Exciting, To-Be-Announced Platform Session...or other commitment. Splunk undertakes no obligation either to develop the features or functionality described or to include any such feature

© 2018 SPLUNK INC.

© 2018 SPLUNK INC. CONFIDENTIAL INFORMATION. DO NOT DISTRIBUTE.

© 2018 SPLUNK INC.

October 2018

Philipp Drieger | Staff Machine Learning Architect

Exciting, To-Be-Announced

Platform Session We can't tell you about it now, but trust us - it's awesome.

© 2018 SPLUNK INC.

PHILIPP DRIEGER

Staff Machine Learning Architect

[email protected]

mailto:[email protected]

© 2018 SPLUNK INC.


During the course of this presentation, we may make forward-looking statements regarding future events or

the expected performance of the company. We caution you that such statements reflect our current

expectations and estimates based on factors currently known to us and that actual events or results could

differ materially. For important factors that may cause actual results to differ from those contained in our

forward-looking statements, please review our filings with the SEC.

The forward-looking statements made in this presentation are being made as of the time and date of its live

presentation. If reviewed after its live presentation, this presentation may not contain current or accurate

information. We do not assume any obligation to update any forward-looking statements we may make. In

addition, any information about our roadmap outlines our general product direction and is subject to change

at any time without notice. It is for informational purposes only and shall not be incorporated into any contract

or other commitment. Splunk undertakes no obligation either to develop the features or functionality

described or to include any such feature or functionality in a future release.

Splunk, Splunk>, Listen to Your Data, The Engine for Machine Data, Splunk Cloud, Splunk Light and SPL are trademarks and registered trademarks of Splunk Inc. in

the United States and other countries. All other brand names, product names, or trademarks belong to their respective owners. © 2018 Splunk Inc. All rights reserved.

Forward-Looking Statements

THIS SLIDE IS REQUIRED FOR ALL 3RD PARTY PRESENTATIONS.

Presentations to Third-Parties (i.e., non-Splunkers)

• When presenting to third parties, this slide must be

included immediately after the title slide.

• Only share confidential information on a “need-to-know”

basis and make sure the audience members are bound

by non-disclosure or confidentiality agreements.

• Before disclosing any customer or other third party

names, logos or use cases, confirm with Marketing that

we have the right consents.

• Don’t bash the competition. If making comparisons

between Splunk and our competitors, stick to the facts.

• Make sure all statements are not overstated and are

supported by facts.

Confidential Information

• If the presentation contains confidential information, a

confidentiality notice must appear on every slide.

Please contact [email protected] for help

implementing this notice: Confidential information. Do

not distribute.

• Examples of confidential information: financial results,

strategic business and marketing plans, product

launches and roadmaps, customer lists and use cases.

NOTICE FROM SPLUNK LEGAL


© 2018 SPLUNK INC.

AI | ML | Deep Learning

© 2018 SPLUNK INC.


Let’s start with AI

Artificial

Intelligence

Machine

Learning

Deep

Learning

► Artificial Intelligence (AI)

The self driving car

► Machine Learning (ML)

Predicting car demand

based on past history

► Example given by BMW

► Deep Learning (DL)

Not in our product

today.

Computer Vision

Interpretation Of

The Road Ahead

Driver Not

Included

MACHINE LEARNING TOOLKIT

…but we still do

NO image processing

© 2018 SPLUNK INC.


Popular ML/DL Frameworks in the Python Landscape

https://www.kdnuggets.com/2017/03/getting-started-deep-learning.html https://www.kdnuggets.com/2018/02/top-20-python-ai-

machine-learning-open-source-projects.html


https://www.kdnuggets.com/2017/03/getting-started-deep-learning.html








https://www.kdnuggets.com/2018/02/top-20-python-ai-machine-learning-open-source-projects.html


















© 2018 SPLUNK INC.

MLTK Container for TensorFlow™

© 2018 SPLUNK INC.


Popular deep learning frameworks help to

extend MLTK for specific use cases.

Freedom for Data Scientists and

Developers to bring in custom code and models

Flexibility to run compute intense model trainings

on GPU accelerated hardware

MLTK Container for TensorFlow™ – Why? Because our customers ask. We listen. Simple as that!

© 2018 SPLUNK INC.

Architecture Overview

© 2018 SPLUNK INC.


persisted model

Search

Splunk > MLTK > Dockerized Deep Learning

OT

Industrial Assets

Consumer and

Mobile Devices

Real Time

Visualize

Alert

| fit y from x* into “model”

| apply “model”

…


© 2018 SPLUNK INC.


Process and interaction flow

… | fit

… | apply

| summary

localhost:5000/fit

> check y, x and parameters (received via POST)

> create keras/tf model

> fit model

> evaluate model

> persist model

localhost:5000/apply

> check x and parameters (received via POST)

> check for cached model or load from saved model

> run model on input data

localhost:5000/summary

> return last fitting history and model information

grey area symbolizes data volume and typical runtimes

© 2018 SPLUNK INC.

Splunk App for the MLTK Containers

© 2018 SPLUNK INC.


© 2018 SPLUNK INC.


© 2018 SPLUNK INC.


© 2018 SPLUNK INC.

Tests and Benchmarks

© 2018 SPLUNK INC.


Fit on 100k DGA training dataset

© 2018 SPLUNK INC.


Apply on test dataset 1.7M events (ca. 11K EPS throughput)

© 2018 SPLUNK INC.

CPU vs GPU Benchmark

© 2018 SPLUNK INC.


Benchmarking Model Training

▶ AWS Instance: p3.2xlarge (64GB, 8vCPU, NVIDIA V100 GPU 16GB)

▶ DGA Dataset: 100K events with 100 dimensions + 1 target dimension

▶ Neural Network: 10 layer deep neural network with 886K trainable parameters,

▶ 100 layer deep with 9M trainable parameters

© 2018 SPLUNK INC.


CPU: 986 seconds (00:16:26) GPU: 66 seconds (00:00:66)

CPU vs GPU > 15x speedup on search runtime 100K dataset | 100 dimensions | 10 layer NN

© 2018 SPLUNK INC.


CPU vs GPU > 25x speedup on model fitting 100K dataset | 100 dimensions | 10 layer NN

© 2018 SPLUNK INC.

Wrap up

© 2018 SPLUNK INC.


Recommendation Matrix consider your ML dataset’s dimensional and computational complexity

computational complexity

dim

ensio

na

l co

mple

xity

Machine Learning Toolkit

In general: for most common ML tasks: use MLTK + MLSPL API

extensibility

Case #1: need for

specific algo / framework

Case #2: need for

distributed / gpu compute

extensibility

© 2018 SPLUNK INC.


Easy install and prebuilt examples in the

MLTK Container for TensorFlow™ App

Customize containerized code for specific use

cases using any ML/DL frameworks of choice

Flexibility to run compute intense model trainings

on GPU accelerated hardware

Key Take Aways What are the benefits of MLTK Container for TensorFlow™

© 2018 SPLUNK INC.

MLTK Container for TensorFlow™

“Available as Splunk Professional Services (Whiteglove Program) that you can engage

as of today – let us know!”

© 2018 SPLUNK INC.

Thanks | Q&A

[email protected]


© 2018 SPLUNK INC.

Don't forget to rate this session

in the .conf18 mobile app

Thank You

© 2018 SPLUNK INC.


Thanks to … … so many amazing colleagues supporting this idea and helping in getting it real

▶ Customers and Partners

▶ ML PMs: Manish, Andrew, Harsh, …

▶ ML Eng: Xander, Lin, Chang, …

▶ Other: James, Duncan, …

Documents

Exciting, To-Be-Announced Platform Session...or other commitment. Splunk undertakes no obligation either to develop the features or functionality described or to include any such feature