Data Centers; a "hot" topic these days!
posted by Bradley Ellison on November 28, 2006
Although processors, with their new multi and many core architectures, are now focused on performance per watt, and new low power platforms are being announced with increasing frequency by many vendors, the attendant problems surrounding data center capacity planning, design, and management will not go away. They are here to stay.
In the press and at many conferences that I have attended, the collective finger is often pointed at Intel and the processors that we design and manufacture. Our competitors have been quick to seize on this opening and use it to gain market share. Granted, we (the collective we of Intel) have been slow to acknowledge the need for low power consumption designs. As recently as a year ago, there was ongoing debate within Intel’s design engineering community regarding the priority for and value of low power designs vs. “the need for speed” (there may still become pockets of resistance here).
But I would argue that Intel is not the culprit here. It is an industry wide information technology issue!
Our newest processor designs (as well as those of our competition) have clearly comprehended the fundamental marketplace shift requiring lower power consumption with high performance. In spite of this, ASHRAE (American Society of Heating, Refrigeration, and Air-conditioning Engineers) projections indicate that increases in thermal density of computing equipment in data centers will continue to rise as far out as 2013 (albeit at a bit slower pace). This includes compute platforms and storage devices. Even more striking, the thermal density of high-performance communications equipment is projected to rise to the top of the list of heat producers over the next several years!
In upcoming blogs, I hope to unravel some of the intertwined causes, the implications and some of the current approaches being used to address the issues. Somewhere down this path, I also want to explore the broader implications to business, the economy, and the environment - but then there is whirled peas (World Peace), and this ain’t no Miss America Pageant! ;-)
Bradley (Brad) Ellison Manager, Global Communications Services Information Technology
Comments (1)
tagged: capacity, Data Center, energy efficient, heat, planning
Comments
Apr 02 | Tony Cucinotta said:
While reviewing an article (Source: Flomerics, Inc) on Energy Fluxes from a Typical Quad Flat Pack I found the below statement to be pertinent to a project I’ve been developing.
Conduction vs. Convection
“With 80% of the heat going straight into the board,it would be easy to assume that we could use a conduction only solution with specified (correlation based?) heat transfer coefficients.
If anything, the opposite is true. Any heat that gets into the board must leave somehow. The only two ways for this heat to leave are conduction through the fabric of the board and convection and radiation from the board surface. We still need to look carefully at convective effects! “
This statement along with your blog comment (“Even more striking, the thermal density of high-performance communications equipment is projected to rise to the top of the list of heat producers over the next several years!”), addresses the current and future concerns with heat dissipation problems.
Assuming that under current interconnection technologies that 80% of heat does go straight into the circuit board, it would appear to more efficient to dissipate the heat through conduction at the source - the printed circuit assembly - and substantialy reduce the load on convective cooling. In effect reduce the energy needed to cool data centers, servers, etc.
If the circuit board assembly could be manufactured that allows for conductive cooling using direct fluid heat exchange techniques (not heat tubes), heat dissipation could be improved dramatically. This would be possible by making the “fabric of the board” heat conductive and have direct heat dissipation channels to the fluid cooled heat exchanger. Source conductive cooling would be more efficient than room cooling, refrigerated server cabinets, etc.
This project addresses the concept of conductive cooling at the source. I will be glad to submit design and product/process detail at your request. There is a patent granted, one pending and one in preparation. After reviewing Intel’s legal information and private policy terms (I am not an Intel employee) I was not sure how much information I could include in this communication.
A. J. Cucinotta