how first to value beats first to market: case studies of fast data success
TRANSCRIPT
page
HOW FIRST TO VALUE BEATS FIRST TO MARKET: CASE STUDIES OF FAST DATA SUCCESS
Executive Webinar Series on Fast Data
page© 2016 VoltDB
EXECUTIVE WEBINAR SERIES: FAST DATA STRATEGY
1. Fast Data for Competitive Advantage: 4 Steps to Expand your Opportunity
2. How First to Value Beats First to Market: Case Studies of Fast Data Success
3. Fast Data Choices: Strategies for Evaluating Alternative Business and Technology Options
2
page© 2016 VoltDB
OUR SPEAKERS
3
Peter VescusoCMOVoltDB
Niall NortonCEOOpenet
page© 2016 VoltDB
DATA IS TRANSFORMING BUSINESS
4
Broad content targeting to generic viewers Smarter, more individualized customer experiences
AUDIENCES
Content Metrics
INDIVIDUALS
Consumer Centric
From: AUDIENCES To: INDIVIDUALS
page
Big Data
“Perishable insights can have exponentially more value than after-the-fact traditional historical analytics.”
Mike Gual2eri, Principal Analyst, Forrester Research
Fast Data
DATA IS TRANSFORMING BUSINESS
page
FAST = ADVANTAGE
6© 2016 VoltDB
page© 2016 VoltDB
• Forrester’s findings:• Businesses can’t get the data they need fast enough
• Data volume and variety are crushing business systems
• The mobile mind shift hinges on data, e.g., metadata that data needs to be classified, linked, and exposed to create “mobile moments”.
7
Niall Norton, CEO, Openet
9 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
• Openet has always been about real-‐=me
• Background is large scale transac=on processing, control and mone=za=on of data for communica=on service providers
• Wanted to take it to the next level to enable smarter engagement for our customers
• This helps communica=on service providers grow to be able to work and beOer compete with OTT and content providers (Google, Facebook, Amazon, Skype, NeTlix, Skype, Spo=fy, etc)
• Enables communica=on companies transform to be Digital Service Providers
Openet – Why Fast and Smart Data is Crucial
10 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
Openet – Mee=ng the Needs of The Digital Service Provider
11 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
• Advanced PCC -‐ The world's most advanced Policy and Charging suite
• Big Data Prepara2on -‐ Turn big data into smart data that delivers real business benefits
• NFV -‐ Openet’s solu=ons are all fully virtualized providing the founda=on for faster =me to market, reduced implementa=on and upgrade =me
• CEM -‐ Having smart data available to provide a holis=c view of all customers as well as understanding customer context in real-‐=me enables personalized marke=ng offers
• Network Op2miza2on -‐ Improve quality of experience, reduce cost and maximize revenue through efficient and proac=ve management of network resources
Openet Exper=se
12 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
Openet Enables Smarter Engagement
13 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
• A higher performance, in-‐memory database that could combine the capabili=es of an opera=onal database, real-‐=me analy=cs, and stream processing in one easy-‐to-‐use plaTorm.
• An in-‐memory database that could handle fast data
• Database technology that would be complimentary to our innova=ve soaware solu=ons and suitable for virtualized deployments.
• A database that was elas=cally scalable and could grow and contract as needed.
• The result – Openet is now rolling enabling smarter engagement at many of the most innova=ve service providers in world.
To Deliver Smarter Engagement Openet Worked with VoltDB to Deliver:
14 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
• Smarter Engagement with Customers – use smart data and enable a beOer customer experience and enable service providers to compete for a bigger share of customers’ digital spend.
Smarter Engagement with Customers
How do you become more relevant to your customers?
15 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
• Smarter Engagement with Real-‐2me Data – understand customer context in real-‐=me. Use this to push personalized, contextually aware offers.
Smarter Engagement with Fast Data
16 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
• Smarter Engagement with Technology – using NFV to run smarter systems, including real-‐=me charging and policy
Smarter Engagement with Technology
17 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
• Smarter Engagement with Exis2ng Systems -‐ reconfigure legacy/diverse networks and systems
Smarter Engagement with Exis=ng Systems
Be Digital Ready -‐ ‘Best of Breed ‘ adjunct approach enables fast track system transforma2on
18 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
Sample Use Cases: Used by Many of the World’s Most Innova=ve Service Providers
Shared Data -‐ Enterprise
Video Op=miza=on
Access type Policy
IN Replacement
Audience Measurement
Tradi=onal Media=on
Time of Day Pricing
Conges=on Management
VoLTE Service Enablement
Spend No=fica=ons and Bill Shock Control
Device Type Policy
Bandwidth on Demand
Fair Usage
Service Tiers
Time-‐based Service Pass
Parental Controls
Dual Persona (BYOD)
Data Volume / Speed Tiers
Data Roaming Service Pass
Data Roaming No=fica=ons
Content Bundles with OTT Services Applica=on Service Pass
Fast Device / Service Rollout
Device Tethering
Real-‐Time Contextual Offers
Shared Data -‐ Mul= Device
Network Selec=on Intelligence
Shared Data -‐ Mul= User Account
Data Giaing
Sponsored Data
19 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
Chosen by Leading Service Providers
20 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
Ope
net
Best of breed
Cross functional for growth & innovation
“Big stack” guys Quality
Applicability
Flexibility
Compatibility
Expansibility
Performance
Quickly redesign services for a dynamic market
Gets along well with other systems
Cloud ready, hardware agnostic
Industry leading
Afterthought or offloading altogether (e.g. NSN)
Closed silo designed for yesterday
Submit change request. Cross fingers.
Vendor lock in
Proprietary
Demand Overwhelms
Why We’re Different
21 © Copyright 2016 Openet – Company Confiden=al For Use Under Non-‐Disclosure Only
• Telecoms is transforming
• Everyone had a strategy but need the flexibility to adapt in =mes of change
• Those who don’t best adapt to change will be lea behind
• Legacy way of doing business and systems will soon be obsolete
• Not just about big data. It’s using data in a fast and smart way to drive change and open new revenue streams
• It’s about enabling change
Summing Up – Openet and VoltDB
page© 2016 VoltDB 22
From Development to Release
First to Market First to Value
From Development to hitting Sales and Profitability Goals
When First to Market doesn’t lead to First to Value, it’s due to either the wrong solu7on or the wrong technology pla;orm.
versus
page© 2016 VoltDB
FIRST TO VALUE WITH FAST DATA – THE CHALLENGES
• Fast data applications have different technology requirements
• Early adoption of technology doesn’t guarantee success
• Many technology options
• Need to pick the business and technology strategy that’s right for you
23
page© 2016 VoltDB
Batch/IterativeAnalytics+
Big DataFast Data
Rapid Data Ingestionand
TransformationStreamingAnalytics
Operational Interaction/ Transactions
COMPARISON OF FAST AND BIG
page© 2016 VoltDB
COMPETITIVE STRATEGY DRIVES TECHNOLOGY AND DATA MANAGEMENT REQUIREMENTS
25
Hyper Personalization
Real-Time Resource Management
Real-time Policy Enforcement
IoT & Sensor Data
page
WHAT’S YOUR CORE COMPETENCY?
- CUSTOMERS AND APPLICATIONS
- DISTRIBUTED SYSTEM INFRASTRUCTURE
26© 2016 VoltDB
page© 2016 VoltDB
EVALUATION CRITERIA
Criteria Considera2ons
Data Volume & Velocity Capacity to ingest, process and export at speed of data
Response speed, Performance Need for interac=ve, real-‐=me
Personaliza=on Batch vs con=nuous event processing
Accuracy, Data Consistency Is data high value, cri=cal?
Scalability Accommodate rapid growth. Cloud-‐ready
Standards SQL for data abstrac=on vs Applica=on heroics
Skill Set Specialty open source skills, e.g., Cassandra
27
page
SOME TECHNOLOGY OPTIONS…
28© 2016 VoltDB
page© 2016 VoltDB
THE “DIY” DATA INFRASTRUCTURE
29
GlueCode
GlueCode
Community Supplied You write this
Zookeeper
page© 2016 VoltDB
THE “DIY” DATA INFRASTRUCTURE
30
GlueCode
GlueCode
Community Supplied You write this
Zookeeper
Implications- Need a specialized skill set- Development: more work to write glue code, test and QA system for potential failure modes- Support: test and maintain “glue” code with each component release
Bottom line: - More $ invested in developing data infrastructure- Longer time to value
page© 2016 VoltDB
THE “DIY” DATA INFRASTRUCTURE VS VOLTDB
• Rigorous testing and QA• 1/4th of the components• Simpler, Faster• SQL and Java• Easier to test, maintain applications
GlueCode
GlueCode
Zookeeper
page© 2016 VoltDB
BATCH PROCESSING VERSUS CONTINUOUS EVENT PROCESSING
• Batch processing is an efficient way of processing large volumes of data• Collect – Process – Report
• Fast data processing involves a continuous process; each event is treated individually• Ingest - Analyze - Act
32
page
BATCH PROCESSING
33
Event OccursAnalyze, Gain Insight
Take Action
Collect Data Process Data Act on the Data
TimeNow Later
page
CONTINUOUS EVENT PROCESSING
Analyze, Gain Insight
Take ActionEvent Occurs
34
TimeNow Later
page© 2016 VoltDB
SQL VERSUS NOSQL
35
• SQL (structured query language) is for relational databases
• Powerful query language
• Standard and widely adopted
• Flexibility - abstracts application from the data
• ACID transactions – ensures immediate data consistency, reliability
• NoSQL
• Analytics are difficult/painful due to ridged data model
• Non-standard programming interface (each product is different)
• Lack of SQL and ACID transaction guarantees drives complexity to the Application
Ø Data integrity becomes the job of the Application developer
page© 2016 VoltDB page
CASE STUDIES
36
page© 2016 VoltDB
Personalized trade recommendations
Business challenges:- “Interactive” speed- Personalized offers- Data accuracy,
integrity (compliance)- Multiple data sources
CASE STUDY: FINANCIAL SERVICES
page© 2016 VoltDB
CASE STUDY: FINANCIAL SERVICES
38
Data Sources
Rules Engine
In-Memory Grid
AppApp App
• Event data from multiple sources
• Each application database replicates to Cassandra and Hadoop
• In-memory grid used to maintain logic and publish ‘state’ back and forth
• Rules engine with fast access to Cassandra
• MySQL used for slow-changing data
page© 2016 VoltDB
BEFORE
Data Sources
Rules Engine
In-Memory Grid
AppApp App
page© 2016 VoltDB
BEFORE
40
Data Sources
Rules Engine
In-Memory Grid
AppApp App App App App
AFTER
Data Sources
page© 2016 VoltDB
CASE STUDY: FINANCIAL SERVICES
Resultsü Simplified system architecture
ü Immediate data consistency
ü Real-time recommendations
ü Faster time to value
41
page© 2016 VoltDB
CASE STUDY: MEDIA AND ENTERTAINMENT
Content Delivery Network Service Provider
Business challenges:- Real-time analytics for customers
- Data accuracy: over/under billing
- Scalability
42
page© 2016 VoltDB
CASE STUDY: MEDIA AND ENTERTAINMENT
43
page© 2016 VoltDB
CASE STUDY: MEDIA AND ENTERTAINMENT
44
Resultsü Simplified system architecture
ü 1/10th the compute resources
ü 100% budget accuracy, eliminated $$$ under/over spending
ü Faster time to value
“We chose to go with VoltDB over other streaming aggregate solu2ons (like Trident) for its SQL interface, real-‐2me Ad-‐Hoc queries over our raw data, and simpler overall design” Behzad Pirvali, Architect, MaxCDN
page© 2016 VoltDB
CASE STUDY: INTERNET OF THINGS
IoT Device Manufacturer Platform- Smart devices, appliances
Business challenges:- High volume and velocity of data from
smart devices- Complexity (multiple ingest points, apps,
databases) - Performance – need to automate action on
inbound data at the velocity of the feeds
page© 2016 VoltDB
CASE STUDY: INTERNET OF THINGS
46
Device Data
Rules Engine
In-Memory Grid
AppApp App
• Device data flows from cloud from multiple devices, appliances
• Each application database replicates to Cassandra and Hadoop
• In-memory grid used to maintain logic and publish ‘state’ back and forth
• Rules engine for intra-day data to trigger actions (e.g., ‘turn lights on’)
• PostgreSQL used for dimension data
page© 2016 VoltDB
BEFORE
Device Data
Rules Engine
In-Memory Grid
AppApp App
page© 2016 VoltDB
BEFORE
48
App App App
AFTER
Data SourcesDevice Data
Rules Engine
In-Memory Grid
AppApp App
page© 2016 VoltDB
CASE STUDY: INTERNET OF THINGS
Resultsü Simplified system
architecture
ü Single ingest point for high-velocity feeds of inbound data
ü Faster time to value
49
page© 2016 VoltDB
WHY VOLTDB?
Faster
Smarter Simpler
Our customers realize exceptional business value
page© 2016 VoltDB
QUESTIONS?
• Use the chat window to type in your questions
• Try VoltDB yourself:
Ø Free trial of the Enterprise Edition:
• www.voltdb.com/Download
• Email us at: [email protected]
51