DC Energy, Efficiency Challenges Choosing a density strategy Implementing a high density environment Maximizing the efficiency benefit Anticipating the dynamic data center Key points on energy usage in Data Centers Munib Khawaja Vice President IT Business Pakistan and Afghanista
The energy dilemma is here to stay The facts The need Energy demand By 2050 Electricity by 2030 Source: IEA 2007 vs CO 2 emissions to 2 avoid dramatic climate changes by 2050 Source: IPCC 2007, figure (vs. 1990 level) Frequent power outages Rising energy prices Climate change Conflicts for resource access & control 2
More than 175 years of history Steel Industry 1836 Creation of Schneider at Le Creusot, France 1999 Groupe Power Schneider & becomes Schneider Electric, focused Control Power & Control 1996 Modicon, historic leader in Automation, becomes a Schneider brand 1991 Square D joins Groupe Schneider 1988 Telemecanique joins Groupe Schneider 1975 Merlin Gerin joins Groupe Schneider Energy Management 2011 Acquisition of Telvent 2010 Acquisition of Areva s distribution activity 2008 Acquisition of Xantrex 2007 Acquisition of APC corp. and Pelco 2005 Acquisition of Power Measurement Inc. 2003-2008 Targeted acquisitions in wiring devices and home automation (Lexel, Clipsal, Merten, Ova, GET, etc.) 2003 Acquisition of T.A.C 2000 Acquisition of MGE UPS Systems 19th century 20th century 21st century 3
Schneider Electric the global specialist in energy management Balanced geographies FY 2011sales billion sales (last twelve months) of sales in new economies (last twelve months) North America 23% Rest of World 18% Western Europe 32% Asia Pacific 27% people in 100+ countries of sales devoted to R&D Diversified end markets FY 2011 sales Utilities & Infrastructure 24% Industrial & machines 22% Data centres 16% Non-residential buildings 29% Residential 9% 4
Providing integrated solutions Integration Make energy visible Make systems work together HVAC control Lighting control Access control Video security Electrical distribution Energy monitoring Motor control Critical power IT data Renewable energies Efficient & productive: Measure and control energy, automate, provide relevant diagnosis Manage processes Make all the utilities of any Infrastructure more efficient Reliable Prevent from power outage & quality variance Safe Protect people and assets Transform and distribute power safely Green: Make the connection of renewable energy sources easy, reliable and cost-effective 5
Data Centers Solution Sub-Systems Reference Design StruxureWare Operation Services Project Services Management & Security 6
Virtualization myths Virtualization always increases efficiency Virtualization requires high density data centers High density and high efficiency cannot coexist I don t need to worry about power and cooling when virtualizing 7
Three ways we can help with virtualization projects Assess the impact of consolidation / virtualization strategy Effect of under-loaded CRACs, chillers, generators Potential hotspots Identify issues in system design or operation that compromise efficiency and recommend solutions: Row-based cooling Scalable UPS Predictive management tools Estimate potential efficiency gains to enable return-oninvestment (ROI) calculations for capital expenditures Data Center Electrical Efficiency Assessment service 8
Which density strategy? New or existing data center? How many high-density servers? How much bulk power and cooling? Can IT devices be moved around? Low-density racks with no virtualized servers Low-density racks with a mix of virtualized and non-virtualized servers High-density racks with virtualized servers Spread the load Consolidate the load White paper Spectrum of post-virtualization densities 46 9
Implementing a virtualized environment Spectrum of solutions for entire density spectrum Blanking panels Supplemental cooling Rear Air containment Quick, least expensive InRow cooling Modular power distribution Hot aisle containment Best efficiency 10
Cooling: Dynamic response to hot spots Row-based cooling allows Short air path between cooling and load Instrumentation for coordinated response to hot-aisle temperature Room CRACs removed CRAC Row CRACs sense elevated temperature and increase fan speed to remove extra heat from hot aisle CRAC Hot spot When temperature decreases, row CRACs decrease fan speed to conserve energy CRAC Row CRACs added Hot spot CRAC CRAC CRAC White paper 130 1 2 3 11
Cooling: Dedicated Pod A high-density island in the room White paper 134 A mini data center with its own cooling Contributes no heat to rest of data center Works with existing room-based cooling Hot/cool air circulation localized within the pod by short air paths and/or containment Achieves optimal efficiency Targeted availability 12
Cooling: Dedicated pod Exhaust air is captured within hot aisle and neutralized to ambient emperature Variable-speed fans optimize efficiency by closely matching performance to dynamic cooling demand Ambient-temperature air is returned to room FRONT CR RAC CR RAC CR RAC Rack Rack Rack Rack REAR Hot aisle Contained hot aisle FRONT CRAC CRAC CRAC Rack Rack Rack Rack InRow cooling unit Equipment racks take in ambient air from front 13
Maximizing the efficiency benefit 4 servers Before virtualization 2 servers After 2:1 virtualization Fixed loss GONE Fixed loss GONE 2 of 4 servers 2 of 4 servers unplugged X X Unplugging servers removes THEIR fixed losses BUT what about power and cooling fixed losses? 14
Another way to look at fixed losses DCiE = 50% DCiE = 33% Pre-virtualization: 100 kw of servers Post-virtualization: 50 kw of servers motors 100kWFan motors Servers 100 kw83% 50% FIXED LOSS Servers 100 kw 50% Total power consumption reduced by server reduction Fan Fan motors 100kW motors (no 100 change) kw 67% FIXED LOSS Servers 50 kw 33% Total electric bill is smaller, but the same FIXED LOSS is now a greater portion of a smaller pie 67% reduction in server power results in 56% electric bill savings 50% reduction in server power results in only 25% electric bill savings 15
Case study Annual electric bill 120 kw data center capacity 90 kw IT load (75% loaded) 59 kw total server load (66%) DCiE = 49% 53 kw IT load (42% loaded) 22 kw total server load 75% servers virtualizable Server consolidation ratio 20:1 DCiE = 39% 60 kw capacity Data center load 88% DCiE =62% $193,123 Before Virtualization 27% savings $140,305 After Virtualization 36% savings savings After physical infrastructure improvements Average 7 kw / rack DX air conditioning No redundancy $0.12 / kw hr Right-sized power & cooling Close-coupled cooling Use blanking panels High-efficiency UPS (96%) Source: TradeOff Tool - TT9 Rev 0 Virtualization Energy Cost Calculator 16
Scalable infrastructure minimizes waste Minimize the inefficiency of oversizing during consolidation and re-growth and be prepared for higher densities to come Power/cooling Capacity Virtualized Load Original Load Virtualized Load Scale DOWN Virtualized Load Load Rack density Virtualized Load Load Scale UP Load Scalable power and cooling results in better DCiE 17
Schneider Data Center Management 18
Operations Suite Gateway Web services Enterprise platform Capacity and Energy Inventory and Reporting Documentation Informed decisions Dashboard Monitoring Data Center On-the-go Virtualization Smart phones and tablets Cloud enablement Web services Gateway 19
StruxureWare for Data Centers Energy and Sustainability Management High level reporting and dashboard Schneider Electric Data Center Facility Management (DCFM) Data Center Power, Cooling & Security Data Center Infrastructure Management (DCIM) Data Center Operations IT Service Management (ITSM)* IT Servers, Storage, Applications Schneider Electric Schneider Electric Cisco 20
Key points on energy usage in Data Centers 1. IDC shows that energy expense associated with powering and cooling the worldwide server installed based has increased 31.2% over the past five years. (IDC, 2010) 2. More than 50% of power going into a typical data center goes to the power and cooling systems NOT to the IT loads! 3. For each 1/10 of a point of PUE improvements in data centers, IDC estimates that $1.2B is saved in energy costs (or 11,500 GWatts of energy, or 17.3 Tons of CO2 emissions or 3M cars off the road per year). (IDC, 2012) 4. The typical 1MW (IT load) data center is continuously wasting about 400kW or 2,000 tons of coal per year due to poor design (DCiE = 50%, instead of bestpractice 70%) 5. Every kw saved in a data center 1. saves about $1,000 per year 2. reduces carbon dioxide emissions by 5 tons per year 3. has a carbon reduction equivalent to eliminating about 1 car from the road White paper 66 6. A 1% improvement in data center infrastructure efficiency (DCiE) corresponds to approximately 2% reduction in electrical bills 21
Thankg Make the most of your energy Schneider Electric Data Centers Business-wise, Future-driven 22