1. Rich Feeds
Policy, the Cloud, and CAP
Barry Demchak
California Institute for Telecommunications and Information Technology (Calit2)
June 18, 2008
2. Overview
• Problems (RESCUE)
– Research data feeds accessible over time
– Needs for particular feeds cannot be
predicted
– Future restrictions and constraints can’t be
anticipated
• Objectives (RESCUE)
– Capture Research Data Feeds
– Expose Datasets
– Remain Flexible and Extensible
3. User View and Status
• Today’s Data Feeds
– Traffic
– Trackable Objects
– UCSD Police Cameras
– CalIT2 Cameras
– Situation Aware
• Today’s Visualizations
– Google Maps
4. Current Goals
• Empower Additional Providers
• Enable Selective Information Dissemination
• Scale to Dynamic Need
5. Current Goals
• Empower Additional Providers
– Allow Owners to
– Define dataset
– Define provider and privileges
– Define consumer and privileges
– Allow Providers to
– Contribute data
– Define consumer and privileges
• Enable Selective Information Dissemination
• Scale to Dynamic Need
6. Current Goals
• Empower Additional Providers
• Enable Selective Information Dissemination
– Allow Consumers to
– Fetch data
– Restrict data based on data or provider attributes (e.g.,
reputation)
• Scale to Dynamic Need
7. Current Goals
• Empower Additional Providers
• Enable Selective Information Dissemination
• Scale to Dynamic Need
– Data capacity
– Incoming data bandwidth
– Query bandwidth
Amazon Web
Services (AWS)
S3 – Simple
Storage Service
EC2 – Elastic
Compute Cloud
9. Logical Architecture w/Policy
• Policies – the key to empowerment
• Owner – Who can store data, who can read data
• Producer – Who can read data
• Consumer – Which providers to view
10. Physical Deployment
• Currently
– Windows Server 2003
– WinXP VM w/Mule, MySQL
• Cloud
– Amazon Servers
– Linux VM w/Mule,
MySQL,Apache/Tomcat
RESCUE Virtual Machine
Traffic
Server
Tracked
Object
Server
Browser,
Javascript,
Google Maps
Internet
Physical Machine
Other VMs
Non-
RESCUE
clients
11. Current Goals
• Empower Additional Providers
• Enable Selective Information Dissemination
• Scale to Dynamic Need
• Common Alert Protocol (CAP)
– Standard disaster alert format based on XML
– Currently used as interchange format
• CAP Server
– Always on
– Quickly scalable
– Provider-driven
– Opportunity for consumer-grade viewers
(e.g., Google Maps, Twitter, SMS, etc)