Upload
others
View
13
Download
0
Embed Size (px)
Citation preview
De data uitdagingvoor service providers
Jeroen Bronkhorst – Innovation LeadOctober 17, 2019
“Alles isData”
DHPA Netwerkbijeenkomst
“Data can transform serviceproviders and their customers”Unlocked through Artificial Intelligence and AdvancedAnalytics
2
DATA IS CREATED AND (RE)USED EVERYWHERE AND INEVERY’THING’!
3
INTELLIGENCE IS KEY FOR EXTRACTING VALUE FROM DATAEXPLOSION
How will you unlock
hidden insights to drive
competitive
advantage, new
revenue streams and
operational efficiency?
Greater than
40ZBsby 2020
4
SERVICE PROVIDER DATA OPPORTUNITIES
1. Optimize service deliverycustomer value – improved service levels
service provider value – better margins & loyalty
2. New (data) servicescustomer value – get better insights from their data
service provider value – new revenue &differentiation
5
CHALLENGES TO UNLOCKING DATA’S FULL POTENTIAL
6
1 Reacting to unexpected problems
Countless hours pouring over logs and files2
3 Vendor support adds to frustration
1 Delays, disruption to applications
Complexity of operations2
3 Human error
Reactive, inefficient IT
Business risk
Results in application delays and lost productivity: The app-data gap
YOU CAN’T AFFORD AN APP-DATA GAP
7
DataApps
Loading
WHAT CAUSES THE APP-DATA GAP?
8
Storage
Network
Virtualization100s ofVariables
Far too complex for humans to solve!
More than
90%of problems
arise from abovethe storage layer*
Reliable & fast storagesolves only half the problem
Server
*IDC, Server and Storage Availability Survey Results, April 2019
Example: HPE Support Services automated by HPE Infosight
OPTIMIZE SERVICE DELIVERY THROUGH ARTIFICIALINTELLIGENCE
9
5 key steps in the machine learning and
predictive analytics process
1. Observing – the 1000s of data points and sensors
built into the HPE server and storage products
2. Learning – applying advanced pattern recognition
to the sensor data collected across all devices
globally
3. Predicting – anticipating problems based on the
observations and learnings
4. Recommending – intelligent decisions that prevent
issues, improve performance, and optimize
resources
5. Acting – automation resulting in game-changing
benefits and outcomes
Global Intelligence enables faster innovation
BUSINESS OUTCOMES?!
101. Analyst White paper by ESG “Assessing the financial impact of HPE InfoSight predictive analytics”, September 20172. Illustrates potential savings based on customer surveys. HPE does not provide financial advice3. HPE Business White paper “Redefining the standard for system availability”, August 2017
AI-Driven Efficiency
By spending up to 85% less
time managing storage1
with over 99.9999%measured uptime2
Always On, Always Fast
with 86% of issues automatically
opened and resolved3
Transformed Support
SERVICE PROVIDER DATA OPPORTUNITIES
1. Optimize service delivery
customer value – improved service levels
service provider value – better margins & loyalty
2. New (data) services
customer value – get better insights from their data
service provider value – new revenue &differentiation
11
DATA SPANS FROM EDGE TO CLOUD
12
Edge Cloud
Where is datagenerated?
How long do youhave to take
action?
What governance & securityregulations do you need to comply
with?
What are yourbusiness goals?
How do you prepare and integratedata for advanced analytics?
What does thatdata consist
of?
FINANCIAL SERVICES DATA PIPELINE EXAMPLE
13
ContactlessPayments
Instant Payments Data Lake Fraud Detection
Bank Systems
Continuous Improvement
DATA PIPELINE MODEL
14
Internet of ThingsEdge processing of data in motion
Data is
• Acquired
• Queued, routed and orchestrated
• Cached and stored locally
• Applied with rules and analyticmodels
Fast DataCore processing of data in motion
Data is
• Ingested
• Restructured, enriched
• Persisted for real-time usage andoffline analytics
• Applied with rules and analyticmodels
Big DataAnalysis of data at rest
Data is
• Hosted in data lakes
• Transformed and restructured
• Aggregated
• Molded by rules, models
• Prepared for Deep Learning
Artificial IntelligenceDeep learning / machine learning
Data
• Trains and builds analytic models
• Creates test models
Analytic Models & Data Science Tools
Business Systems
Edge Cloud
Learnings from HPE customer interviews in 1HFY19
CUSTOMERS STRUGGLE TO ADOPT DATA ANALYTICS
15
Data analytics tools sprawl- Selecting the right tools takes time- Tools & versions often changing- Complex tool deployment- Data scientists expect latest and greatest
IT unable to support data analyticsfrom the agile development to the production process- Mode 1 Infrastructure (bi-modal IT, Silo’ed)- No self service options- Unpredictable growth hard to facilitate- Traditional way of working (not DevOps)- No Hybrid IT support strategy
Security hard to manage- Support multiple users and groups (internal and external)- Right level at right phase of the project- Data management (data integrity and compliancy of
sources)
Data specialists are rare- Tough to find, hard to keep (motivation)- Important to keep busy (infrastructure & tool availability)- Keep them busy with the right work
(not: installing tools, getting access to data sources)
Where do you want to play?
NEW (DATA) SERVICES OPPORTUNITIES
16
Datacenter
Server | Storage | Network
Server- | Storage- | Network virtualization
(Container) Operating System
Container 2 Container 4Container 3
Infr
aas
aS
erv
ice
Co
nta
iners
as
aS
erv
ice
Cloud (Container)Infrastructure
Service
Container 1 Container 5
Hybrid IT InfrastructureManagement
Operating System Management
Container Runtime, Packaging,Orchestration & Cluster
Management
Virtualization & CloudManagement
DataAnalytics
App 1
DataAnalytics
App 5
DataAnalytics
App 4
Self Service Portal & Catalog
Pla
tfo
rmas
aS
erv
ice
/D
ata
An
aly
tic
sas
aS
erv
ice
DataAnalytics
App 3
DataAnalytics
App 2
Build- | Deployment Automation
App Data 2 App Data 4App Data 3
Application Life CycleManagement
Continuous Integration /Continuous Delivery
Build / Deployment Management
Portal- | Catalog- | APIManagement
Data Management
App Data 1 App Data 5
Persistent Distributed (Big) Data