Data Centers: How DCIM Improves Your Daily Operations

Speakers

Martin Matse

Global Business Development Manager – Data Centers

ATS Global

Dereck Saunders

Application Engineering Manager

Inductive Automation

Christopher Markstein

Strategic Industry Account Executive

Inductive Automation

The public image of data centers is changing. Increasingly, there is a common perception that data centers use too much power, water, land, and renewable energy, and hardly create any local jobs.

Administrations around the globe are responding with new directives, regulations, etc., forcing data centers to optimize and report. Even though every country/region has its own rules, they still have a lot in common.

In this webinar, experts from Inductive Automation and ATS Global will look at those common requirements and present how an open data center infrastructure management (DCIM) solution based on Ignition can help you to comply, and maybe even change the public opinion about Data Centers in the long term. Don't miss it!

• See the new Ignition demo for data centers
• Learn about data centers’ public image and how governments are responding
• Explore specific and common demands written in the current and new directives
• Find out how DCIM, in combination with Balanced Scorecard, can help you to comply

Transcript:

00:01
Dereck Saunders: Alright, hello. Welcome everybody. I think we can go ahead and get started. So welcome to today's webinar, "Data Centers: How DCIM Improves Your Daily Operations." We're glad that you've all joined us here today. Today's presenters are Dereck Saunders. That's me. I'm the Application Engineering Manager at Inductive Automation. We also have Martin Matse on the webinar today. Martin is the Global Business Development Manager at ATS Global, which is a Premier Integrator with a lot of experience working with data centers. And we've also got Chris Markstein here and he's the Strategic Industry Account Executive at Inductive Automation. Martin and Chris, let's each take a minute to tell the audience a little bit more about ourselves. We'll start with Martin and then Chris and then me.

00:44
Martin Matse: Okay. Good morning, good afternoon, good evening everybody. So my name is indeed Martin Matse. I'm working for ATS Global, I'm out in the Netherlands office. And my role is business development for our data center activities. So looking at me personally, I started in this market in 2007 and now we have around 200 different data centers using our solution in the area. Thank you very much.

01:11
Christopher Markstein: Yeah, thanks Martin. Hi everyone. Good morning, good afternoon, good day. Chris Markstein here. I've recently joined Inductive Automation and my scope is to build and drive the growth plans for the verticals we've identified as being uniquely positioned to benefit from a strategic focus and investment. And I guess that's especially those that might be a bit out of the mainstream for Inductive Automation traditionally, the first of those industries and our focus here today is data centers. So I'm super excited to be here and looking forward to the discussion. Back to you, Dereck.

01:49
Dereck Saunders: Great, thanks guys. And I'm Dereck Saunders. I'm an Application Engineering Manager here at Inductive Automation. I co-manage the application engineering team, which is part of our Sales Engineering Division. My team works closely with sales engineers and the sales team to help develop Ignition solutions for customers as well as contribute resources to the Ignition Exchange. And we also develop and maintain Ignition demo projects. And in a few minutes here we'll take a look at our latest demo project, the Data Center Demo. So here's the agenda for our webinar today. We'll give you a quick intro to Inductive Automation and Ignition software. Then we'll introduce the new Data Center Online Demo and the upcoming Data Center Industry Pack. Next we'll learn about DCIM and how it supports your data center's operations and we'll end it with a Q&A session. If you have any questions as we go along here, just type them into the questions area of the GoToWebinar control panel and we'll answer as many of them as we can.

02:40
Dereck Saunders: If for some reason we're unable to get to your question today, please do reach out and one of our account representatives will be happy to help you answer those questions. Also, this webinar and the slides will be made available within the next few days. If you wanna go over it again, share it with somebody who wasn't able to make it, or just refer to any of the URLs that we share in the slides here. So what is Ignition? Here's a quick summary. Ignition is a universal industrial application platform for HMI, SCADA, MES, and IIoT. Ignition is used by 57% of the Fortune 100. It acts as a central hub for everything on the plant floor and beyond. You can use it to create pretty much any kind of industrial application. It is web-based, web-managed, and web-deployed. It has an unlimited licensing model and it's cross-platform.

03:23
Dereck Saunders: It offers industrial-strength security and stability, and it's trusted by thousands of companies worldwide. So we're gonna start out here by sharing some of the new resources for data centers that we've been working on here at Inductive Automation. In our online demo of Ignition, we have demo apps for several industries, but last month we actually added a demo app for the data center industry. So this is a new demo app. It's very realistic, it's a detailed view of what a real Ignition project for data centers could look like. The demo screens are fully functional, they're interactive, and they are completely mobile-responsive. The demo includes a variety of different screens, including a system overview and EPMS single line, a BMS heat map, alarms and trends, and an architecture screen as well. So this is the first iteration of this new demo project, but I'm excited to say that we plan to release further iterations of this demo, expanding on scale and features sometime in Q1 of next year. So for now, I'll turn it over to Chris and he'll take you through some of the specifics of our new Data Center Demo. Chris, take it away.

04:24
Christopher Markstein: Thanks Dereck. Starting with the overview here, and we have some site-level KPIs on the left, and in the main content pane, there's a diagram showing a section of a data center. This specific view focuses on electrical infrastructure and includes power quality meters, generators, UPSs, and PDUs. The red-colored lines designate the power is on or hot and that the equipment is energized and each piece of equipment is also clickable, which brings up a pop-up to show more detail and metrics for that specific item, including any active alarms. As Dereck mentioned, we expect in future releases of this demo, that we will scale out the application and this overview will likely depict multiple sections within a data hall or even multiple data halls as you might find in a hyperscale facility. The design concept here is to show key metrics and status at a glance without the extra complexity and detail of a single-line diagram or engineering schematic.

05:24
Christopher Markstein: But of course those detailed single-line diagrams can be really useful in certain cases, so we've included that in our demo as well. Let's take a look. So here we have our single-line diagram, as you would commonly find in an Electrical Power Monitoring System or EPMS. This includes additional data points that our at-a-glance overview doesn't, including breaker status, a more detailed electrical diagram. And this screen is built on an SVG image of the single-line diagram itself. On top of that, we've layered Ignition components to show real-time status for breakers and equipment. As you can see, this page leverages pan zoom capabilities. These diagrams and schematics can be quite large and intricate, so having the ability to zoom in and out, to move around the diagram while also showing real-time status of equipment is quite useful.

06:17
Christopher Markstein: You also have the ability here to switch to an interactive mode and you can click on equipment to bring up more detailed information in a pop-up just like on the overview page previously. The pan zoom capability isn't native functionality in Ignition just yet. Instead we leveraged a tool our application engineering team developed and we've made this pan zoom tool available to the public for free on the Ignition Exchange. So I'd encourage you to check it out. Dereck will give some more information on that in just a few minutes. Next, let's step over to the BMS heat map.

06:52
Christopher Markstein: So here we have a heat map showing temperature status for the hot and cold aisles in this section of the data hall. As typical of what you might find in the building management system, when the time-select bar below the aisles is green, the aisles and temperature sensors are depicting real-time values. However, a neat thing that we've added to this screen is the ability to play back historical data. Selecting an alarm from the table at the bottom of the screen will show the aisles and their temperature values at the time the selected alarm occurred. The time-select bar will change from green to blue. That indicates we're looking at historical values instead of real-time. And the temperature sensors and alarm will also highlight with a white border to make them immediately identifiable. You can also simply drag the time-select bar to any time of interest and see the historical values, then drag the bar back to the current time and you'll return to real-time display. We've also included some metrics, average hot and cold aisle temperatures in the bottom left. And in the future, we'll add additional sensors and metrics including humidity, differential pressure, rates of change, and other vital environmental data. Alright.

08:03
Christopher Markstein: Next, we'll take a look at alarm and trend screens. And these two screens are pretty straightforward. The alarm screen leverages Ignition's native Alarm Status and Alarm Journal components for viewing and managing alarms. The trend screen leverages our native Power Chart Component. And we're displaying some site-level KPIs at the moment here. But of course, any tags can be trended here ad hoc for further analysis, troubleshooting. And there's a great example of that in the BMS Demo app in our public web demo that I'd also encourage you to take a look at. And finally, here in our demo, let's take a look at the architecture screen. This view depicts the specific Ignition architecture for our Data Center Demo. And you can see two frontend servers hosting Perspective UI sessions connected to a backend server for all of the tags, the alarming, historization.

09:01
Christopher Markstein: You'll also notice that the backend has a redundant server as a standby. And that's in case, for any reason, the main backend server would temporarily go down. The backend server is also connected to a database to store history. It's connected to remote I/O via Ignition Edge gateways. And you can click, as you see, on any of these servers to then get more detailed information and metrics on the health and the status of both the Ignition gateway and the actual server that it's running on. That's a brief overview of the Data Center Demo. Excuse me. But we also wanted to mention that leveraging the power of Ignition's Perspective Module, the entire project is completely mobile responsive, as you can see here.

09:49
Christopher Markstein: You can try the demo for yourself, go to the URL on the screen, or simply go to demo.ia.io, click demo apps, and then data centers. We look forward to bringing you future iterations of this project, expanding on the scope, the features, and our next release is planned for Q1 of 2024. So stay tuned. And with that, I'll turn it back over to Dereck, who will talk briefly about something else coming in 2024, the Data Center Industry Pack.

10:23
Dereck Saunders: Thanks, Chris. As Chris just mentioned, something we expect to release sometime in the first half of 2024 is the Data Center Industry Pack. The Data Center Industry Pack will be one of several vertical packs that we develop over time, each focusing on a specific industry. Since we launched the Ignition Exchange back in 2019, it's been used for finding and sharing community-made resources to help you build Ignition projects faster and better. Now, with the launch of industry packs, we're expanding the use case of the Exchange to offer collections of resources that are designed for the needs of specific industries.

10:52
Dereck Saunders: The idea here is that Ignition as a platform gives you immense flexibility to build whatever you can dream up for your projects and solutions, but sometimes having prebuilt solutions can really help kickstart the development of a project or supplement an existing project. That's where the Exchange and soon industry packs comes in. The industry packs will provide several resources specific to an industry like water/wastewater, oil and gas, or in this case, data centers, in the form of Perspective views, UDT tags, pop-ups, scripting resources, anything like that that can be deployed in a project to help jumpstart new projects or supplement existing ones. Also as Chris mentioned earlier, the pan zoom tool that we leveraged in our Data Center Demo developed by Application Engineer Mike Bordyukov is available for download on the Ignition Exchange.

11:37
Dereck Saunders: It's called the Pan Zoom iFrame, and it can be found by browsing the Ignition Exchange or directly from the URL provided here. One last thing that I would like to mention is that our dev team is designing a prototype container for Perspective that will natively provide similar functionality in the future, but it's currently in early development and prototyping, so it's likely a ways out, it's probably [Ignition] 8.3 and beyond. So this Pan Zoom iFrame that we have in the Exchange is still a great solution for that functionality for now. I invite you to check out that resource and all of the great Ignition resources that we have available on the Ignition Exchange after this webinar. Now we're going to switch gears. Martin from ATS Global is going to talk about Data Center Infrastructure Management, or DCIM for short, and he'll explain how DCIM can support your operations and unlock potential for greater success. Martin, over to you.

12:23
Martin Matse: Well, thank you, Christopher. First... That's going fast. First, two slides about our company, ATS Global. We see ourselves as the leader in the smart Digital Transformation. If you look to the left side of the slides, you can see the areas in which we are active, so robots, controls, IoT, MES/MOM, cybersecurity, and data centers. In the middle, you see the markets in which we are active: automotive, aerospace, medical, pharma, government, food, and of course data centers, and we support our customers from cradle to cradle, so from the design phase, configuration phase, implementation phase, operation phase, maintenance phase, until the end of the process. We have offices in 26 countries. We are with I think at the moment even with 1,400 people. And if you look at this map, we are mainly based in APAC, EMEA, and the Northern part of America. This is about ATS. Okay, if you know about the image of data centers and about responses from legislators. So if you talk with people about your profession and they say, "Okay, I work in data centers." And then you have two types of responses and 30% of people will say, "Data centers. What's that?"

13:36
Martin Matse: No idea. And 70% will say, "Oh, data centers. I know that." Data centers are using so much power, and all the renewable energy is going to data centers. Just to give you an example, I live in the Netherlands and we built a wind farm at sea and almost all the power is going to a very big cloud provider. They use so much land, they use a lot of water, and to be honest, nobody works there. So there's only consuming our goods and doesn't bring anything. So that's quite a negative public image that data centers are having. I think legislators, administrations think more or less the same. And they're also adding all kinds of new rules, regulations, and directives for data centers they created.

14:21
Martin Matse: So if you look to the left part of the slide, you can see that every region creates its own directives and regulations. And if you look to APAC, they even create directives per country, which are actually all more or less the same, but still they have some differences. A very interesting one is the EN 50600. It's a European standard, but the nice part about that 50600 is it is quite complete. And if you look to the other directives, they only say you need to do this and this and this and this, but they don't tell you how you're going to do it. And if you look to EN 50600, they also explain to you how you can reach their demands. And the nice part is that other regions like APAC and like maybe even the USA are adopting the EN 50600. So it can be a very good base for creating your reporting and other demands. So if you look to the EN 50600, for example, you take the PUE. Sorry. The PUE. PUE is normally total power divided by IT power. And I see a lot of different PUEs in reality.

15:35
Martin Matse: So some say, "Okay, we have a windmill, so we subtract the power from the windmill." They all say, "Well, when we start, when we're running on genset, we're not going to add to that because it's not fair." And the other say, "When we're testing our UPS, we have to unload and load our batteries." So we're going to subtract it again from the PUE. So there are many, many interpretations of the PUE, but the EN 50600 is quite clear how you should interpret the PUE and also adding those other things. So liquid fuel, gas fuel, everything is added to that PUE. And they've also other metrics, which you can use like renewable energy, energy reuse, cooling efficiency, carbon use, CO2 footprint. They tell you exactly what you need to report and how you can create the reports and what's needed to come to the data.

16:19
Martin Matse: And they give you some tools to optimize and to improve your data center. So it's quite an interesting standard to use next to the one, which are mandatory in the region where you're active, because it gives you many answers on those other standards. For example, there's another one, the EERD, the Energy Efficiency Reporting Directive. It tells you more or less the same. So you have to create reports based on traceable data. It has to be very detailed. You have to identify optimization measures and you should identify how you can decrease your energy consumption, cost saving. And what I think is very interesting, at least it's for me very interesting, you really need an EPMS or a certified EPMS. If you have a small data center, you can do it with a simple EPMS. And if it's a little bit bigger, you need a certified one. And what is small and what is bigger?

17:11
Martin Matse: If you look to the standard, they say a 350 kilowatt data center is a small data center. So actually that's about 30-50 RECs. And a 3.5 megawatt data center is a big data center. And that's around 200-300 RECs. And actually for me, that's still a very small data center. So I think in reality, you cannot do without an EPMS in a data center. You might know the Uptime Institute. They have a yearly survey among data centers. And every year they have somewhere between 800 and 1,200 data centers responding to the survey. They started in the year 2007. And this is the 2022 survey. And what's very interesting, if you look to the 40, 35, and 31%, it's all saying, help me with reducing my energy consumption, help me to improve my CO2 footprint, help me to find renewable resources. So it's all about energy using and optimization. The 30% is a bit worrying because they ask, "Can you help us to understand how we should apply an effective sustainability strategy?"

18:20
Martin Matse: So this tells me in general that data centers are searching, how can we comply with standards and how can we improve? At the moment, they don't have a good idea and not good tools to do so. And this survey helps, confirms that image. This is done by one of the biggest international hardware vendors in data centers. And I don't know the number of respondents, but it's quite a lot. And what I found very interesting, so they asked their customers, "Okay, what can we do to help you to comply and to improve your data center operation?" And if you look to the last one, the customers were asking for a more energy efficient UPS, that was 78%. I looked at their product range and I did find out their UPSs have an efficiency between 93 and 97%, depending on the type and the load and so on. So data centers are asking, "Can you add 1, maybe 1.5% more efficiency to help us?"

19:19
Martin Matse: And actually, if you look to the losses on UPSs, it's already quite small. And to be honest, it doesn't bring much. So this also confirms that for me, the data centers are still searching, how can we improve our operation? Like I said before, the PUE is still the Holy Grail in data centers. And the PUE Power Usage Effectiveness is total power divided by IT power. So it'll always be above one. And again, this is coming from the Uptime report, 2023 report. When they started in 2007, the global average PUE was 2.5. And if you look to the report in 2013, it was 1.65, and after that, the PUE flatlined, so it didn't improve anymore. So the last 10 years, it was more or less stable. And a good question is why is that? And actually, if you look the steps that data centers have taken to improve, they only take one step in general and that step was hot air, hot aisle or cold aisle containment. So that was a big step. And after that, they did... That's almost nothing.

20:29
Martin Matse: They have no good idea how to improve anymore. So that's why it's flatlining through time. And 1.5, 1.6 still means that around 30% of your energy, 35% of your energy is being used, spoiled on other things than IT equipment. Then they ask to data centers. "Okay, what kind of metrics are you using?" 88% said, "Okay, we're using power consumption. That's for us the most important metrics." And 71 said PUE. The other ones are much more and others are not even mentioned. So you see again power and PUE is still the only method the data centers are using knowing to improve. Of course, if you look to our system this is an Ignition-based system. We have the PUE, the iPUE, the partial PUE, and it's split up in different levels and parts. So you can see the background and it's also containing the ER, but cell is only PUE because this is what customers in general ask for.

21:28
Martin Matse: "Can you show us the PUE?" Yes, we can. Okay. Now back to you and your data center. So imagine you have your own data center. You're the CEO of a data center. You are responsible for it and you have some demands. You want to improve your asset utilization. You have some outages, quite frequent, quite lately. So you want to decrease the risk of a new outage. You want to manage your CapEx and your OpEx. You want to improve your performance, reduce your costs, and you want to have more information about your energy balance because you have to report to it and you want to comply with new standards. And your conclusion is, okay, we need a DCIM solution to help us to reach that goal.

22:09
Martin Matse: Okay. So that data center got a visit from the local administration. So this is data center. It's using, as you can see, 4.1 megawatts. So this is a so-called big data center. So they need to have a certified EPMS system. And the local administration said to the CEO, "Listen very good. Your current PUE is 1.46. And you know, if you look to your permit your maximum allowable PUE is 1.3. So your current PUE is way too high. Now you have 12 months to bring your PUE below 1.3 because this is not acceptable. You're exceeding your permit limits." So the next step is the CEO went to the operation manager. He said, "Well, I had a visit from a local administration, our PUE is too high and it should become below 1.3 coming month." So the operation manager responded to the CEO. "Well, that's very nice but if I look to my EMS, my best part of my PUE is 1.3."

23:11
Martin Matse: "My worst part is 1.6. And what I already did, I did cold aisle containment. The vendor of the CRACs and the air handling units and the chiller came by. They optimized their equipment. So I think I did everything I could. And to be honest, if you look to the design PUE of a data center, it's 1.25, but the reason that it's so high is not because of me, because I did everything I could. It's because of sales. I see that room two is hardly filled and sales should sell more. And if the room two is also filled, then we are within the range. So go to sales. There is nothing I can do." And actually this is quite a normal situation because in general an operation manager has mainly his PUE to manage and control and optimize his data center. So he has no more ideas how to improve anymore.

24:03
Martin Matse: What you need is an integrated DCIM system, like the one we are using based on Ignition. If you go to the top right part, you see the white space. In the white space, you see the RECs and the temperatures and the humidities and the CRACs. So you have a good idea about your SLA contracts and so on. The left part is your EPMS. So the Electrical Power Monitoring System, you can see your power consumption. The bottom part is your BMS. So you know everything about oil usage, diesel usage, water usage, and so on. And the right bottom part is the CMS, the Cooling Management System. So now if I change, on the CMS part, the water temperature of the chiller, you can straight away see what it'll do on your energy consumption. And you also see straight away what it does on your whitespace.

24:52
Martin Matse: So you have a good idea if you change something, what the final result will be both on efficiency and both on white space. If you have an integrated system you can do much more with it. One thing you can do, you can put all the information in a balanced scorecard. So this is a balanced scorecard, again, built in Ignition, in the middle there's a composite performance index. But if you zoom in to the balanced scorecards, which I will do step by step, it starts with the PUE, you can see that my PUE is 1.46 as showing in my previous story. And you can see the two parts, 1.25 and 1.62, for the two rooms.

25:31
Martin Matse: And the... So together it'll follow the PUE. And if you do the math again, like it did in the previous slide, you can see, 4.1 megawatts power consumption, 2.8 megawatt to IT. So that will result in a PUE of 1.46, and it should be below 1.3.

25:55
Martin Matse: Now we're looking at the EER. The EER, the Energy Efficiency Ratio, tells me something about my cooling efficiency. It says 6.2. And if you want to calculate the EER, it's IT power divided by the power needed for cooling. And you can see in room one I have an EER of 8, and in room two I have an EER of 1.6. But I'm using a DX cooler and a direct expansion cooler. And the vendor told me that an EER of 7.5 is feasible and everything is perfect. It can even go up to 9. So I know in this case my EER is too low. And I also see in section two, room two, that my EER is 1.6, which is extremely bad. So I know at least something is wrong in room two.

26:44
Martin Matse: Okay, and looking to my part of my UPS capacity, I see two things. I see that my total load is around 30%, 30-31%. So that's quite low. And UPS is never efficient when it is so low. And the other thing I see is that there is an imbalance. So between phase one, two, and three, there are differences. And the differences in the phases will result in zero occurrence. And zero occurrence is 100% loss. So two things are wrong, balance and load. And if I look to the total UPS part, including my cabling, my UPS efficiency is at 87%, which is not very good, I think.

27:29
Martin Matse: The cooling part again, this really scares me. Because what's happening? I have two rooms, section one and section two. And I know in each room, I have five CRAC units. So room one has five CRAC units and room two has five CRAC units. And I see that my load in room one is 83%. So I know I've lost my redundancy over there. So that's a big issue. Because that also does lead to an SLA violation. If I look to my temperature, my humidity, and my rate of change of temperature are still okay. But I've lost my N+1 estimation, which is quite worrying. Because if one CRAC unit now will fail, then I will have an outage on that room. And that's something I really want to prevent.

28:14
Martin Matse: What I did? I went to my customers and the customer from room one. And I found some very nice customers. And they permitted me to move from IT from section one, from room one, to room two. So that's the step I'm going to take. I'm going to move a part of my IT equipment from room one to room two. Now something interesting has happened, as you can see. My 83% in my section two is now lowered to 61%. So I have restored my N+1 again. And the second interesting part is, if you look to a CRAC unit in general, it's the most efficient one that runs around 60% load. If you go lower, it's less efficient. And if you go higher, it's less efficient. So it's a double win.

29:00
Martin Matse: That's late compliancy. My N+1 is back. So my redundancy is back again. And it will also straightaway have an influence on my EER. So now my section two has increased, as you might remember, it was 1.6 to 8.2. So instead of having a total EER of 6.2, I now have an EER of 8.3, which is above the target of 7.5, which my vendor said was the minimum I could reach with my EER. So I have a much more better situation right now. And the other thing is, that's also quite interesting, my PUE has lowered from 1.46 to 1.3, just by moving some IT equipment. And that's what this dashboard was already showing me.

29:50
Martin Matse: So I'm almost right, almost below or equal to 1.3. So what else? The other step I'm going to take is decreasing my UPS capacity and my load balancing. Normally, you use this kind of dashboard not to decrease capacity of your UPS, but you use normally more UPSs, and you're going to meet the, make sure that your UPS size equals the load you have. But in this example, we are using a moderate UPS. And so we're going to detach some modules. So then in that case, we're going to decrease the UPS capacity. So we just switch down some of those modules, do not use the batteries anymore. And the second step we're going to take is load balance. Sorry.

30:34
Martin Matse: And normally, if this is a load chart coming from Energy Star, UPSs, of course, the form of the trend will be more or less the same. The numbers will vary on type and size and manufacturer. And in these two, you can see if a UPS has a 30% load, it has a 87% efficiency. And if you go to an 80% load, you can say you will run up to 93%. So you can see this is a normal trend. Normally, when you come above 80%, it will decrease a little bit again. So you know your most efficient parts will be around 80% load. So detaching a part of the UPS capacity will help you to improve your performance.

31:14
Martin Matse: And the next step is load balancing. So the way we are showing it, the darker blue lines are the actual loads. And you can see an imbalance. And the green one are the failovers. And that's your redundancy. So in this case, it's not balanced, and you have lost your failover redundancy. That's wrong. And by balancing, you will save a lot of money, because then the zero currents will disappear. And that means, actually, theoretically, your cable length will be cut in half. So your losses will be gone in this situation.

31:46
Martin Matse: So that's the next step, load balancing. So make sure that every feed has the same load. So not those big steps anymore. So that's what I did. Detached from a modular UPS segment, we balanced the load. And what's interesting right now is that our efficiency went up from 87% to 94%. So we also gained quite some on the UPS part. And now our PUE is below 1.3. It's even 1.25.

32:17
Martin Matse: And do you remember in the beginning that the operation manager said, "Well, I did everything I could and there is nothing I can do?" Well, actually, he was right because he had no idea. He didn't have the tools. But if you have an integrated DCIM system and you put everything, in this case, a balanced scorecard, you have so much more knowledge and tools to help you. And there are many other tools, by the way. There are many other tools you can use to optimize your data center. I think the balanced scorecard is a big step forward because it gives you so much more information as just using the PUE, which normally a data center does at the moment. Okay, I also told you, what does it bring?

33:00
Martin Matse: So show me the money. This is a little bit bigger data center. It's a colocation site from 12,000 square meters. They're using a 15 megawatt hour. It's three times as big as the example. And the energy purchase price is $10 per kilowatt. And if you have a PUE of 1.46, your yearly energy consumption will be 13.1 million US dollars. If you take the step we just showed you, and the PUE is lowered from 1.46 to 1.25, your energy consumption is lowered from 15 to 12.8 megawatts. Your new energy cost will be 11.2 million US dollars. That means you will be saving 1.8 million dollars yearly by just taking some simple steps.

33:48
Martin Matse: In this case, a DCIM system will have an ROI of a couple of months, three, four, five, six months, and you comply with your local legislation. It's actually, it's very worth the money to buy a DCIM, but really use it as where it is meant to be for. Okay, that's my part of the presentation. I think back to Christopher.

34:11
Dereck Saunders: Yeah, I can take it away. Thank you, Martin.

34:14
Martin Matse: Welcome.

34:18
Dereck Saunders: Before we start the Q&A, I just want to let everyone know that you can get a free trial of Ignition today. It's quick to download, takes about three minutes, and you can use it in that trial mode for as long as you want to. And you can dive right in and explore the platform that's made it possible to build all the unique projects that we saw today. We also want to tell you about Inductive University, which is a free online training website. It has hundreds of training videos so you can learn Ignition step-by-step at your own pace, and there's also a comprehensive online user manual as well that you can refer to at any time.

34:43
Dereck Saunders: For those of you outside of North America, we want you to know that we have a network of international Ignition distributors who provide business development opportunities and sales and technical support in your language and time zone. If you want to learn more about the distributor in your region, please visit the website listed here on the screen, or you can contact our International Distribution Manager, Yegor Karnaukhov. If you'd like to speak with one of our account representatives at our headquarters in California, please call 800-266-7798. To reach our new office in Australia, please call 1300-10-8088. Now let's get to the Q&A. As a reminder, you can type any questions into the questions area of the GoToWebinar control panel, and I think we do have a few questions here. So first one, "How does one provide ROI in areas of DCIM other than HVAC control? Energy savings can be straightforward, but asset management and tracking is less so."

35:44
Martin Matse: Okay, so what I was showing was just one example. If you look to our own customer base, I think the biggest group of our customers are our colocation data centers, and what we did find out at colocation is, especially the multi-tenant colocation data centers, they have customers housing in, housing out every day, and a very big gap is found on the billing system. So we have a very complete billing system, which all kinds of checks and balances, and in general, we found that out at around 5-10% of the energy that's being used is not being invoiced due to broken meters, misconfigured meters, RECs not attached to contracts, different number of RECs than assumed. So there are so many things that go wrong. So if you build a data center, say 20,000 square meter data center, I think to build it, it will be around 250 million US dollars to build it, and you can imagine how much energy will be billed every year. So saving, well, increasing your revenue by 5-10%, just by having a right billing system, that saves you really millions per year.

36:53
Martin Matse: So there's so many things on which you can gain a profit, and the other one is, for example, is managing your capacity. There's always a very big gap, there's a capacity available and capacity sold. And if you eliminate the gap, we did it for some customers, they had a saving on CapEx for 50%, so they bought 50% less equipment to manage a data center, and they just match, they didn't match the equipment based on their sold capacity, but on the real capacity and a growth rate. And if you don't install equipment, you don't maintain that equipment, so you also save a lot on OpEx. So there are many ways in which a DCIM ecosystem can help you to save money if you apply it properly.

37:38
Christopher Markstein: Just to add on to that quickly, 'cause there's a specific note about asset management in the question there, I think Ignition, by really hitting that next level of IT/OT conversions through concepts like a unified namespace, enables the connectivity of those asset management tracking systems, or it could be done directly with the database connectivity of Ignition, further enhanced by tools like the native barcode scanner in the mobile app, for example. So there's definitely a lot of interesting options to explore there.

38:20
Dereck Saunders: Great, thank you both. Another question we have here that probably for you, Martin, "What is the criteria for budget estimation of your offering for DCIM?"

38:26
Martin Matse: Oh, that's a very difficult one. So, I think if you look to difficult projects, somewhere between $50,000 and $1.5 million, it depends on size and functionality. So it's just simple, straightforward measuring it. It's quite simple, but it's very difficult to give a budget estimate for this kind of thing. So it really depends how big, how many devices. So in general, if you look at our data centers, they have 50-60,000 infrastructure devices, so that's quite a lot. So down here we'll spend maybe 6-700 euros, 6-700,000 euros dollars for that part. But if you want to include billing and maintenance and asset tracking and optimization and they stopped winning. So it's not an answer. You can't just say, "Well, it'll cost you like this." But it all depends on size and wanted functionality.

39:17
Dereck Saunders: Great. Thank you Martin.

39:19
Martin Matse: Welcome.

39:19
Dereck Saunders: A few more questions that we have here. Does the Ignition Data Center Pack have a flavor for a colocation data center? So the industry pack is probably going to be somewhat generic. The idea is to kind of have some building blocks to build off of. So I don't know that it will be that specific.

39:38
Christopher Markstein: Dereck, I know we haven't committed the full scope of that just yet. But definitely there's a number of features that are highly relevant to colocation service providers. So yeah, we'll see what ultimately ends up in there, but I definitely see tools, even some of which are already in the demo, that colo providers will enjoy.

40:00
Dereck Saunders: Great, thank you, Chris. Another question here. So there was a question, "How do you implement action one?" So it'd be for you Martin.

40:07
Martin Matse: I presume that action one is moving IT equipment. Was that action one? Is that...

40:12
Dereck Saunders: I believe so. Yep.

40:13
Martin Matse: Yeah, yeah. Well, I think in general that's not very easy. Depends on the type of customer. So we have customers who are using, schools and universities, are housing in and they only run during day hours or office application only run during day. They can be moved. And we see customers who upgrade their hardware equipment and instead of, and then gonna use to go to other RECs. But if it's a 24/7, IT application, then it's of course more difficult. But the idea is, this was just in the presentation, the idea is that you fill your data center recording the knowledge you have and not first let it go wrong and then try to adjust it. But if you do it good from day one, so I see customers will really give penalties if customers create an imbalance on their feeds and that kind of things and they will say, "Okay, no, you cannot use this room anymore. You have to go to this room because of this reason." So it's more when growing and shrinking. And of course also in colocation, if you have a gaming customer, it will grow, it will shrink, and just when it's growing, make sure that it will grow somewhere else. And if it's shrinking, it's shrinking in the right space. So all depends on the type of customers you are having and the flexibility and so on.

41:20
Dereck Saunders: Great. Thank you Martin. Alright, I think we'll get to a couple more here and then we'll wrap it up. So ATS, DCIM is energy-as-a-service a solution?

41:32
Martin Matse: Sorry, I'm afraid I don't understand that question.

41:36
Christopher Markstein: Is the ATS DCIM solution, is that classified as an energy-as-a-service offer? And I believe the answer is no, but I'll let you expand on that.

41:47
Martin Matse: Yeah, no, it's not, no.

41:50
Dereck Saunders: Okay. And then what percent of data centers do you think have partially implemented systems versus fully implemented systems?

42:00
Martin Matse: If you look to the numbers, and that's what I found very interesting, and you can find those market reports. And what I noticed is that the growth rates of DCIM softwares was almost double compared to the growth rate of data centers. So there are still a lot of data centers having a very limited application running. So only a building management system or only a cooling management system. And I even see manual billing on power consumptions at very big colocation site, which really amazed me that that customer still accepted. So it, I don't know exactly percentages, but not all the running data centers have the right systems and it's, if you look to that new legislation and it's also recognized by the Uptime Institute, many data centers are not able to comply with the demands that are coming. So many of them still have to add a lot of additional functionality and build additional system to be able to comply with all that new directives, corporate sustainability, energy efficiency, etc., legislation. So that's also what's confirmed by the Uptime. They see a lot of problems coming for the data center market and especially in the colocation data center market.

43:07
Dereck Saunders: Great. Thank you Martin. And so we do have a question. "Is it possible to share the developed source program?" So we will share the slides that we show here. So that'll be available sometime after the presentation. The demo you can find at the website that we listed on the url there, you can go to demo.ia.io. That demo that we showed earlier, it's not available for download yet. But when we release that next iteration in Q1 of 2024, that will be available to download on the Exchange. So you can import all of that into your designer and check that out on your own time there as well. And then let's take one more. "What impact do you feel increased or proper water flow and water temperature will have?"

43:48
Martin Matse: Yeah, the what I... I'm, I think it goes about, this goes about cooling, isn't it? I think that's, but that's my personal opinion. I think the bigger the temperature difference will be, the more efficient it will become. Because if you look to the efficiency of heat exchanges, they become more and more efficient when the delta temperature increases. And, but luckily also if you look to IT supplies, they'll start delivering IT equipment who are able to run a much higher temperature. So I think that, the delta T will increase and what the ideal water temperature will be also depends on, I think the location and the region you are in. So if you're somewhere in the Nordics, you'll have much lower than water temperature than somewhere maybe in the desert, but still the temperature difference on water between inland and outland, should increase and that will make things much more efficient. And then it is also more valuable if you want to use, reuse heat from data centers. So if you have a high delta T on your water temperature, it's much easier to use your, reuse your heat. And your sustainability will grow quite strongly, I think. I hope that will answer the question.

45:00
Dereck Saunders: Great. Thank you Martin. And if we didn't get to your question here today, we will try to reply to that after the webinar. So, thank you. We will be back in January with a new Ignition webinar, so keep an eye out for registration for that coming soon. Until then, stay connected with us on social media and our weekly news feed emails. You can also stay up to date on Ignition and industrial automation through our blog, articles, case studies, and more. There's a ton of helpful content for you to explore on our website, so be sure to check that out. Thank you so much for joining us here today. Take care and have a great day.

45:29
Martin Matse: Thank you all.

45:30
Christopher Markstein: Thank you. Take care.
 

Related Article:

Empowering Data Center Growth: Leveraging Ignition for Scalability and Efficiency