I had the pleasure of delivering a customer briefing for a power company. Needless to say, when I asked my standard question about what is more important - server density with higher power usage versus lower density and lower power usage, the answer was not surprising - for them, floor space is a bigger issue than getting additional power.
We had a great discussion which included the three areas of the Thermal Logic story: Reduce, Reclaim, and Extend.
Reduce is how we reduce the power consumption of systems. That includes things like more efficient power supplies, adaptive fans, processor power improvements, and more.
Reclaim then looks at the more efficient environment, and asks how we can free up stranded capacity. What I mean by that is that many data centers provision the power in their data center based on the face plate requirements of the server - maximum load at the maximum configuration. We all know that most servers don't use that much power ever, let alone on average. Through power calculators monitoring and management tools such as Insight Control power management, we can more accurately provision power to systems and free up the stranded, unused power capacity in the data center.
Third, we can extend the life of the data center. Technologies and services such as HP Data Center Environmental Edge, HP Modular Cooling, and HP Critical Facilities Services all help our customers get the most out of their existing data centers, or build new ones if needed.
The best part of the briefing isn't my presentation of HP's offerings, but the questions.
First, what is the adoption of power management technologies by customers today? So, the question for our readers - do you use reduce, reclaim, or extend types of technologies or solutions today? If so, which ones?
The second interesting question is what are the biggest power consumers in the data center? Is it network, servers, or storage? Where do we see improvements happening (my answer - all areas). So, once again, where do you think power is being consumed in your data center? I should mention that this is something that I know has been discussed in at least one of The Green Grid's work group calls.
Comments and feedback are always appreciated.
I finally got out to see the movie Avatar in 3D last week with a bunch of friends. I enjoyed the movie, and the animation and 3D effects were wonderful.
What is more interesting is the infrastructure behind the movie. A lot of the work was done on HP BL2x220c blades in HP Bladesystem c7000 enclosures. Imagine racks of them - or about 40,000 processors and 104 Tb of memory, according to the Reality Check - Server Insights blog posting at http://www.communities.hp.com/online/blogs/reality-check-server-insights/archive/2010/01/13/the-it-behind-the-big-blockbuster-sci-fi-hit-avatar.aspx.
This is a great example of scaling out to do a huge job. But, the nice thing is that the same infrastructure can also be used for HP Integrity BL860c and BL870c blades . Even the HP Non-Stop servers are available in a bladed architecture, and they have some of the highest levels of uptime in the industry.
Compared to rack-mount servers, our customers save floor space, power and cooling, and even management time and expenses compared to the save compute power in a rack-mount chassis. So, whether you are scaling out to render a new movie, running a mission critical database instance, or helping run a commodities exchange, a HP BladeSystem infrastructure can help run your business today.
While Kurt is off visiting customers and Lorraine is reflecting on the nature of mission-critical computing, I am also joining the blogging team. Let me introduce myself. My name is Jacob Van Ewyk, and I work in Business Critical Systems Marketing. I'm currently focused on Power, Cooling, and Cloud Computing. I spend some of my time on the HP IT Transformation story, and spent about 5 years working on server virtualization. Between these areas, I have had the opportunity to meet many of our great customers, partners, and sales reps around the world, hear some great stories of how technology is improving businesses and peoples’ lives, and get asked a lot of questions about the mission-critical business, HP, and the technology industry in general.
This week, I'm preparing for a training event. Needless to say, Green IT, and specifically Power and Cooling, will be a topic of discussion, and I'm looking forward to both sharing what I've discovered, as well as getting feedback on possible future directions.
The first topic that I'll be discussing is around server density versus power efficiency. Here is the scenario: we can build servers (such as blades) that pack more performance into a smaller space. Alternatively, we could perhaps use a little more space, but run the system at a higher temperature so that they use less power and cooling, but get the same computing power. Why is this important? ASHRAE (American Society of Heating, Refrigeration and Air-Conditioning Engineers - www.ashrae.org ) recently increased the maximum datacenter temperature recommendation to as high as 27°C (80.6 °F) . Imagine if servers could run at, say 45°C/113°F - would you even need cooling in your datacenter?
So, what do you think? Do you think server density is a higher priority than power efficiency, or vice versa? Or, are the both roughly equal priorities in the data center?
And this just came out: Newsweek rates Green companies, and HP ends up #1! http://greenrankings.newsweek.com/