Fast Innovation requires Fast IT
Fast Innovation requires Fast IT
Cisco IT - Analytics Implementation
Vineet Jain
Cisco IT
Dramatic Internet Growth Occurring Through New Connections
Source: Cisco Internet Business Solutions Group, 2012
Internet of Things
“Fixed” Computing
Mobility / BYOD
Internet of Everything
200M
50B things
10B
Taking Advantage of Increasingly Relevant and Valuable Connections
Connections Availability, comprehensiveness, accuracy, timeliness, relevance, and richness
People, data, things Context awareness, increased processing power, greater sensing abilities
Source: Cisco IBSG, 2012
Intelligence Convergence
Visibility Security
Who Will Drive Big Data? IT Becomes More of a Strategic Business Partner
Globally, 73%say IT will drive Big Data
…in partnership with other groups
Research & Development Finance Operations Engineering
Sales & Marketing
Source: Cisco Connected World Technology Report - 2012
Cisco IT Response Data Innovation Accelerator
We will deliver…
Collaborative Use of Information in Cisco’s Ecosystem to Drive Decision Making, Operational Efficiencies and Creation of Value
Revenue/Cost Optimization Innovation Maturity Time to Value
Business Value Technology Culture
Data and Analytics Policy and Rules Framework Hierarchy Services
Data Innovation Building Blocks
Hierarchy
Management Platform
Hierarchy as a Service
Datawarehouse
Big Data
In-memory Analytics
Ops Intelligence
Data Virtualization
Rules Engine and
Repository Framework
Collaborative use of information to drive decision making, operations efficiencies, and value creation End State
Data Analytic Platforms
Integrated & Consistent Data
Data Warehouse
• Data Warehouse Appliance
• EDW – Enterprise Measures SSOTs
• Standard Enterprise BI Reporting
• Pre-canned & ad-hoc analysis
• Data mining & exploration
Platform
What Is it?
Use Case
Big Data In Memory Ops Intelligence
• Distributed Computing with Inexpensive Servers
• ~ Infinite Capacity
• ALL your DATA
• Unstructured data (Content, Documents, Logs)
• Large Structured Data
• Offload EDW ETL processing
• High Performing In-memory database
• Self Service
• Limited High Value Data
• Real Time Data
• Predictive Analytics
• What-Ifs, Modeling
HANA
UCS
• Machine Log monitoring, analysis and reporting tool
• Used for immediate action of log events
• Proactive Monitoring
• Incident Root Causing
• Design & Architecture Feedback
Data Virtualization
• Connect different data models
• Data federation, combined result sets
• Real time data across sources
• Data governance
Big Data Definition (Gartner)
VOLUME
VELOCITY VARIETY
Big data is high volume, high velocity, and/or high variety information assets that require new forms of processing to enable enhanced decision making, insight discovery and process optimization
Use Case 1 - Dynamic Insights for Sales Executives
10
Dynamic Insights for Sales Executives (DISE)
Goal Capability
Drive top-down sales management accountability for sales performance
Increased pipeline & funnel visibility Predictability in sales forecast and bookings
Use real-time data and analytics to drive execution
Sales measures by multiple lenses of the business (customer, channel, product, etc)
Insights: • Track sales funnel and forecast
accuracy • Perform P&L analysis • Negotiate with customers and
channel partners • Manage selling resources
Rapidly Adapt to Business
Needs
Accelerate
Execution
Timely Insights
into Business
Mobile Accessibility
Sales Performance
Indicators
MBR: My Business Reports
• $$ of Bookings by Level • Overall Plan Attainment
SFDC & OMF:
• Pipeline • Forecasting
PowerPoint & Excel:
• All other information, e.g. Architecture & Partner
• Limited ad-hoc drill-downs
One System: Multiple Data Sets
• Plan, Forecast, Bookings and Pipeline data
Summary and Drill-Down Views:
• Ability to show performance, comparisons and trends
• Theater, Country, Segment, GTM • Customer, Partner, • Self serve data analysis & reports
Mobile Accessibility:
Today: Desktop, iPad & iPhone
Dynamic Insights for Sales Executives (DISE)
Legacy DISE
Bookings (Every 15 minutes)
Client
Load Balanced
Self Service
IT Authored
Forecast (Every 5 Minutes)
Misc. Data
Pipeline (Every Hour)
EDW
OMF Dashboard
Explorer
Cisco HANA – High Level Architecture
DISE Run The Business (RTB) Dashboard
The DISE RTB Dashboard enables automated, self-serve access to Plan, Forecast, Bookings and Pipeline information in a consolidated view set for
Cisco Sales Management.
710,830,879 rows of Pipeline data
553,884,879 rows of Bookings data
76,756,764 rows of Forecast data
Plus all the contextual data (customers, products, etc.)
• HANA Blazing Fast (Sub-second) … Experience Platforms Slow (3-4 seconds)
• Managing data consistency across multiple data analytic platforms
• Customized views and levels of detail combined with real-time data
• Time-to-capability reduced and custom view sets built by end users in less than a day
• Summary to transaction level details in seconds
• Mobile access (iPad & iPhone) and compatible with all browsers
Value
Learnings
Value and Learnings
Use Case 2 - Cisco Partner Annuity Opportunity Management
16
Cisco Partner Annuity Opportunity Management Business Opportunity: Low value / high volume service opportunities
Designed for High Availability, Resiliency and Performance
Process and Integration of ALL Data Assets Creates Higher Value Services
Big Data PaaS Enables the Cisco Services Smart Analytics Strategy
Enhanced Opportunity Management
Ease of Doing Business Increased for Cisco Partners
Drive Incremental Bookings
Future Potential
Business Value
Daily Input Data Daily Output Data
End User
UI Tools
Data Storage and Processing
Business Rule Engine
Source
Systems
Cisco Customer
Care
EDW Reference
data
EDW
Validation Result Set
New SSOT Opportunity
Orchestration Service
Cisco Tidal Enterprise Scheduler
Install Base and Service
Contracts
Product Data Set
Service Opportunity
Data
Logical Architecture
Rules Rule to Data
Entity Mapping
Hadoop Physical Architecture
19
A new architecture called “spine and leaf” creates a maximum of one routing hop among sets of Big Data servers, with servers connected to an access switch (“leaf”), connecting to a non-
blocking, fully interconnected mesh of switch/routers (“spine”).
• Cost effective and scales linearly with addition of more nodes to the cluster. Cisco’s Common Platform Architecture(CPA) for Hadoop allowed for fast installation and efficient operations.
• Invest in Hadoop Skills. Used Hive to leverage existing SQL skills
• $40m/yr incremental service bookings estimated in first year
• Processing time reduced from 50hrs to 5hrs with 50% additional data. More than billion records processed daily.
• Business Rules Engine provided business agility
Value
Learnings
Value and Learnings
Use Case 3 – IoE in IT Operations
21
Business Need Need: Improve the Effectiveness of IT Support
Challenge
Manual Log Monitoring
No Alerts on Errors/
Exceptions
No History of Logs
Limited Visibility to Logs =
Missing Correlation and
Improper Root Cause Analysis
Solution
Proactive Performance and
Availability Monitoring
Incident Root Causing
Design and Architecture
Feedback
Business Insight
Cost Reduction for Operational
Excellence
Machine Data to Operational Intelligence
Unix, UCS, VMware Syslog
Configuration
Tickets
Application Servers
Databases
Web Servers
Splunk: The IT Data Engine No predefined schema, no custom connectors, no RDBMS
Virtualization & Cloud
Windows Linux/Unix Applications Databases Networking
• Web logs • Log4J, JMS,
JMX • .NET events • Code and
scripts
• Configurations • syslog • SNMP • netflow
• Configurations • Audit/query
logs • Tables • Schemas
• Hypervisor • Guest OS,
Apps • Cloud
• Configurations • syslog • File system • ps, iostat, top
• Registry • Event logs • File system • sysinternals
Logfiles Configs Messages Traps Alerts
Metrics Scripts Tickets Changes
Customer Facing Data
Outside the Datacenter
• Click-stream data • Shopping cart
data • Online transaction
data
• Manufacturing, logistics…
• CDRs & IPDRs • Power
consumption • RFID data • GPS data
• Uniform logging standard and log data quality is critical
• Most effective for log analytics when immediate action is needed
• Broken link reduction 95%
• Availability 99.1->99.9
• Performance SLA 98.2->99.4
• Saved 76 mins RCA per incident
Value
Learnings
Value and Learnings
Big Data Opportunities @ Cisco
Business Opportunities
• Fault prediction
• Threat intelligence
• Performance predictor
• Multimedia content analytics
• Video/audio
IT and Operations
• Performance monitoring
• Operational benchmarking
• Compliance
• Pervasive Security
Sales and Marketing
• Opportunity Mining
• Recommendation/Guided Selling
• Campaign effectiveness
• Customer segmentation
Engineering and Manufacturing
• Customer insights
• Inventory management
• Distribution and logistics optimizations
• Informed supplier negotiations
• Software Recommendation
10101010
10101010
10101010
10101010
10101010
Big Data -> Smart Data The Edge is Infinitely Scaling, But The Center Cannot
Business Process , Rules Management and Analytics
• Connect the Edge to the Center with business process, rules & analytics
• Analytics at the Edge enables faster decision making
• Doing so reduces the volume of data and therefore cost at the center
Easy Access to Platforms / Data
Self Service… with Help The
“Data Scientist”
How Is Cisco IT Encouraging Data Innovation?