Home > chip makers, data center energy measurement, design challenges, designers, developers, DFM, Intel, Semiconductors > How Intel manages IT through downturn — Server and data center optimization!

How Intel manages IT through downturn — Server and data center optimization!

September 15, 2009

Ever wondered how Intel is managing IT through the downturn — Server and data center optimization? According to Kenny Sng, data center engineering manager, Intel Technology Asia Pte Ltd, there are three key things that Intel IT does. These are:

• Internal efficiencies are critical in freeing up resources and reducing operational costs.

• Server refresh is a key strategy to ensure IT runs efficiently.

• Intel continues to look at innovation in DC operations for reducing running costs

Server and data center optimization? Intel's Kenny Sng, data center engineering manager, making a point!

Server and data center optimization? Intel's Kenny Sng, data center engineering manager, making a point!

How can IT make a difference?

* Drive employee productivity — by way of mobile client refresh

* Drive business productivity

* Continue IT efficiencies — by way of data center and server refresh

Intel data center profile

Intel has four major groups currently driving individual data center requirements (DOME).

Design:

Support the chip design community

Design Computing: Has most of the servers in Intel

Office:

Supporting typical IT and customer services

General Purpose

Manufacturing:

Manufacturing computing supporting fabrication and assembly

FAB/ATM

Enterprise:

Enterprise applications supporting eBiz and ERP

About 80 percent of servers in Intel are in D. And, 20 percent of servers in Intel are in O, M and E, categories.

Intel IT’s approach to data center optimization

Intel’s approach is very simple — standardize, improve and optimize.

Standardize

* Supply and demand forecasting

* Processes and design specs

* Overall data center structure

All of this  enables IT and consolidations, prevents unnecessary spending and ensures consistency in the overall data center structure.

Improve

* Batch processing pools via grid computing (DCV) – (D)

* Virtualization (DCU) – (O) & (E)

* Replace single core with quad-core servers

* Information Lifecycle Management

* Intel “Green” data center initiatives

* Containerized Data Centers

These go on to reduce server spending and storage/hardware expenses, Contain costs (network, power, space), simplify the environment and well, improve energy efficiency by at least 6x.

Optimize

* Close inefficient and unnecessary data centers

* Assure batch and virtualized servers are in optimal data center locations

What do these do? One, maximize data center utilization in all locations, and two, maximize server asset utilization across the world.

Data Center Balancing Act

Optimize the data center. Maximum utilization, lowest total cost! Intel generally adopts a three-pronged strategy.

Strategy 1: Data Center Optimization

Strategy 2 : Optimization + Technology

Space, floor loading, power and cooling are vital capacity components in data center utilization.

Strategy 3: Growth and location

Growth: Modular and scalable.

Locations: Hubs with local support.

Optimal efficiency: Refresh + Virtualization + Consolidation + Modularization!

“Green” data center considerations

These include:

* Air-flow management and optimization

* Economizers

* Rack level cooling

* High ambient temperature operation

* Power efficient systems

* DC power distribution

* Application productivity link

* Server power management

* Containerization (CC systems integration)

Intel IT’s server environment and refresh strategy

Office and enterprise ~105k Servers

• Currently ~20 percent utilization

Design engineering

• Grid computing

• #91 and #115 on Top 500 Supercomputer list

• ~100k Batch jobs every second, 24×7

• ~80 percent utilization

Server refresh strategy — Accelerating refresh of all servers >four-year old!

According to Kenny Sng, there is value in buying higher-end server processors. There are advantages through reduced data center space, reduced power, and lower application licenses.

Advertisements
%d bloggers like this: