Mike Manos is Back, Blogs about his Microsoft Exit to Digital Realty

Mike Manos is back blogging on Loosebolts, and starts off his first discussing his decision to exit Microsoft.

The first question of course is - Why did I leave Microsoft for Digital Realty Trust?

First we need to pull our heads out of the tactical world of data centers and look at the larger emerging landscape in which data centers sit.  Microsoft, along with Google, Amazon and a few others are taking aim at Cloud Computing and are designing, building, and operating a different kind of infrastructure with different kinds of requirements. Specifically building ubiquitous services around the globe.  In my previous role, I was tasked with thinking about and building this unique infrastructure in concert with hundreds of development groups taking aim at building a core set of services for the cloud.   A wonderful blend of application and infrastructure.  Its a great thing.  But as my personal thought processes matured and deepened on this topic flavored with what I was seeing as emerging trends in business, technology and data center requirements I had a personal epiphany.  The concept of large monolithic clouds ruling the Information-sphere was not really complete.  Don’t get me wrong, they will play a large and significant role in how we compute tomorrow, but instead of an oligarchy of the few, I realized that enterprise data centers are here to stay and additionally we are likely to see an explosion of different cloud types are on the horizon.

The problem Mike is tackling is the role of data centers as information utilities.

In my opinion it is here in this new emerging space where the Information Utility will ultimately be born, defined, and true innovation in our industry (data center-wise) will take place.   This may seem rather unintuitive given the significant investments being made by the big cloud players but it is really not.   We have to remember that today, any technology must sate basic key requirements.  First and foremost amongst these is that it must solve the particular business problems.  Technology for technology sake will never result in significant adoption and the big players are working to perfect platforms that will work across a predominance of applications being specifically developed for their infrastructure.   In effect they are solving for their issues.  Issues that most of those looking to leverage cloud or shared compute will not necessarily match in either scale or standardization of server and IT environments.    There will definitely be great advances in technology, process, and a host of other areas, as a result of this work, but their leveragability is ultimately minimized as their environments, while they look like each other’s, will not easily map into the enterprise, near-enterprise, or near-cloud space.   The NASA space program has had thousands of great solutions, and some of them have been commercialized for the greater good.  I see similar things happening in the data center space.  Not everyone can get sub 1.3 Average PUE numbers, but they can definitely use those learnings to better their own efficiency in some  way.  While these large platforms in conjunction with enterprise data centers will provide key and required services, the innovation and primary requirement drivers in the future will come from the channel.

Why Digital Realty?

In Digital Realty Trust I found the great qualities I was looking for in any company.   First, they are positioned to provide either “information substation” or “enterprise” solutions and will need to solve for both.  They are effectively right in the middle of solving these issues and they are big enough to have a dramatic impact on the industry.  Secondly, and perhaps more importantly, they have a passionate, forward looking management team whom I have interacted with in the Industry for quite some time.

…

Another area where there is significant alignment in terms of my own personal beliefs and those of Digital Realty Trust is around speed of execution and bringing capacity online just in time.   Its no secret that I have been an active advocate of moving from big build and construction to a just in time production model.

I think we are going to see more of Mike Manos out there now that he is at Digital Realty Trust.

In the end, my belief is that it will be companies like Digital Realty Trust at the spearhead of driving the design, physical technology application and requirements for the global Information Utility infrastructure.  They will clearly be situated the closest to those changing requirements for the largest amount of affected groups.  It is going to be a huge challenge. A challenge, I for one am extremely excited about and can’t wait to dig in and get started.

As much as there is curiosity about Google, Microsoft, and Amazon data centers, these companies will only share what passes internal approval processes.  I challenge any internal group to limit what Mike can talk about. Mike knows part of being innovative comes risk, and he took a risk leaving a nice safe Microsoft position building data centers for its own needs, to building data centers as information utility infrastructure.

We need more people like Mike out there talking about changes in the industry.

Add Loosebolts to your rss readers as i am sure Mike will be discussing all kinds of things. http://loosebolts.wordpress.com/feed/

Read more

Data Centers as Cities, and Servers as Buildings – IT ecosystem

I just had an hour discussion with data center construction company and another hour discussion with Intel on the 5500 chip and data centers. Thanks to these two conversations and some good input from Intel’s Allyson Klein, Director of Server Technology Leadership Marketing, I came up with the following metaphor.

You could think of data centers as cities and buildings as servers.  Allyson was quick to point how Portland which is in Intel’s backyard of Beaverton where she is based, has taken a sustainable approach to the city.

Welcome to the new Bureau of Planning and Sustainability

In January 2009, Portland City Council merged the Bureau of Planning with the Office of Sustainable Development to create the new Bureau of Planning and Sustainability. The Bureau of Planning has had an exceptional record of guiding Portland's growth and development toward the thriving, livable city that it is today. The Office of Sustainable Development pioneered policies and programs that integrate environmental, economic, and social benefits.

This new bureau will ensure that sustainability principles are integrated into the core of Portland's planning, urban design and government operations, strengthening Portland’s position as the global epicenter of sustainable practices and commerce.

If you have a Sustainable City plan, then green buildings fit as well.

Green Jobs High performance green buildings Healthy environment

City of Portland Proposed High Performance Green Building Policy

High performance green building presents one of the best solutions to improve environmental performance while strengthening the local economy and keeping buildings affordable in the long term. Recognizing the many benefits of green building, in 2007, Portland City Council directed the Office of Sustainable Development to develop policy options to improve the environmental performance of commercial and residential buildings community-wide. The resulting proposed High Performance Green Building Policy also addresses City Council's goal to identify steps to reduce greenhouse gas emissions 80 percent from 1990 levels by 2050.

The proposed High Performance Green Building Policy seeks to accomplish the following goals for buildings and the sites they occupy in the City of Portland:

  • Reduce greenhouse gas emissions that cause climate change.
  • Maximize energy efficiency and cost savings.
  • Keep housing and commercial buildings affordable over time.
  • Decrease consumption of potable water, especially during summer months.
  • Increase on-site stormwater management.
  • Reduce waste during construction and operation.
  • Improve indoor environmental quality, occupant health and productivity.
  • Increase the number of local living-wage jobs.

Take these ideas and apply the metaphor to a green data center plan with the right kind of servers (performance per watt and power management).

After talking to Allyson more it became clear how Intel could fund its data center innovation efforts like Data Center Efficiency Challenge, as Intel has encountered situations where customers want to buy and deploy Intel’s latest products, but power and cooling capacities were the limiting factor. In 40% of the scenarios when customers postponed purchasing, data center power and cooling capacity was a limiting factor. Helping customers improve the energy efficiency of their data centers frees up power and cooling for more equipment. 

Intel is amongst a crowded data center services market, but they have a unique position as where others can buy or build buildings (server OEMs), Intel can make servers be more productive and more efficient with processors and support chips.  The Intel 5500 chip has some new power management features, and added instrumentation and controls to allow higher levels of power management.  Intel’s Data Center Manager is Intel’s latest effort.

Flexible data center hierarchy support

  • Supports management simultaneously at all levels of data center hierarchy

Power and thermal data aggregation
and trending

  • Monitor node power and inlet temperature data in real time
  • Aggregate power and inlet temperature data
  • Stores trend data for up to 1 year

Intelligent group power capping

  • Supports multiple policies depending on user power threshold target or goal to minimize power consumption
  • Maintains group power capping while dynamically adapting to changing server loads
  • Accepts SLA priority as policy directive
  • Automatically manage rack and group power consumption and safeguard from sudden power spikes

If you follow the data center/city and server/building metaphor, then DCM creates neighborhoods to control power use.

I like the city, neighborhood, building metaphor for data centers.  It seems like a good way to think about the problems in greening a data center and the interdependencies.

Read more

Left Microsoft 3 years ago, and not looking back

Today is May 1, 2009, and 3 years ago after 14 years at Microsoft I decided it was time to leave.  Having spent my whole working career at big companies (HP 5 yrs and Apple 7yrs), it was a big leap to be on my own.

But after 3 years I don’t regret the move at all.  I still keep in touch with many Microsoft people who are working on Green IT/Data Centers, and have met many more people who want to solve sustainability IT problems.  In 15 min, I have a series of conference calls with a construction company and then intel to discuss its Data Center Efficiency Challenge.

The weather is nice today in Seattle, and I am going to work my dock for as long as the laptop battery will last. Wifi to my house gives me high speed internet, so I can even write this blog entry.  The only problem I have is if a sea plane takes off, and I won’t be able to talk on my cel phone.

IMG_0815_edited-1

I am lucky to talk and work with some of the brightest people in the industry who want to build and operate green data centers.

And, my blog are my notes for what I discover and part of my research on topics.

Thanks for reading my blog.

-Dave Ohara

Read more

Power Consumption of Information Services is Apple’s Future Competitive Advantage

There is a lot of news out there about Apple hiring chip designers.  The WSJ.com has their version.

In Major Shift, Apple Builds Its Own Team to Design Chips

By YUKARI IWATANI KANE and DON CLARK

Apple Inc. is building a significant capability to design its own computer chips, a strategy shift that the company hopes will create exclusive features for its gadgets and shield Apple's work from rivals.

The Silicon Valley trend-setter has been hiring people from many different segments of the semiconductor industry, including engineers to create multifunction chips that are used in cellphones to run software and carry out other chores.

Apple could use the internally developed chips to sharply reduce the power consumption of its hit iPhone and iPod touch devices, and possibly add graphics circuitry to help its hardware play realistic game software and high-definition videos, people familiar with its plans say.

There is WSJ blog with more specifics on the chip designers.

There’s been buzz about Apple’s interest in microprocessor designers ever since the company bought the Silicon Valley startup P.A. Semi last year. But there’s ample evidence that the company’s hiring of chip-heads started much earlier, and is continuing. The question: what is Apple going to do with these guys?

One goal of CEO Steve Jobs is pretty clear–developing powerful, energy-efficient microchips for its iPhone, iPod touch and other future devices. Postings by Apple’s recent hires on the Web site LinkedIn include plenty of people who previously worked at cellphone chip makers, such as Texas Instruments and Qualcomm. There are plenty of references to power reduction in Apple’s job postings–spanning both hardware and software technologies–on its Web site and those gathered by search sites such as Indeed.com. Typical language: “The Iphone software team is looking for great engineers to help us achieve our ambitious battery life goals for the iPhone and iPod touch.”

The Apple staff has reached the point where software and hardware need to work better together to achieve higher energy efficiencies. Power consumption is one of the biggest differentiators now, and I would bet Apple knows if it shares its hardware innovations on the open market, their competitors will be able to reverse engineer how to reduce power consumption.

Steve Jobs being the ultimate showman wants to tell how long you can use his devices vs. the competition before you plug it in.

The only competitor who has the resources to compete against Apple is Microsoft, but given their monopoly status Microsoft has a hard time creating proprietary closed solutions.  Whereas Apple can hire the best and brightest chip designers to drive new power consumption performance.

One engineer I remember from my Apple days is Mike Dhuey.  Here is what Mike’s been up to.

Focus on alumni:
Michael Dhuey

image

When he was a boy growing up in the ’70s, Michael Dhuey had an idea that computers could be used for more than crunching numbers and playing games.

Little did he know that he would help create tools that would change the way people live.

Dhuey has been at the forefront of several computer breakthroughs. During his 25 years at Apple Computing, he worked on the Lisa Project, which laid the foundation for the Macintosh. He also developed hardware for the little tool that’s become omnipresent: the iPod.

Now a technical lead at Cisco Systems, he’s a big part of the team that has produced TelePresence, a videoconferencing system that’s like a sci-fi dream from the ’60s. He was a finalist for Design News magazine’s Engineer of the Year in 2006 and 2007.

…

“That’s one thing I always loved about Apple,” Dhuey says. “The company is willing to spend more on research and development than many places are. A lot of the projects fail, but there is always something of value that comes out of them.

“And in many ways, the Apple of that time was the last refuge of the true computer designer. Hardware and software—we invented it all.”

 

To be more efficient requires thinking about how the software and hardware work together.  A performance per watt gets you thinking down this path.

Apple is thinking this way in consumer devices.  Who is thinking this way in the enterprise data centers?

Read more

Microsoft’s Daniel Costello, Engineering Approach to Solve Data Center Design

Microsoft’s Daniel Costello has a good post on an engineering approach to solve data center business problems.

Before I get into what Daniel wrote, let’s contrast what typically happens in a data center project.

1) Collect the stakeholder requirements for the data center.

2) Forecast capacity requirements to determine how big the data center needs to be to meet the requirements.

3) Sell the project internally. Data centers are mission critical and meeting the requirements is #1 priority.

4) Bid out the project to industry experts.

Now let’s look at  Daniel’s steps.

1) Time to Market

2) Cost

3) Efficiency

4) Flexibility and Density

And the goals of the Microsoft team.

The Goals our Engineering Team Set

· Reduce time-to-market and deliver the facility at the same time as the computing infrastructure

· Reduce capital cost per megawatt and reduce COGS per kilowatt per month by class

· Increase ROIC and minimize the up-front investment for data centers

· Differentiate reliability and redundancy by data center class and design the system to be flexible to accommodate any class of service in the same facility

· Drive data center efficiency up while lowering PUE, water usage, and overall TCOE

· Develop a solution to accept multiple levels of density and form factors, such as racks, skids, or containers

Why take this approach?

Perhaps most importantly, with Generation 4 we can quickly add capacity incrementally in response to demand. Gone are the days when we had to wait 12-18 months for a large data center to be built, only to use a small portion of its capacity while we waited for demand to catch up to capacity. In short, our Generation 4 design delivers a revolution in terms of time to market that the data center industry has never seen before.

Read more