52
© 2020 SPLUNK INC. PLA1703C – Splunk as IoT Analytics Platform Leading technology group establishes Splunk as data platform for manufacturing process improvements Dr. Martin Döring Data Scientist | Heraeus Ronald Perzul Staff Consulting Sales Engineer | Splunk

PLA1703C – Splunk as IoT Analytics Platform

  • Upload
    others

  • View
    13

  • Download
    0

Embed Size (px)

Citation preview

Page 1: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

© 2020 SPLUNK INC.

PLA1703C –Splunk as IoT Analytics PlatformLeading technology group establishes Splunk as data platform for manufacturing process improvements

Dr. Martin DöringData Scientist | Heraeus

Ronald PerzulStaff Consulting Sales Engineer |

Splunk

Page 2: PLA1703C – Splunk as IoT Analytics Platform

During the course of this presentation, we may make forward‐looking statements regarding future events or plans 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 may differ materially. The forward-looking statements made in the this presentation are being made as of the time and date of its live presentation. If reviewed after its live presentation, it may not contain current or accurate information. We do not assume any obligation to update any forward‐looking statements made herein.

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

Splunk, Splunk>, Data-to-Everything, D2E and Turn Data Into Doing 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. © 2020 Splunk Inc. All rights reserved

Forward-LookingStatements

Page 3: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Data Scientist | Heraeus

Dr. Martin Döring

Page 4: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Staff Consulting Sales Engineer | Splunk

Ronald Perzul

Page 5: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Agenda 1) About Heraeus

2) Digital Transformation

3) Benefits and Selected Use Cases

4) Architecture

5) Data Onboarding

6) Summary and Next Steps

7) Technical Appendix

Page 6: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

About Heraeus

Page 7: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Facts & Figures

Page 8: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Portfolio

Page 9: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Fields of Activity

Page 10: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Digital Transformation

Page 11: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Starting Point:Digital Transformation Initiatives

End of 2016: first data analysis initiatives“We have a lot of data, where can we put them?”“We have a system for that, let us have a look!”

Mid of 2017: first large use case“Let’s analyze a whole year’s data of our production line!”

Start of 2018Founding of the “Heraeus Digital Factory”Department specialized on digital transformation

Start of 2019“We need an enterprise-grade IoT platform!”Vendor selection based on defined assessment criteria

4

3

2

1

Page 12: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Today: Impact-Driven Approach

Identify use casesApply Six Sigma

methodologyProvide consulting

services

Evaluate impactBusiness unit rates

impact of use cases

Get commitmentExpected impact is

underwritten by relevant stakeholders• Business unit

management• Business unit controlling• Corporate controlling

Conduct projectStandardized project

runbook• Connectivity• Data onboarding• Data visualization• User training

Page 13: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

The Splunk Journey

Aug 2016Introduced

Splunk for Security and IT Operations

Aug 2017First license

extension

May 2019Additional

Splunk instance for use in SOCLicense Volume:

Splunk Enterprise 300 GBSplunk ES250 GB

Oct 2019Decision Splunk

for IoT

Oct 202020 IoT Use

Cases deployedLicense Volume

IoT: 50 GB

Page 14: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Benefits and Selected Use Cases

Page 15: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Impact and Benefits

Number of use cases: 4Savings: ~5M

Number of use cases: 12Savings: ~2M

Number of use cases: 4Savings: ~0.5M

Process optimization

Machine status & capacity

utilizationDefect

statistics

Of the 20 use cases implemented so far

Page 16: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Process Data OverviewAudience

• Process engineers• Quality engineers• R&D

Benefits • Analyze machine sensor data as

basis for production process optimization

• Establish real-time alerting in case of anomalies

Page 17: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Machine StatusAudience

• Shop floor• Lead workers

Benefits • Achieve fine-granular

transparency about machine status

• Use machine status data for OEE visualization

• Decrease downtimes and increase yield and throughput

Page 18: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Capacity UtilizationAudience

• Shop floor• Production planning

Benefits • Achieve real-time transparency

about and gain insights into capacity utilization using machine sensor data

• Increase yield and throughput by optimizing capacity utilization

Page 19: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Defect StatisticsAudience

• Process engineers• Quality engineers

Benefits • Gain insights into product quality

using data from quality systems• Correlate with machine sensor

data to optimize production processes

Page 20: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Architecture

Page 21: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

MQTT:• Lightweight, publish-subscribe protocol to

transport messages between devices

IoT Platform and Protocols Primer

IoT platform:• Multi-layer technology that enables straightforward

provisioning, management, and automation of devices and industrial assets

IoT analytics platform:• Sub-component of an IoT platform with the following

capabilities: reporting, visualization, analytics, alerting, rules, machine learning, and augmented reality

OPC Unified Architecture (OPC-UA):• Machine-to-machine communication protocol for

industrial automation

MQTT Broker

Sensor

Server

Mobile Device

publish 15W

Publish to topic ”Power Consumption”

Subscribe to topic ”Power Consumption”

Page 22: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

End-to-End Architecture (High-Level)From PLC to analytical layer

• Real-time availability of data• Strong dashboarding capabilities• Ingestion of massive amounts of data• Interface to data science platform

Machine

Gateway

MessageBroker

IoT Platform

SAP PCo

• HiveMQ distributes machine data in the Heraeus network

• Ensures that all data points reach the IoT platform

• Dell gateways are available globally in all Heraeus locations

• Using SAP PCo ensures compatibility with SAP MES in the future

Page 23: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

Indexers SHHF

End-to-End Architecture (Detailed)

Page 24: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

Indexers SH

PLC

HF

End-to-End Architecture (Detailed)

Page 25: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

Indexers SH

Gateway

PLC

HF

OPC-UA

End-to-End Architecture (Detailed)

Page 26: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

Indexers SH

Gateway

PLC

HF

OPC-UA

End-to-End Architecture (Detailed)

Page 27: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

Gateway

PLC

HF

MQTT

OPC-UA

End-to-End Architecture (Detailed)

Page 28: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

Gateway

PLC

HF

MQTT

OPC-UA

HF w/ MQTT

HF w/ MQTT

End-to-End Architecture (Detailed)

Page 29: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

Gateway

PLC

HF

MQTT

OPC-UA

HF w/ MQTT

HF w/ MQTT

End-to-End Architecture (Detailed)

Page 30: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

Gateway

PLC SAP ERP

HF

MQTT

OPC-UA

HF w/ MQTT

HF w/ MQTT

End-to-End Architecture (Detailed)

Page 31: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

DataikuGateway

PLC SAP ERP

HF

MQTT

OPC-UA

HF w/ MQTT

HF w/ MQTT

End-to-End Architecture (Detailed)

Page 32: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

DataikuGateway

PLC SAP HANA BWSAP ERP

HF

MQTT

OPC-UA

HF w/ MQTT

HF w/ MQTT

End-to-End Architecture (Detailed)

Page 33: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

DataikuGateway

PLC SAP HANA BWSAP ERP

HF

MQTT

OPC-UA

REST

HF w/ MQTT

HF w/ MQTT

End-to-End Architecture (Detailed)

Page 34: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

DataikuGateway

PLC SAP HANA BWSAP ERP

HF

MQTT

OPC-UA

REST

HF w/ MQTT

HF w/ MQTT

End-to-End Architecture (Detailed)

Page 35: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

DataikuGateway

PLC SAP HANA BWSAP ERP

LB

HF

MQTT

OPC-UA

REST

HEC

HF w/ MQTT

HF w/ MQTT

End-to-End Architecture (Detailed)

Page 36: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

DataikuGateway

PLC SAP HANA BWSAP ERP

LB

HF

MQTT

OPC-UA

REST

HEC

HF w/ MQTT

HF w/ MQTT

File system

End-to-End Architecture (Detailed)

Page 37: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

DataikuGateway

PLC SAP HANA BWSAP ERP

LB

HF

MQTT

OPC-UA

REST

HEC

HF w/ MQTT

HF w/ MQTT

File system

UF

End-to-End Architecture (Detailed)

Page 38: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

DataikuGateway

PLC SAP HANA BWSAP ERP

LB

HF

MQTT

OPC-UA

REST

HEC

HF w/ MQTT

HF w/ MQTT

File system

UF

End-to-End Architecture (Detailed)

Page 39: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

HF

HiveMQ

HiveMQ

LB Indexers SH

DataikuGateway

PLC SAP HANA BWSAP ERP

LB

HF

MQTT

OPC-UA

REST

HEC

HF w/ MQTT

HF w/ MQTT

File system

UF

Custom code(if required)

End-to-End Architecture (Detailed)

Page 40: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Data Onboarding

Page 41: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Data Onboarding

Some design guidelines:• Decide whether to use events or metrics to

ingest MQTT payloads (data interval, data types)

• Always use index time field extractions.• Establish proper transformations rules.• If possible, standardize on message payload to

reduce amount of transformations needed.• Use log2metrics to convert MQTT messages to

Splunk metrics.– See Appendix for detailed props.conf and transforms.conf

example

Events/Metrics

Yes

No

• Index time field extractions• Convert boolean

(true/false) to 1/0• Create additional

descriptive fields (batch ID, asset)

• Index time field extractions• Convert boolean (true/false)

to 1/0• Convert machine state (e.g.

running) to numerical values• Create dimensions for

descriptive fields (batch ID, asset)

Data interval <= 10 sec

Metrics index Event index

Page 42: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Data Onboarding

SAP ERP data• Order information, batch information, quality inspections

Data sets are frequently updated• Only the newest data is required for analysis and visualization• Events/metrics are not suited well for that purpose• Use KV stores instead

KV Stores

REST API Saved searches

SAP ERP

SAP HANA BW

Dataiku Splunk DB Connect

KV store

JDBC JDBC

Page 43: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Summary and Next Steps

Page 44: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Why Splunk?

Evaluation based on defined assessment criteria.Based on this, Splunk was

selected.

Criteria Rating CommentData onboarding via MQTT (and OPC-UA) ++

Time-series visualization of key process parameters

++

Visualization of OEE +

Self-service dashboarding capabilities 0 Requires user training

IoT asset management -- No management

Integration with existing machine learning platform

+ Possible, but custom dev

Scalability ++

Pricing +

Rating of Key Requirements

Page 45: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

1. Conduct advanced trainings for business users

2. Onboarding of additional machines (approx. 500 until end of 2021)

3. Establish internal Splunk community with forum functionality

4. Investigate live dashboards with real-time refresh

Next Steps

Page 46: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Outlook

Additional self-service capabilities

Analytics Workspace

Extend SOC for OT security

OT Security

Provide access to data when

information is needed

Augmented Reality

Leverage new Splunk

Dashboard App

Dashboards

Direct integration with HiveMQ (partner SVA)

HiveMQ

Additional uses of Splunk for IoT

Page 47: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

1. Splunk expert can help avoid complex proof events

2. Leveraging existing Splunk environment and know-how reduced implementation time

3. Using Splunk can help avoid more complex IoT platform architectures

4. Splunk can compete with and beat legacy IoT platform vendors regarding costs, benefits, and capabilities

Key Takeaways

Page 48: PLA1703C – Splunk as IoT Analytics Platform

SESSION SURVEYPlease provide feedback via the

© 2020 SPLUNK INC.

Thank You

Page 49: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Appendix

Page 50: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Convert MQTT Messages to MetricsUse log2metrics feature to convert MQTT messages to metric data points

High-level transformations to create metric data points:1. Identify timestamp prefix, format, and time zone2. Strip quote characters, replace “true” / “false” by 1 / 03. Extract key-value pairs (regex with repeated matching)4. Identify metrics dimensions (e.g. “oven_no”)

See next slide for detailed props.conf and transforms.conf example

Page 51: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Input Transformation ExampleTransform events (as received from MQTT add-on) to metrics

Example payloadTue Jul 21 16:30:33 CEST 2020 name=mqtt_msg_received event_id= topic=<production_area>/<mqtt_client>/<production_process>/machinedata msg={ "oven_no": "oven14", "timestamp": "2020-07-21T14:30:28.153", "performance": 34.41088, "volume_flow": 24.89, "temperature": 2033, "running": true }

props.confTIME_PREFIX = timestamp": "TIME_FORMAT = %FT%H:%M:%S.%3QTZ = UTCSEDCMD-modifypayload = s/"//g s/true/1/g s/false/0/gTRANSFORMS-extractpayload = extract_dimensions,extract_fieldsMETRIC-SCHEMA-TRANSFORMS=metric-schema:metrics_for_iot_data

transforms.conf[extract_dimensions] [extract_fields]REGEX = topic=(.*/(.*)/machinedata) msg= REGEX = \s([\S]+):\s([^,]*)[,\s]FORMAT = topic::$1 production_process=$2 FORMAT = $1::$2WRITE_META = true REPEAT_MATCH = true

WRITE_META = true[metric-schema:metrics_for_iot_data]METRIC-SCHEMA-MEASURES = _ALLNUMS_METRIC-SCHEMA-WHITELIST-DIMS = topic,production_process,oven_no

Page 52: PLA1703C – Splunk as IoT Analytics Platform

© 2020 SPLUNK INC.

Top Search CommandsAs used in all IoT dashboards so fareval 4919 54.95%

stats 783 8.75%

table 561 6.27%

timechart 393 4.39%

sort 257 2.87%

streamstats 197 2.20%

search 193 2.16%

where 178 1.99%

fields 157 1.75%

delta 141 1.58%

mstats 124 1.39%

lookup 118 1.32%

appendcols 111 1.24%

autoregress 101 1.13%

rename 94 1.05%

eventstats 79 0.88%

chart 63 0.70%

fillnull 62 0.69%

transaction 58 0.65%

addinfo 45 0.50%

addtotals 43 0.48%

append 40 0.45%

bin 34 0.38%

dedup 27 0.30%

rex 23 0.26%

gauge 22 0.25%

transpose 22 0.25%

filldown 20 0.22%

fieldformat 17 0.19%

reverse 16 0.18%

makemv 10 0.11%

replace 9 0.10%

foreach 6 0.07%

join 5 0.06%

union 5 0.06%

mvcombine 4 0.04%

mvexpand 4 0.04%

nomv 4 0.04%

rangemap 2 0.02%

top 2 0.02%

regex 1 0.01%

untable 1 0.01%

xyseries 1 0.01%

totaling 93%