Microsoft's Mike Manos comments on CADE metric from McKinsey and Uptime

Mike Manos posted his comments on the CADE metric presented by McKinsey and Uptime.

May 05

Struggling with CADE, McKinsey / Uptime Metric

I guess I should start out this post with the pre-emptive statement that as a key performance indicator I support the use of CADE or metrics that tie both facilities and IT into a single metric.  In fact we have used a similar metric internally at Microsoft.  But the fact is at the end of the day I believe that any such metrics must be useful and actionable.  Maybe its because I have to worry about Operations as well.  Maybe its because I don't think you roll the total complexity of running a facility with one metric.  In short, I don't think dictating yet another metric, especially one that doesn't lend itself to action, is helpful.

People were quoting Mike at Uptime to just start measuring something.  I want to add a correction that given Mike’s experience he would never randomly measure something which is what other presenters were suggesting to take action.  Mike knows what he wants is effective measurements, and knows whether he is picking a measurement that is useful.

The recommendation should be modified  to “pick something to measure that you think is useful in the long term, start anywhere you want, pick up a clipboard.”

Measurements need to be thought as part of a closed loop feedback system where the measurements are indicators of how you are meeting operation service goals and whether modifications you are performing are effective.

The presenter at Uptime for the McKinsey study joked that he had a process that had you running in circles, and CADE would probably do that.  As you keep on measuring CADE, it will have you running in circles, chasing what you need to do to make it better. The CADE numbers don’t have the expected result as Mike also points out some flaws.

  • As you cull out dead servers in your environment, your utilization will drop accordingly and as a result the metric will remain unchanged.  The components of CADE are not independent. Dead servers are removed so that Average server utilization goes up then Data Center Utilization goes down showing proportionally so there is no change and if anything PUE goes up which means the metric may actually go up. Keep in mind that all results are good when kept in context of one another.
  • Hosting Providers like Savvis, Equinix, Dupont Fabros, Digital Realty Trust, and the army of others will be exempt from participating.  They will need to report back of house numbers to the their  customers (effectively PUE).    They do not have access to their customers server information It seems to me that CADE reporting in hosted environments will be difficult if not impossible.  As the design of their facilities will need to play a large part of the calculation this makes effective tracking difficult.  Additionally, overall utilization will be measured at what level?
  • If hosters exempted, then it gives CADE a very limited application or shelf-life.  You have to own the whole problem for it to be effective.   
  • As I mentioned, I think CADE has strong possibilities for those firms who own their entire stack.   But most of the datacenters in the world would probably not fall into "all-in" scenario bucket.

    • This was my first Uptime Institute, and I have dozens of observations, I'll be making over the next couple of weeks.

    As McKinsey and Uptime also promoted the idea of the energy czar. I disagree with the idea, and bounced a better idea off of multiple people at the conference that energy metrics should be integrated as part of capacity planning  and reporting. Data Center staff and executives need to get used to seeing power numbers along with other key performance indicators for their data centers.

    The last thing you want in the data center is to have an energy czar/nazi who measures their performance on energy savings who does not appreciate the relationships of efficiency projects to overall operations.

    So, even though it was a nice effort to introduce the idea of a data center efficiency metric, it will not work as I first thought. The data center efficiency metric makes more sense to be developed by a group like the The Green Grid where the consensus driven model for publishing content will make sure a metric has industry support before publishing.

    Read more

    Network Vendors compete for Green Data Center Market at Interop

    Whenever I talk about Green efforts in the data center, I try talking to the network guys, but the response is usually the power costs for our network are not significant. Well at Interop Green IT has arrived with Nortel pushing the competition.

    ComputerWorld reports on Interop and Green IT's Appeal.

    May 2, 2008 (Computerworld) LAS VEGAS -- Green IT was a major theme at Interop this week, as vendors outbid one another on how well their networking gear could lower power consumption to help save the planet.

    Some of the green IT claims were not-so-veiled attacks on the biggest switching and routing vendor, Cisco Systems Inc. In response to comments from competitors, Marie Hattar, Cisco's vice president of network systems and security, demurred in an interview, "We didn't want to go with the 'greenwashing' treatment."

    Hattar's comment aside, Nortel Networks Corp. cited a January study by The Tolly Group Inc. that said some of Nortel's networking gear uses 50% less power than Cisco gear. Joel Hackney, president of enterprise solutions at Nortel, said in an interview that Tolly and other studies have shown Nortel data center gear, in addition to being more energy-efficient, can offer 20 times the performance and seven times the resiliency (meaning submilliseconds of recovery time from an outage) of Cisco's and other competing products.

    The market even has vendors developing network measurement tools.

    The assessment of green IT was picked up by at least one other vendor, Calabasas, Calif.-based Ixia, which demonstrated a proof-of-concept testing system for measuring power efficiency in relation to network and application load. Areg Alimian, technical product manager at Ixia, said tools are lacking for making such measurements. The company is providing the technology used by Miercom in its Certified Green testing process.

    Cisco believes in the Green Messaging as well.  See the Al Gore and John Chambers webcast.

    Al Gore, John Chambers and Sue Bostrom will come together in a virtual unified communications environment before a live audience of technology leaders to examine the critical role that innovation can play in mitigating climate change. The virtual discussion will also be streamed live via a webcast session for general viewing.
    This event marks a technological breakthrough showcasing a new way of communicating and collaborating in which the panelists will appear as if they are in the same room even though they will be thousands of miles apart. This unique virtual discussion will take place via the Cisco TelePresence system with John Chambers, Al Gore and Sue Bostrom addressing live audiences simultaneously in London, England, Orlando, Fla.,and several other locations around the world.

    Read more

    Why I Named my Company Green M3, What is the Green M3 Method

    One of the interesting things I learned at Uptime Institute is the timing was right to explain why I named my company Green M3.  I told the story of how I started the company in response to a WSJ article. And, next let me tell you why Green M3.

    I have as my tag line Monitoring, Metering, Managing ...

    But, these 3 words starting with M where created after I figured I wanted to call my company Green M3.

    Nicholas Carr's post triggered my own post on containers.

    UPDATE: There's more on the data center philosophy of the 'Soft Boys in this interview with Manos and one of his colleagues. On the attraction of containers: "One of the things we like about them is we can take a bunch of servers and look at the output of that box and look at the power it draws. At the end of the day, we can determine, 'What is the IT productivity of that unit? How many search queries were executed per box? How many emails sent or stored?' You can get into some really interesting metrics. A lot of people say you can't look at the productivity of a data center, but if you compartmentalize it - not as small as the server level, but at some chunk in between - you can measure productivity."

    Now that people are talking about containers and thinking down the line of evaluating performance per the container, it makes it easier for me to explain Green M3.  I tested this on a bunch of people, so let's put in down in writing.

    In coming up with a method to evaluate Green performance in the data center, I needed a way to think and have the right mindset in looking at the problem. Then I hit upon abstracting to what it is you wanted, and computing needs to be measured in a universal way.  Processor technology, RAM, HD, network, various I/O technologies confuse the issue, and cloud the ability to evaluate things side by side. Computing is delivered in a box - 1U, 2U, 4U, Rack, but side by side comparison's are still hard as solutions required multiple servers.

    How do you evaluate the IT solution implementation in one technology vs. another?

    Then I started to think of things as the space it takes to encompass the solution. Characterize the space in abstract terms - KW required, Cooling required, network connectivity, compute power, and other TCO factors. Define the space, how big it is, what it does, and what it costs.

    What is a universal unit of measure? The cubic meter. The M3 is meant to communicate cubic meter = meter x meter x meter, not Monitoring, Metering, and Managing. But, this is too hard to put in a tag line. 

    A rack is 1.2 cubic meters.

    A 40 foot container is 67.5 cubic meters.

    A 20 foot container is 33 cubic meters.

    This gave me the right framework, architecture method to evaluate different implementations.

    If you apply the Green M3 method you can calculate the cubic meters of white floor space and compare it to the cubic meters required in a container, and associate a cost and performance per cubic meter. People tend to think of square meters/feet. But, to be efficient you need think about the metrics per cubic meters.

    The good thing is I named my company back in Nov 2007, before containers became popular, and Christian explained to Nicholas Carr a benefit of containers. Also as background, part of what helped me understand this method is working as a distribution logistics engineer 20 years ago for HP and Apple where I learned to think in terms of boxes, pallets, and shipping containers. I remember working at Apple, when the Mac Plus was being introduced, and people were getting all excited about the Mac Plus vs. the Mac 512k. But, in distribution I told people it is just another box, it is no bigger, doesn't weigh anymore. It doesn't make any difference to us, other than we'll ship more of them.

    When you abstract a technology how does it effect the cubic meter of data center computing?

    I'll work on more on documenting the Green M3 Method, but here is the start.

    Read more

    Digital Realty Trust Adopts PUE as Standard Metric for Green Data Center

    Digital Realty Trusts has taken a leadership position to publish PUE for its datacenter facilities. PUE enables customers to evaluate how efficient a facility is versus theirs.

    Digital Realty Trust Publishes Industry's First Data on Datacenter Energy Efficiency

    Company Adopts PUE Metric as Standard for Measuring Datacenter Efficiency

    SAN FRANCISCO, May 1 /PRNewswire-FirstCall/ - Digital Realty Trust, Inc. (NYSE: DLR), a leading owner and manager of corporate and Internet gateway datacenters, announced today that it has published the industry's first energy efficiency data about its datacenter facilities. Digital Realty Trust will use the Power Usage Effectiveness (PUE) metric as the methodology for measuring and reporting energy efficiency in its portfolio of facilities in North America and Europe. Digital Realty Trust is the first company to provide customers with detailed and actionable information about how datacenter operations are meeting their corporate green IT objectives. The Company will also publish benchmarks that will support industry-wide initiatives to make datacenters greener and reduce expenditures on energy.

    PUE is an emerging standard promoted by The Green Grid and others in the datacenter industry to provide a simple and consistent method of measuring the ratio of power delivered to IT equipment to the total amount of power used by the datacenter facility. PUE provides reliable information about the energy efficiency of datacenter facilities by calculating how much power is devoted to driving the actual computing/IT components (servers, for example) versus the ancillary support elements such as cooling and lighting.

    This follows through with their Green Datacenter Vision.

    Our Green Datacenter Vision

    Digital Realty Trust is committed to developing practical standards that reduce datacenter power requirements and negative by-products while increasing their operational effectiveness. Our datacenter facilities are designed to LEEDs (U.S.A.) and BREEAM (Europe) certification standards. Digital Realty Trust is a voting member of the Green Grid consortium.

    The interesting effect is how customers can now compare their own data centers vs other hosting companies as Jim Smith, VP of engineering states. Imagine a customer asking for the PUE for a hosting facility and the sales rep, says "PUE? What is that?"

    "Digital Realty Trust has been a leading voice on this issue since well before it gained the prominence it has today, and has achieved a number of important industry milestones related to green datacenters. However, I believe that reporting the PUE metric for our datacenter facilities is the most significant milestone yet because it will provide a model and benchmarks that can be used industry-wide to achieve dramatic increases in energy efficiency and can be used by customers to compare the energy efficiency of different datacenter facilities."

    It will be interesting to see what kind of numbers Digital Realty Trust publishes and whether they publish the range of PUE vs different times of the day and year as well as how load can effect PUE in their facility.

    Remember PUE is a dynamic number, not static, and is expected to fluctuate in a range.

    Read more

    Sun visits Rackable's Ice Cube, leaves Green with Envy

    At Uptime Institute Conference, Rackable Systems has their Ice Cube Container parked in the parking lot open for all to visit. While talking to the Rackable staff, it became clear how brilliant they were to have their container easily accessible.  Their traffic has been high, because they are open during the whole conference vs, the few hours the exhibit area is. This means they are open the whole conference to potential customer's visits with their undivided attention with no competing traffic. Rackable staff says this is one of their best shows for interest in containers from qualified customers.

    While I was in the Rackable container, Mark Monroe, Sun's Director of Sustainable Computing visited the container to get a little tour. I asked the Rackable guys how the conversation went. They joked, "he left green with envy."

    It's funny, and he may have.  I am blogging from the session on modular "containerized" computer room concepts with Sun, Rackable, Intel, and Verari participating in a panel discussion.  There is lively engagement discussing the issues of using containers. The IBM and Dell guys were in the audience as well.

    What is interesting to hear is all the different scenarios where people are thinking containers make sense for data center deployments.

    One particular interesting piece of data is Verari Systems said the Microsoft Container Data Center deployed in Boulder, CO is operating at a PUE of 1.3, and should be expected to go to 1.2.

    Read more