Feb. 19 to Feb. 21, 2019

Hyatt at Olive 8, Seattle, Washington, USA

IMPACT Speaker Sessions

Keynote: Real User Performance Monitoring at Netflix Scale.

Martin Spier   |   Senior Performance Engineer at Netflix

Netflix runs on thousands of different devices, from smartphones to laptops, smart TVs and gaming consoles. These devices are spread all around the world, have a vastly different set of capabilities, resource constraints, and run under a multitude of network conditions. Understanding how real users perceive Netflix’s app performance is key to optimizing the right things, finding issues, and improving user experience.

AI: Where we came from, where we are, and where we are going.

Bryan Knouse |   Grand Prize Winner, Need for Speed Hackathon

AI: Where we came from, where we are, and where we are going with a particular focus on natural language technology. This talk is for everyone: non-technical individuals will appreciate the conceptual breadth and CTOs will enjoy the dips into present theory powering the space.

Hack your life - Solutions for a High-Performance Existence

Denise Kalm   |    Long-time IMPACT Speaker

Performance isn’t just about IT. How do you do your best work or have a great quality of life if your ‘systems’ (body and mind) aren’t working optimally? In this session, learn a few ways you can improve the quality of your health and your life with small tuning changes that will pay off big.

Is my VMware VM getting the CPU it needs?

Debbie Sheetz   |   Greater Boston Area

A holistic approach is defined, identifying metrics from each layer of the virtualization, i.e. ESX cluster, host, VM, the VM’s operating system (Linux or Windows), and the application in order to determine if the guest needs to be resized or the ESX environment requires changes. CPU is the focus.

Project Enigma - Exploring the AI Continuum

Rick Lebsack | IBM and the CMG Board

AI has moved from domain of univerities to business. It is easy to get lost in the new terms, not to mention the deep math, that let machines assist humans – be it phone unlocking, password protection or detecting criminals. This talk gives you a foundation to understand AI’s entry to general IT.

Is Your Load Generator Launching Web Requests in Bunches?

Jim Brady   |    Reno, Nevada Area      |

A load generator that launches web requests in bunches produces incorrect system performance results and until now load testers had no tools to identify the problem. In this talk a free downloadable tool is used to evaluate request timing quality by comparing a test run stat with real user behavior.

Jim is a Capacity Planner for the State of Nevada.

Capacity Planning Under Cloudy Skies: The Times Are A-Changing

Brian L Wong, Capital One      |     Richmond, Virginia Area                

Much of the world is moving to clouds, and many significant parts of IT are changing at the same time. Capacity planning is far from obsolete, but practicing it in a different and much more rapidly changing environment will be very different than in the past. Here’s where the field is going.

z/OS CPU Measurements Demystified

Scott Chapman, Enterprise Performance Strategies            

z/OS CPU measurements seem to confuse many people, especially when trying to understand how all the different CPU metrics relate to each other. In truth though, it’s all just CPU time and simple math.

Planning For Your Next Mainframe Processor

Scott Chapman, Enterprise Performance Strategies            

When you are considering upgrading your mainframe it’s always best to do your own analysis of the options instead of just relying on your business partner’s recommendation. But since it’s an analysis that’s usually done only every few years, it’s not a process that most sites are well familiar with.

Combining Capacity Management with Machine Learning

Charles Johnson, Syncsort, Inc.      |     Greater Detroit Area              

Capacity Management within the enterprise continues to evolve.  In the past we were focused on the hardware; now focused on services. The amount of data available has increased significantly.  To be successful with this discipline going forward, we need the machines to do more of the heavy lifting.

Continuous Availability: The role of Nonfunctional Testing in Assuring Continuous Operations

Silvio Orsini, Pietro Tiberi, Marco Capotosto, Banca d’Italia   

In this paper, we present a concept on automated NFT running on a H24/7D application and focus on a couple of  cases implemented (Disaster Recovery, Chaos Monkey), with the aim of building a predictive and repeatable model for operating conditions under disturbances.

What you need to know about Startups

Elisabeth Stahl, IBM      |      Cleveland/Akron, Ohio Area   

Interested in the culture at startups? Need to work with or market to a startup? Want to learn more about what makes startups successful? Attend this session for the inside view of startups and best practices and recommendations!

Why is this web app running slowly?

Mark B. Friedman, Demand Technology Software                

The session looks at how the YSlow scalability model influenced the development of other web application performance tooling, culminating in the W3C specification of a navigation and timing API that provides access from JavaScript to web application performance measurements.

Integrating Performance into the Software Development Life Cycle

Mark B. Friedman, Demand Technology Software

This talk discusses Best Practices for integrating performance into each and every phase of software development, from specifying performance requirements to instrumenting applications during developing to track progress against performance goals, to stress testing during the QA, through DevOps.

Web Performance Analysis Quickstart

Paul Calvano, Akamai

You need to optimize your site’s performance, but where do you start? Do you focus on the backend or frontend? Which tools should you use? And what can you really expect out of 30 minutes? During this talk you will learn some techniques that will get your performance audit off to a productive start.

Catching Anomaly and Normality in Cloud by Neural Net and Entropy Calculation

Igor Trubin, Capital One Bank

The Neural Network (NN) is not a new machine learning method. Now, because computing power is increased dramatically, more tasks can be done by using NN. The goal of the presentation is to demystify the technique in some simple terms and examples to show what it actually is and how that could be used for Capacity and Demand management.

Capacity and Risk - The Coffee Cup and the Lake


Risk is one of the elements that are overlooked when discussing performance monitoring or capacity management. Risk is best understood by the concept of vulnerability vs convenience. How will performance monitoring and capacity planners deal with this correlation? Listen to this presentation.

Answering Questions About the Web w/ the HTTP Archive

Paul Calvano, Akamai

Have you ever wondered how your site’s performance and construction compares to the web as a whole? Or how different technologies are used on the web? Paul Calvano will explore how the HTTP Archive works, some ways people are using it, and ways that you can leverage the data for your analysis today.

Compression on the Web

Paul Calvano, Akamai

How you configure compression matters to your customers’ experience and your bottom line. With webpages growing >3MB, gzip is no longer enough. During this session, Paul Calvano will dive deep into best practices and techniques that you can apply to reduce your page weight.

Basic Queueing Laws and Understanding Weird Performance Test Results

Andre Bondi, Software Performance and Scalability Consulting LLC

Strange features of performance test results can usually be diagnosed by examining how and why the performance measures deviate from standard queueing and performance laws. The structure of the test may also be informative. We shall illustrate this with some examples.

A Neural Model to Detect & Diagnose System Anomalies

Girish Gundale & Vamsi Pentakota, Tech Mahindra Ltd

A discussion on an intelligent system developed for automatic detection of the system anomalies followed by a rational root cause analysis during performance testing using artificial neural network algorithms. This model auto-analyzes and co-relates key metric like response times, server utilizations