predictive models to evaluating and visualizing results.

This tutorial is sponsored by Salesforce and is open to all Velocity Conference pass types.

Photo of Jasmin Nakic

Jasmin Nakic

Salesforce

Jasmin Nakic is lead software engineer on Salesforce’s Frontier Scale performance team, where he helps analyze enterprise customer workloads and simulate large-scale benchmarks. Jasmin focuses on bringing advanced predictive analytics and machine-learning methods to massive amounts of system performance data. Previously, he did database and application development at Teradata, KickFire, KLA-Tencor, and Oracle. Jasmin started his computer science adventure in high school in his small hometown in Bosnia, where he wrote short programs to analyze results from astronomical observations on HP calculators. Later, at the University of Sarajevo, he studied system programming, compilers, advanced architectures, and networks and wrote a thesis on object-oriented approaches to database programming.

Photo of Samir Pilipovic

Samir Pilipovic

Salesforce

Samir Pilipovic is a senior software engineer on the Performance Engineering team at Salesforce. His primary interest is performance and optimization of large enterprise deployments in the cloud.

Leave a Comment or Question

Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?

Join the conversation here (requires login)

Comments

Picture of Jasmin Nakic
Jasmin Nakic | LEAD SOFTWARE ENGINEER
06/22/2017 4:04am PDT

We want to thank all who attended the tutorial. Hope that it provided some value to you and your organization and that you may apply some of methods and processes covered during this session.

Correction: When mentioning the history of linear regression, we incorrectly stated that it was invented by Leibniz. It was first defined and used by Carl Friedrich Gauss in 1798 while he was 21 years old. The first known application was to predict the path of asteroid Ceres in 1801.

Picture of Jasmin Nakic
Jasmin Nakic | LEAD SOFTWARE ENGINEER
06/18/2017 5:00am PDT

For those attending tutorial “Predictive system performance data analysis” on Tuesday at 9AM, please check additional information about the prerequisites visiting https://github.com/sfperfdemo/vel2017-ml-wave. If you hit a problem installing the software, please let us know. See you at the tutorial!