#ConnWeek Data Pedigree: Can you trust your data?.

#ConnWeek

Data Pedigree: Can you trust your data?

#ConnWeek Santa Clara, CA May 23-26, 2011

Automation Data Issues

• Accuracy and Precision• Over and Under Sampling• Scaling and Unit of Measure• Labeling and Searching• Sequence and Timestamp• Lost and Missing Records• Tampering and System Faults

…not an exhaustive list

2


Findings from the Field

3


Solution Example UC Davis

• Highlights:– Spark lines as visual cues– User interface via familiar map– Good data assurance through trends

4

#ConnWeek Santa Clara, CA May 23-26, 2011 5


UC Davis – Map View

6


Week on Week Power

7


Solution Example UC Rochester

• Highlights:– Classic “plant” process flow view– Very dense display of indicators and calculations– Internal and external data– Trend Forecast

8


University of Rochester

9


Plant Data

10


6 Day Forecast Link

11


Harvard Medical School• Facilities Management

– 18,000 Maintainable Assets– 24/7 Call Center– 200 Node BMS

• Energy Usage– 15 Mw Electric– 70K lbs/hr Steam– 10k+ Tons Chilled Water

• Sustainability Initiatives– Approx. 60 Primary Meters– Approx. 112 Tenant and Sub-meters– 172 Total CHW / Steam / Electric / Air


Greenhouse Gas Models

• Analyze Energy Usage

• Model Building Performance

• Long Term Data Trends


Anywhere Access

• Live Data on Blackberry for monitoring critical situations


eBay Data Center Approach

• Key Factors– Monitoring ‘baked in’ at design time– Utility integrated control of primary 138kV feeds– Constantly adding/removing racks (tech refresh)– Requires real time data and notifications

15


eBay - State of the Art Data Center

Integrated Monitoring:•Mechanical•Power•IT•Security

1,000 Devices100,000 Indicators

16


Yes – you can Trust your Data!

17

Industry Average -vs- eBay Utah Data Center


Summary

• Enable users• Provide context• Automate analysis• Get all the data• Manage change• Keep data forever

18

#ConnWeek Data Pedigree: Can you trust your data?.

Documents