greenlight data collection architecture
DESCRIPTION
Data collection architecture overview by Ingolf Krueger, Claudiu Farcas, 2010TRANSCRIPT
GreenLight Project
Data Collection meetingData Collection meeting
Claudiu Farcas, Filippo Seracini, Ingolf Krueger
2
GreenLight Project Our objectives
• Create an energy model for the BlackBox– Understand energy consumption related to task execution– Optimize task scheduling to decrease the environmental
impact– Provide to user different modes of computation (i.e. max
performance, max energy saving, min computational cost, etc)
• Create the cyberinfrastructure to – Manage and control the Blackbox – Run tasks– Provide green data related to task execution
3
GreenLight Project First steps
• Requirement engineering with the different teams involved in the project– What are the data sources?– What can be measured?– How is data stored?– How is data represented?– Who wants what?
4
GreenLight Project
Stackeholder ProviderConsumer
Networking (Brian)
Ingolf
DC Power Testing
Tom/Phil (storage)
Tajana
Ingolf
Falko
Gupta
Amin
Research Distribution
• GreenLight Researchers are interested in both producing and consuming greening data such as temperature and power measurements
Greenlight Instrument
Single Instrument
Aggregate Instrument
*
Research
Greening ON
Greening OFF
5
GreenLight Project Existing Capabilities
• Default Sun Monitoring interface
6
GreenLight Project Initial Domain Models
• Data Collection Points
Node
Server ConveyStorage
Rack
Thumper Database
GPGPU
Processor
Hard Drive
Memory
Network Interface
Temperature
Heat Exchanger
Fan
Fan Slice
Speed %
5
Panel
A B
Amps
Volts
Watts
2
PDU
IP Address
Phase
3
Phase 1
BlackBox
FrequencyPhase 2
Phase 3
2
1
8
Relative Humidity
Dewpoint
Outlet Inlet
Water
BB1
BB2
BB3HumidityExternal
Internal
7
GreenLight Project Resource Allocation
• Requirements elicitation at the 2nd meeting
8
GreenLight Project Next Steps
• Define a model for “greening data”
• Define a resource model for GreenLight resources (e.g., CPUs, VMs, nodes, etc)
• Model GreenLight resources in the context of the Rich Services framework (SOA derivate)
• Identify usage policies and adequate scheduling algorithms to improve efficiency
• Create the Rich Service infrastructure to use and manage the Blackbox
• Expose the “green data” as web services & portal