The Data Center in Real Time: Monitoring Tools Overview & Demo


Published on

Real-time monitoring provides unprecedented visibility into power and cooling infrastructure. Through the technology, data center professionals can baseline energy use, diagnose inefficiencies, and take corrective actions to improve performance and cut costs.

To view the recorded webinar, which includes a demonstration of a Real-Time Monitoring solution, please visit

Published in: Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • And today our objective will be to introduce and explain some of the data center measurement solutions in the marketplace. Measurement and monitoring are intriguing topics in the data center arena that are garnering a lot of attention. These tools represent a movement to proactively managing and improving the efficiency and performance of a data center. We are excited to present this to you this afternoon.
  • Our agenda is rather simple. We will first discuss what’s driving the need for monitoring. We will then take a look at the 2 most common types of monitoring in the data center, the point in time and the real time approach. We will then expound on the virtues of utilizing a real time solution and follow that with a demo of one of the available real time solutions in the market place. As always, we will close with a Q&A session, so without further ado, let’s jump into the presentation…..
  • Thisslide lists survey results for a group of enterprise data center owners/operators. This poll was released in Spring of 2009. As you can see, when these data center professionals were asked to comment on their concerns, Efficiency and Monitoring were at the top of the list.
  • This same group is polled twice a year. The results of the 2nd survey note that availability returns to it’s place as the number one concern for data center professionals. In 2009, moving into 2010, data center professionals remain focused on uptime, while trying to balance their availability concerns with efficiency and monitoring mandates.
  • Sothe questions needs to be asked, why do we need to monitor? I think it is captured best in these 5 bullets. Monitoring addresses the major challenges facing data centers from an overall $$ perspective. The cost of power continues to rise, and more and more facilities are running out of power. There are extreme costs associated with inefficiencies in the data center whether they be conversion losses in the power stream or issues concerning poor cooling strategies and the resultant by-pass airflow. These issues can lead to hotspots and failures and eventually downtime. Monitoring solutions help bridge the gap between IT and facilities, giving both groups the visibility they need to make informed decisions. And obviously monitoring tools assist in keeping SLAs.What is their perceived value?While Downtime is a word no DC Manager wants to hear, the cost and availability of power has become a key issue across the board. Many DC managers are realizing that cost and availability of power goes hand in hand with inefficiencies. The easiest way to remediate these inefficiencies is to control the DC in a holistic manner. As for SLA’s, anyone in our audience can tell you that not meeting your SLA requirements can become very costly to any organization through lost revenue and confidence in the infrastructure. Now we’ll take a look at a couple of monitoring approaches.
  • We mentioned 2 approaches to monitoring, the point in time and the real time. We will take a look at both approaches, and stack them against each other
  • One of the most common names for the point in time approach is a sneaker net. The name, while humorous, explains the approach. Basically an engineer is commissioned to investigate the data center and collect data. They are typically looking at critical pieces of infrastructure like the UPS and cooling systems, taking readings off a building management system if there is one in place. This data is usually compiled on a spreadsheet and used to in some very rudimentary calculations. When a 3rd party entity is brought in to provide this type of work a report is generated with results and recommendations to improve efficiency in the data center.What should we be looking at when taking this approach?In order for a walk through assessment to be of value you should really take a look at all aspects of the DC. This includes everything from the raised floor construction, the components of the actual infrastructure from cooling, power and power distribution all the way up to lighting and fire suppression. Once this data is collected you need to format the data into understandable deliverables that an end user can then act upon. Lets take a look at just few key Infrastructure components.
  • While it does not appear to be a comprehensive, holistic view into the data center, there are some key benefits to taking the sneaker net approach
  • These assessments are very cost effective and simple to execute with little advanced planning. There is no issue with interrupting operations as it is a non-invasive approach. At the heart of the results, we do have the ability to provide a level of risk mitigation with regular measurement.Below you’ll see a couple of findings from an assessment completed by one of 42U’s engineer’s.Talk to grounding – integrity of the electrical ground scheme, CRAC efficiency and PUE as areas that need to be investigated for remediation. “While great information is gathered in a DC walk through with various measurement tools, it is simply a point in time. It does not take a holistic view and makes it very difficult to consolidate the information into a user friendly process of remediation. Multiple walk throughs must be done to ensure any changes are effective and if any changes have had a negative effect”
  • Theshortcomings of this approach are obvious. At the top of the list it is simply a snapshot. A sneaker net does not take into consideration the ebbs and flows of a business. We know there are peak processing times that are not accounted for in a simple snapshot. It also requires manpower. In the last year we’ve seen a drastic reduction in available resources. DC Professionals just don’t have the time to take an engineer of a project to do the manual check.There are issues surrounding the various systems in the data center and making them speak the same language.And of course, what do you do with the data? How is it reported, tended, and aggregated? Who owns the process and how often do they do it?
  • All these questions lead one to ask, is there a better way. Let’s turn our attention to that better way, real time monitoring.
  • What you see here is the power of visualization. Monitoring solutions give you the ability to track conditions in the data center that have been elusive in the past. Air flow throughout the entire room as well as humidity. Power consumption from the panel all the way down to the outlet. We can get granular visuals down to the chiller level, watching flow if we need to
  • What you see here is the power of visualization. Monitoring solutions give you the ability to track conditions in the data center that have been elusive in the past. Air flow throughout the entire room as well as humidity. Power consumption from the panel all the way down to the outlet. We can get granular visuals down to the chiller level, watching flow if we need to
  • This is where real time monitoring begins to differentiate from the sneaker net. We are now giving you real time information that covers:Availability- And what we are really talking about here is Risk Prevention. Is my data center operating within industry standardsEfficiency- Gives me the data I need to establish a baseline and track my improvements. These improvements = Lower Operational Costs (we can actually save money)And finally Monitoring- which essentially brings availability and efficiency together through the power of Visibility
  • As you can see from this slide there can be an immense amount of information available depending on manufacturer and models of equipment. What we need to do to take a holistic view is to consolidate this information through wireless, wired, ModBus, BacNet or any other means available and comprise this information into a user friendly, meaningful format.
  • Here we have an actual screen shot of one of the many live thermal images available to the end user. You can see that we can instantly recognize areas of concern, in this shot it happens to be thermal strata at the tops of the racks or cabinets within this zone. With a simple click of the mouse we can pull up multiple maps and even animate the screen shots over a user defined period of time, which will give the user a live historical visual representation of the activity of a particular map. Now lets take a look at shot of some empirical data and how it relates to the real time monitoring solution for a group of air handlers.
  • I really like this information for a couple of reasons. You see that we have name plate values for sensible cooling, or the amount of cooling the manufacturer has spec’dthese air handlers. By monitoring the return and supply side of the actual coil we are able measure the Delta T across the coil and thus calculate the overall efficiency of each unit. As you can see from this information these units are running at a very inefficient level. These are prime candidates for some investigation and remediation to not only increase their efficiency but also re-coup a large amount of energy being wasted. Now let’s take a look at an actual Energy node installation.
  • The Data Center in Real Time: Monitoring Tools Overview & Demo

    1. 1. © 2009 42U All rights reserved<br />1<br />The Data Center in Real Time: Monitoring Tools Overview & Demo <br />Presenters:<br />Daniel Skrove – Sales Engineer<br />Steve Lewis – Director of Sales<br />
    2. 2. © 2010 42U All rights reserved<br />2<br />Agenda<br />The Need for Monitoring<br />Monitoring Approaches<br />Point in Time<br />Real Time<br />The Value of “Real Time”<br />Live Demo<br />Questions<br />
    3. 3. © 2010 42U All rights reserved<br />3<br />Emerson Data Center Users Group Survey- Spring 2009<br />“Forty-seven percent of respondents listed energy efficiency as one of their top facilities and network concerns, while 46 percent listed adequate monitoring in their top three concerns.”.<br />
    4. 4. © 2010 42U All rights reserved<br />4<br />Emerson Data Center Users Group Survey- Fall 2009<br />“[Availability] jumped from the fourth most important concern just six months earlier to the number one concern.”<br />
    5. 5. The Value of Monitoring<br />COST OF POWER…………..$$$ <br />COST OF INEFFICIENCY….$$$<br />COST OF DOWNTIME……..$$$$<br />OWNERSHIP……………......IT or FACILITIES<br />SLAs…..<br />Can you afford not to? <br />© 2010 42U All rights reserved<br />5<br />
    6. 6. Monitoring Approaches<br />Point in Time Vs Real Time <br />© 2010 42U All rights reserved<br />6<br />
    7. 7. Sneaker Net <br />Point in Time Measurements<br />One day investigation<br />Readings from critical components: UPS, Panels, HVAC, BMS<br />Clipboard readings entered into spreadsheets<br />3rd party: report & recommendations<br />© 2010 42U All rights reserved<br />7<br />
    8. 8. What to Record?<br />© 2010 42U All rights reserved<br />8<br />Temperature<br />Humidity<br />Fan Speed<br />Alarms<br />Setpointvs Production<br />HVAC<br />Volts<br />Amps<br />Battery Operation<br />Power Factor<br />KW(A) in / KW(A) out<br />Efficiency<br /> UPS<br />Depending on the model:<br />Volts<br />Amps<br />Per Whip values<br /> PDU/RPP<br />Depending on model:<br />Amps per PDU<br />Amps per Socket<br />Amps per Circuit<br />Rack PDU<br />
    9. 9. Sneaker Net Benefits<br />© 2010 42U All rights reserved<br />9<br />42U Point In Time Assessment Findings:<br />“Raised access floor has no supplemental grounding system. The raised access floor should have a grounding grid wire for better high frequency noise and static electricity protection”<br />“The existing CRACs have been problematic… Cooling capacity far exceeds heat generation today but the reliability isn’t optimal “ <br />“(Customer’s) IT load is 95kW and the input power to the data center is 226.699 kW yielding a PUE of 2.386”<br />
    10. 10. Sneaker Net Cont’d<br />© 2010 42U All rights reserved<br />10<br />How do we maximize the data?<br />Is it reported, trended?<br />Aggregated?<br />Quality of Analysis?<br />
    11. 11. © 2010 42U All rights reserved<br />11<br />Is there a better way?<br />Containment Video: LiveImaging Showing LBNL Data Center Overheating Issue <br />
    12. 12. Real Time Measurement<br />© 2010 42U All rights reserved<br />12<br />Branch Circuit Monitoring<br />Real Time PUE<br />
    13. 13. More Real Time Tools<br />© 2010 42U All rights reserved<br />13<br />Temperature<br />Zone Alerts<br />Chiller<br />
    14. 14. Previous Survey Concerns<br />1. AVAILABILTY<br />Real time tools ensure operation within designated guidelines such as ASHRAE or other industry accepted parameters. <br />2. EFFICIENCY<br />Real time tools establish an operational baseline that can be historically trended for remediation control. <br />3. ADEQUATE MONITORING<br />Real time tools measure proactively, protecting & trending all facets of physical infrastructure. <br />© 2010 42U All rights reserved<br />14<br />
    15. 15. An Array of Data<br />© 2010 42U All rights reserved<br />15<br />Min, Max, Average HVAC supply temperatures <br />Min, Max, Average HVAC return temperatures <br />Min, Max, Average HVAC temperature delta <br />Air Loss Ratio<br />Bypass Air Ratio<br />Recirculation Air Ratio<br />Fan Energy<br />HVAC<br />UPS In<br />UPS Out<br />IT Load<br /> UPS<br />Current<br />Voltage<br />Power<br />IT Load<br /> PDU/RPP<br />Cold Aisle Temperature<br />Hot Aisle Temperature<br />Humidity<br />Thermal Maps<br />Rack<br />
    16. 16. Reporting to a Common Database<br />PDU<br />CRAH<br />Rack<br />Chilled Water Flow<br />Sub-floor<br />Disparate vendor products and protocols now speak the same language: efficiency.<br />16<br />
    17. 17. Real Time: Thermal Map<br />Other Maps Available<br />Temps below 55F<br />
    18. 18. Real Time: Thermal Report <br />© 2010 42U All rights reserved<br />18<br />Values based on user-defined time periods <br />
    19. 19. Real Time: Cooling Energy<br />GATEWAY<br />ENERGY METER<br />WIRELESSSENSOR<br />REAL TIME POWER VALUES FOR THIS PANEL <br />
    20. 20. Remediation- Next Steps<br />Cooling Infrastructure<br />Aisle Containment (Cold or Hot)? <br />Raising Temperature Setpoint? <br />Variable Frequency Drive Retrofit (if CHW)? <br />The framework is in place to visualize the effects and quantify the results. <br />© 2010 42U All rights reserved<br />20<br />
    21. 21. Real-Time Implementation<br />Data Center Survey<br />Best practices & ROI<br />Potential rebates<br />Executive managerial buy-in<br />Establish efficiency team & goals<br />Plan<br />Execute<br />Install instrumentation & software<br />Perform energy utilization audit<br />Calculate current benchmark <br />Measure<br />Data Collection<br />Environmental Monitoring<br />Analysis<br />Report<br />Recommendations<br />Reconfiguration plan<br />Efficiency improvements<br />Remediate<br />Low Hanging Fruit<br />Progressive Changes<br />Savings Reinvested<br />Ongoing Improvement<br />
    22. 22. Live Demo<br />© 2010 42U All rights reserved<br />22<br />
    23. 23. Takeaways<br />© 2010 42U All rights reserved<br />23<br />Meeting of industry thought leaders convened on January 13, 2010 “to agree on datacenter energy efficiency measurements, metrics, and reporting conventions.”<br />Results of the Meeting: <br />“Power Usage Effectiveness (PUE) using source energy consumption is the preferred energy efficiency metric for data centers.”<br />When calculating PUE, IT energy consumption should, at a minimum, be measured at the output of the uninterruptible power supply (UPS). However, the industry should progressively improve measurement capabilities over time so that measurement of IT energy consumption directly at the IT load (i.e., servers) becomes the common practice.”<br />SOURCE:<br />
    24. 24. The Future is Now<br />© 2010 42U All rights reserved<br />24<br />TheProgressive Improvement is available today: <br />Rack level power consumption<br />Ongoing PUE measurement<br />Informed decision making with non-invasive installation<br />Impactful at all data center sizes<br />Branch Circuit Monitoring<br />Real Time PUE<br />
    25. 25. Questions?<br />For a copy of today’s presentation, please email<br />For additional information on 42U’s products and services, please visit:<br /><br />Or Speak to your Solutions Consultant Today<br />
    26. 26. Dedicated to Improving Data Center Performance<br />Thank You!<br />