measurement and management technologies...

© 2010 IBM Corporation

Measurement and ManagementTechnologies (MMT)

Data Center Solutions for a Smarter Planet

IBM Systems Lab Services and Training

© 2010 IBM Corporation2 World Wildlife Fund Climate Savers Meeting

IBM Corporate Environmental Affairs – Measurement and Management Technologies

AGENDA

� IBM Energy Efficiency Program & Accomplishments

� IBM CO2 Reduction Commitment

� Featured Solution: Measurement and Management Technology 1.5 System



IBM’s Commitment to Energy Efficiency

� IBM first established its Corporate Policy on Energy Conservation in 1974– Formalized energy conservation program continues unabated to this day

� IBM’s Environmental Management System objectives include achieving and sustaining progress in

– improving the environment by maintaining a position of leadership in energy conservation; and

– reducing costs and increasing competitiveness and shareholder value through gains in energy efficiency.

� Achieving energy efficiency is one aspect of IBM’s five-part strategy to reduce the company’s operational greenhouse gas emissions

– IBM has a corporate goal to achieve annual energy conservation savings equal to 3.5% of IBM’s total energy use

� IBM’s experience in energy management and demonstrable results inform and contribute to the company’s efforts in developing capabilities and solutions to bring value to its clients



IBM’s Operational Energy Efficiency Performance

� Between 1990 and 2009, IBM saved 5.1 billion kilowatt-hours (kWh) of electricity consumption, avoided nearly 3.4 million metric tons of CO2 emissions (equal to 50 percent of the company’s 1990 global CO2 emissions) and saved over $370 million through its annual energy conservation actions

– IBM’s energy conservation goal recognizes only one year’s MWH savings from identified projects that actually reduce or avoid the consumption of energy in its operations

– Reductions in energy consumption from downsizings, the sale of operations and cost avoidance actions are not included in the energy conservation performance

� In 2009, the energy IBM saved through over 1700 energy conservation projects was equivalent to 5.4 percent of its actual energy use.

– 246,000 megawatt-hours (MWH) of electricity; – 410,000 million BTUs of fuel oil; – nearly $27 million savings in energy costs– Saving Focus Areas:

• Time of Day Scheduling and More Efficient Hardware for Lighting & HVAC Systems• Building Recommissioning Efforts• Data Center Thermal Management and Server Virtualization Projects• Central Utility Plant and Building Control and Analytics Systems



IBM CO2 REDUCTION COMMITMENT

� CORPORATE GOAL:

Between 1990 and 2005, IBM’s energy conservation ac tions reduced or avoided CO2 emissions by an amount equal to 40 percent of its 1 990 energy use. To further extend this achievement, IBM set itself an aggressive “2nd generation” goal: to reduce the CO2 emissions associated with its energy use 12 per cent by 2012 against a 2005 base year through energy conservation and the procuremen t of renewable energy

� IBM CLIMATE SAVERS GOAL:– Corporate CO2 Reduction Goal– Data Center Energy Efficiency Improvements:

• Thermal Management with MMT 1.0/1.5 (IBM Featured Solution)• Server and Storage Virtualization• Dynamic Cooling • Data Center Design Innovations

� Today’s Presentation will Focus on the Results of the MMT 1.5 Sloution


DATA CENTER ENERGY EFFICIENCY



Rack Design

Product Design & Workload/Energy Management

Thermal Load Management

Power Management :

Measure

Cap

Adjust to

Workload

Virtualization

Power & Workload Management:

Data Center Network

Integration Maximizes Power Savings

Rack/Room Layout:

Air Flow Management

Liquid Heat Removal:

Variable Cooling Delivery :

Facilities SystemImprovements

Integrate Cooling Delivery with Equipment Demand

Optimize Delivery of Power and Cooling to The Data Center(Reduce Wastage)

Reduce System Power Losses Through Utilization of More Efficient Equipment

DataCenter Energy Efficiency Opportunities

Component Power Efficiency

FACILITIES

IT SYSTEMS


SOLUTION:

MEASUREMENT & MANAGEMENT TECHNOLOGY 1.5



What is MMT?

Data Center Measurement and Management Technologies• Real-time and High resolution Measurements

• Measurement-based Modeling • Management and Control

High resolution measurements are made using IBM’s Mobile Measurement Technology



MMT provides both high time & spatial resolution co mbining� High resolution measurements / assessments for base model generation, sensor placement etc.� Real-time sensing for feeding dynamic models

� External and internal (via AEM, ITMfEM)� Wired and wireless solutions� Sensor grid solution with 1-wire protocol � Access points every 2-4’ for thermal, flow, pressure , and

relative humidity (all with the same network) � Acoustics, power, current and corrosion capabilitie s are

being developed.

� 3D dimensional mapping tools for detailed environme ntal measurements

� Adapters to existing data sources (BMS, asset db)

Dynamic Monitoring and Management

(MMT 1.5)

Measurement Survey and Analysis

(MMT 1.0)

Measurement and Management Technologies



3 Implementation of “ Best Practices ”

� Realize air transport energy savings

� Realize thermodynamic energy savings

� Achieve reduced energy consumption

� Potential for deferring new investments

1 Measure

� Capture high resolution temperature data, air flow data and infrastructure & layout data

2 Analyze / Model

� To identify improvement opportunities model the data center and use optimization algorithms (“best practices rules”)

Solution Approach – Three Steps

MMT 1.0 Measurement and Survey Analysis



Savings Results From an MMT 1.0 Survey

� Surveyed 50,000 square feet of Raised Floor Space�One-half million thermal, humidity and flow measurements�1000 rack inlet temperatures with ~ 5000 servers were measured �measured air flow from all perforated tiles (total of >1600 flow measurements)�inspected all 54 / 54 (active) ACUs*�measured / estimated all relevant DC power levels

� Energy saving opportunities of up to 3.3 MW (~17% o f IT)� improve Air Conditioning Unit (ACU) utilization level from 59 % to 75 % (turn 14 ACUs off)� higher ACU utilization reduced plenum temperature by 3 F� reduce hotspots by 7 F in key areas using air flow management techniques� increase chiller set point to increase chiller coefficient of performance� leverage plate heat exchanger to bypass chiller and increase free cooling durations

$17.70$27.54$11.84Energy Savings/sf for 3 years

$903,000$1,404,000$603,000Potential 3 yr savings

$301,000$468,000$201,000Annual potential energy reduction of 3.3 MWh

$5.90$9.18$3.94Energy Savings/sf per year

$0.09 / kWhour

National Avg

(estimated)

$0.14 / kWhour

Germany and Denmark

(estimated)

$0.06 / kWhour

(actual)

6-month payback, 100% ROI on energy alone at $0.06 / kWhour



MMT 1.5 Dynamic Monitoring and Management

Development of a real-time Measurement and Manageme nt Technology (MMT 1.5)� Data centers are dynamic: MMT 1.0 is sparse in time but dense in space � Real-time sensors are sparse in space but dense in time

� MMT 1.5 provides high time & spatial resolution� 3D temperature distributions with cross-sections, inspections etc.� alarm/threshold services � efficiencies in real-time and corresponding cooling zones



� Process for MMT 1.5 Set-up�Use MMT 1.0 service to Identify Efficiency Improvements & Lay Out Sensor Network �Install (real-time) sensor network (thermal, flow, humidity, pressure) �Create MMT base model and divide Data Center into cooling zones

� Benefits:�Monitor/compute local efficiency changes in real time with Dashboard & Weekly Reports�Optimize Cooling Load as Equipment is Refreshed/Added or IT Power Varies�Raise Overall Raised Floor Temperate to New ASHRAE Standard – Reduce Cooling Reqs

Dynamic Monitoring and Management – Process and Benefits



Dynamic Monitoring and Management – Installation

� Sensors

� End of row rack with sensor at 5.5 ft

� Temp/flow sensor

� Under floor cable mounted close to the raised floor

� Tsense mounted on door

� Temp/humidity sensor

� Temp/pressure sensor



Dynamic Monitoring and Management - Deployment

� Sensor access points are installed every ~ 2-4 feet (T-boxes)

� Install strategy allows maximum flexibility for sensors to be moved, removed and added

� Sensors are bar-coded and location / ID is tracked on MMT web client

� Thermal / humidity / pressure / flow sensors are supported on the same network

� Wired and wireless sensors are supported

Sensors

Sensor Hub

T-box



Dynamic Monitoring and Management – Grid Design

� Existing area

� Build-out area

� Background from AutoCAD� Red = existing racks� Blue = future racks

Hub Ch 1Hub Ch 2Hub Ch 3Hub Ch 4

Temperature SensorPressure SensorHumidity SensorFlow Sensor

T-BoxHub



Dynamic Monitoring and Management - Cooling Zones

Zone boundaries

Zone is 56% efficient

Bar heights indicateefficiency of

respective zones

ACU Info



Realized Energy Savings

Additional possibleEnergy Savings

Possible capital Cost avoidance

Dynamic Monitoring and Management– Active Data-Center Management (Client example A)

Four types of Savings� Energy Savings� Delay of capital investments� Increased reliability (cost avoidance)� Lost revenue avoidance

Customer Example:- DC power changes all the time- without active management, ACU utilization followed power density (blue curve)- customer started to use MMT 1.5 to actively manage ACUs to maintain constant 75% utilization level



• rapid ROI can be accomplished by active management to a prior ACU utilization level • after ~ 30 weeks more than $1 / sqf of hard savings (@ $0.12 per kW hour )• another 20 % more energy savings are possible if managed to design point• operation at 85 % of ACU utilization allows also increasing the power density from 77 to 93

W/sq ft, which can save an additional ~ $25 / sqf per year * in capital cost avoidance

�* @ ~ $1000/sqf construction cost and 5% COC over 25 years

Dynamic Monitoring and Management – Hard Savings (Client example A)



IBM Internal Deployment

Measurement and Survey Analysis successfully deploye d and implemented in more than 1M sq. ft. of internal DC s pace� More than 20% cooling energy savings� Almost 8M kWhour of savings today� Over 25 IBM strategic data centers assessed

Dynamic Monitoring and Management successfully deployed at large-scale production DC� To actively manage large-scale dynamic DC� Additional installations underway � Wider rollout in NA and in EMEA

Dynamic Monitoring and Management is part of the Raleigh NC Leadership Data Center design and architecture� Full deployment in this new Leadership DC� Integrated into ITD data models & asset

management systems, Maximo and Tivoli



NEXT STEP: INTEGRATED MANAGEMENT OF DATA CENTER ASSETS AND SYSTEMS

� INVENTORY and TRACK IT ASSETS and LOCATIONS:– Power Use Distribution– Match Available Rack Space with Available Cooling– Manage Equipment Refreshes

� DYNAMIC CONTROL OF COOLING DELIVERY– Algorithms to Control Cooling Delivery Based on Sensor Data

• IT Equipment Power Management will Vary Power Load with Time• Workload Management Based on Energy Use will enable Idling of Equipment at Low

Workload Times.• Integrate Data on Facilities and IT Equipment

– Track IT Equipment Power Use through Active Energy Manager (Equipment Based)

� OPTIMIZE DATA CENTER ENERGY USE ON A REAL TIME BASIS.



MMT integrated with Building Management System, IT to improve availability, capacity planning & energy management

IT and Facilitiesmetrics and dashboards

Alerting to business impacts

Optimal asset placement

Optimize energy withthermal mapping

IBM Tivoli MonitoringITM for Energy Management, TBSM, Netcool

Active Energy Manager, Director

OSISoftPI

IBM Asset Lifecycle Management

Key data, facility requirements

IBM MMTThermal measurement,

modelling, spatial planning

IT Server data; Asset data

DATA CENTERRoom sensors; CRACs; PDUs

BUILDINGSiemens BMS; SquareD



MMT Variable Frequency Drive Control

MMT Sensor and control network

Variable capacity AC unit (ACU)

� >2x reduction in (ACU) cooling power with no impact on inlet temperatures 51.1 F 93.2 F

Inlet Temperature Histogram

max. Temp.

Inlet temps [F]

# of

occ

uran

ces lower temps with

less energy

BEFORE: ACU Power Consumption = 11.5 kW(ACU#1 @ 100% + ACU#2 @ 100 % = 24800 cfm)

Inlet Temperature Histogram

max. Temp.

Inlet temps [F]

# of

occ

uran

ces

AFTER: ACU Power Consumption = 3.5 kW (ACU#2 @ 84 % = 10416 cfm)



• MMT Measurement and Survey Analysis can Deliver:

• Up to a 10% Reduction in IT power costs

• 10- 20% Reduction in cooling power costs

• Opportunities Greatest at Data Centers that have not Managed Cooling (most)

• MMT Dynamic Monitoring and Management Offers Continuous Optimization

• Manage Cooling Through Equipment Refreshes and Changes in Data Center Operation

• Integrate Asset Data

• MMT Dynamic Monitoring and Management Can Match Heat Generation to the Thermal Load:• Maximize Benefit of IT Equipment Power Management Capabilities.• Functionality is being integrated into IBM’s Software stack

Summary



Program Contacts

� Greg MallareOpportunity Manager, Data Center Services

[email protected]

� Greg HarwickManager, IBM Systems Lab Services and Training – Data Center Services

[email protected]

Or for more information contact us at: [email protected] .com



Special notices

IBM, the IBM logo, ibm.com, Let's Build A Smarter Planet, Smarter Planet and the planet icons are trademarks of International Business Machines Corporation, registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the Web at www.ibm.com/legal/copytrade.shtml.

The results presented are illustrative only and not guarantees of future performance. Actual results may vary.



Dynamic Monitoring and Management - Benefits

� Maintain high saving rates � Implement new ASHRAE guidelines� Provides additional means to increase ACU utilization

• Currently average ACU utilization levels are still below 50% (2x over capacity) • 1% improvement of ACU utilization can save 0.5 % of power

� Manages change in data center• Equipment placement• Dynamic cooling capacity to match cooling to IT power

� Provides unique visibility and control into the DC space• Dashboard / summary views developed

� Global IBM Systems Lab Services and Training support & Tivoli Maximo SW support� MMT 1.5 is an open/integrated system with ongoing IBM Research development



Dynamic Monitoring and Management - Architecture

HOST

data



Dynamic Monitoring and Management - Implementation

� Detailed ways to manage the data center using an MMT layout editor� Real-time, 3D temperature, humidity, pressure distributions with cross-sections etc.� Alarm / threshold services � Efficiencies in real-time and corresponding cooling zones � Reports and energy efficiency summaries� Available in Maximo for Energy optimization (MEO) as a product and as a service



Dynamic Monitoring and Management - Alarms

� User-specified thresholds� E-mail alarm alerts



12/8 1/19 3/2 4/1370

80

baselinemonitoring

result monitoring

Hotspot mitigation work increase of setpoints

turn off ACUsand adjust settings

Mod

eled

inle

t tem

pera

ture

[F]

date

� MMT provided energy savings without compromising in let temperatures (reliability of IT equipment)– hotspot mitigation was very effective driving down average inlet temperatures– specific actions reduced significant hotspots locally

- in one example, MMT alerted the client to adjust “sub-optimum” ACU settings, which hadcaused significant hotspots

- the BMS did not detect problems with the ACU settings – decommissioning of ACUs reduced inlet temperatures even more

(low-utilized ACUs discharge warm air into the plenum)– combination of turning off ACUs and hotspot mitigation allowed safely increasing set points

Dynamic Monitoring and Management - Hard Savings (Client Example B)

Modeled Average Inlet Temperature Estimated Energy Savings

21/1 1/5 2/9 3/16 4/200

10000

20000

30000

40000

50000

60000

70000

80000

90000

100000

110000

120000

turn off ACUs

setpoint increase

Est

imat

ed E

nerg

y S

avin

gs [k

Who

ur]

wea

ther

- n

orm

aliz

ed

date

1/521/1



MMT Integration – Active Energy Manager, ITMfEM, Maximo

IBM Tivoli Monitoring for Energy Management

Graphical view with key energy & thermal

metrics

IBM Systems Director Active Energy Manager

INFORMATION TECHNOLOGY FACILITIES & PROPERTY ASSETS

IBM Maximo Asset Manager for Energy

Optimization

IBM Tivoli Business Service

Manager (optional)

Role-based energy dashboard

Control & Power

Management

Centralized Energy Views

Energy Reporting

Delivering key energy management metrics through an integrated solution

measurement and management technologies...

Documents