Call in to speak with the host
Welcome back to Raritan Radio everyone. Today, we've got two guests, Yee Liaw and Jawahar Swaminathan mispronouncing them, but welcome back guys. You are just back from the Green Grid Conference in Santa Clara, California and you are joining us to recap your experience. We're glad to hear. Thanks for joining us.
Yeah. Thank you Steve. We are happy to be here.
No. Its always fun. And you guys were at the great conference, so you got a lot to tell us of that, go ahead and jump in. The theme of the Green Grid Forums website, if I am correct, seems to be resource efficient IT. Jawahar, is that pretty well some of what you experienced there?
Yeah. That's being their founding principle and I have been attending the Green Grid Conference from the first one I think back in 2005 or 2006 and their conference today is a lot more inclusive of the facility, it's my observation. I think they have built enough with IT efficiency gains and it's all the low-hanging fruits have been addressed that efficiency gains that you can do with blades and virtualization and cloud so now that could be part of the reason why the Green Grid is expanding there coverage to the facility side of the equation whether its containment or cooling or pre-cooling or site selection and so on. So staying through to their core I think the IT efficiency is still the primary focus but they have expanded pretty well and to the other side of the PUE equation.
Very interesting. Another thing I find on their website they've mentioned their four pillars of the Green Grid Conference that were covered and. Yee, did you notice anything about the challenges and opportunities of big data that's the one they list as the first.
Oh yeah. Yes. Okay, oh a big data, huge amount with data. Right now we are in we are information age as huge amount of data being collected in real time; Text message, email, you name it, online purchase, visit to doctor's office those data have been collected and among with data being collected is counted, not by gigabyte anymore or terabyte, it is now a petabyte. I don't even know what that is and I don't know how many zero for now, okay. But just think about to analyze this huge amount of data into useful information for whatever purpose you need, requires huge amount of storage and computing power. So in other words, there are lots of lots of extra skill a lots of data center have been built to process all of this big data or huge amount of data, okay. What's that mean in terms of Green Grid consortium? Well, you build a data center, how do you plan it? How do you operate it? What's the concern about the resource efficiency, okay? So this is exactly the challenge and opportunity that the Green Grid consortium wants to help improve. So this is an important topic for everybody to consider.
It's interesting, as I was driving back from lunch preparing for this show I was thinking about a new social media that we are playing with called Vine and its -- you probably haven't heard of. It just launched in January and it creates, it's a really cool social media that creates video it's very quickly, and it's catching on like crazy and I was thinking holy cow even more data heading our way.
That has video via cell phones and all these other devices, it's just flooding data centers. I am sure that's just gonna continue to be a big issue, big data. I am sure, right?
Absolutely. Yeah. Actually you mention Steve the data including the audio and video, so all of the processing power intelligent to recognize the video like a face recognizing or voice recognizing. They all require huge amount of computing power. Okay.
Just to begin the process, right exactly. Jawahar one of your interest in the show you mention was launching up for effectiveness of energy efficiency programs can you tell us what you saw with regards to that?
Yeah absolutely. (Crosstalk) Yeah so they've had a couple of sessions that was focusing on the energy efficiency rebates and that utilities provide to see what has been influential and who has been influenced the most based on this rebates that the utilities provide and to gauge their effectiveness and how to improve them and how to make them more attractive to end-users. So the session they had one of them was even actually named this Green Carrots. So they were looking at -- they did a survey among 5000 members about what kind of utilities are really influential and how much do they influence your decision on what kind of projects you undertake towards energy efficiencies. And the interesting part of the findings is that they have more than the end-users it's the vendor community that's driving the message on utility rebates and incentive programs. So the vendors is a lot more interested in making sure that the customers gain all the rebates that they can possibly get for an efficiency project in the data centers and on the same talking when they put the reasons why an end-user considers an energy efficiency program or other facility upgrade, the incentives that they get from the utilities happen to be not the top of that pyramid as to their top contributing refunds.
So the influence of utility rebates on an end-user it's not as tires we all thought at least in the vendor community. Let's focus to the question, so what is not working? They did another study to see so what are the other reasons or the influences when it comes to customers embarking on energy efficiency programs and it happens to be power upgrades at the biggest influence so in other words it's data center this is running out capacity that is the biggest driver for them to upgrade the facility rather than going at the energy efficiencies as this made the reasons for the upgrade. So that's kind of a eye opening for most for most of the vendors I would presumed at least for the housing and they also tell us clearly that what they have been doing in the field which is to talk about all the utility rebates and the incentives that you can get from the utility to the end-users having very effective in our phase and this study has this fluent that it's the vendors that are pushing the efficiency story and the rebates to the end-users more proactively than the end-users themselves going after this and I think there is a lot of strategy that needs to be guarded by the utilities themselves to make it easier on the end-users, but in the mean time, I think the vendors are doing their part to breach that gap.
That's great. You're right. They raises a lot of questions as of what's their motivation and obviously it's still power, power, power, right?
Mm -- mm. Yes. So the utilities have their own reasons why they are very much into providing its rebates because the cost of producing a kilowatt hour of electricity is far higher than the cost that providing an incentive to say what's it or whatever as electricity. So it's very clear for the utilities that they would rather stop the dice or incentivize an end-user to save energy than going after and making meal, but they also have the additional cost of making sure the energy that produced is sustainable and also it's efficient. So they can completely award that situation or taking this instead how the end-users save that or whatever.
Right. And the areas you come is you are still experiencing a lot more pressure on the other end like from a public relations point of view as well, correct? Was just how much power they're producing and what they can do to promote more efficient use. So you guys are coming in and sort of doing that for them which is great.
Mm -- mm.
So Yee, the third one is software and IT hardware efficiencies. What did you pick up at the Green Grid Conference regarding those?
Yes. Yes. This is definitely the major improvement evolution of Green Grid. Jawahar have mentioned that initially the Green Grid focused on the low-hanging fruits so when we use to be -- when we're walking to a data center, right? The temperature is near freezing. Very, very cold in order to keep those IT equipment safe and operation properly, but normally how we learn, it does not needs to be that. You don't need to spend that kind of cooling power to cool the data centers that much. The IT equipment can operate safety reliably in their much warmer temperature. So that's a low-hanging fruit we'd learned it and that improved the PUE. So the next focus is on the IT equipment all of those server running storage and CPU do all the computation so that's together how efficiently they operate running software. Guess what, based on the survey and the study, the server utilization is less than 15% used to be much lower, that's mean many many server are sitting idle doing nothing, that's number one. But what's there impact to the far efficiency. A server despite sitting idle doing nothing not running any useful program it consume 35% of the peak power. You spend huge amount of power doing nothing work. On the other hand, if server running 10% of CPU utilization you use 50% of the power, when it go to 70% you are running 79% of power so this statistic will give you there is a more utilize the server to run useful application at to his small efficient way of using power, okay?
Certainly. Yeah, that's really interesting. And software shows the way?
Software makes that possible?
That's correct. So we have lots of cal computing right? So the software running the cal computing got to have the intelligent to built it how to distribute the work properly among server that a next in mind that utilization of computing power, okay and that's the software intelligent we need to continue to improve.
You'd bring up another point that we have mentioned in probably at least one other reason -- probably to, which is all of this calls for a more granular metrics and data at the rack doesn't it in order to get that efficiency to work via software.
Absolutely yes, yes.
Very interesting. Jawahar, the fourth of these pillars of the Green Grid Conference they had referred to is Financial and Operational Efficiencies, what did you pick up there, on that?
Yeah that's a very interesting observation that I had with the financial side. So it seems to be recurring team and you know actually being the fourth pillar of the Green Grid conference on listed there is a financial motivation or an incentive. All good intentions are just that, just good intentions so everything is tied back to the profitability of the business and so the financial incentives are far out the image that the companies and organizations want to project in the marketplace and the sustainability goals they have. So influence since when they had a session about side selection versus or sustainability. It's a collision course which one wins or can they both coexist. Can you do a site selection with sustainability in mind throughout from day 1? Now the interesting factor is when they did a survey of member company and organizations, they have found that the biggest influence that rendered come to the side selection for a brand new data center that's gonna go out is not renewable energy or sustainable energy its actually power and telecommunication. So what was that mean so when it's comes to power it's actually the cost of power that's ranked the highest and then the availability of power that's also reliable and that is the highest priority and highest influence of a rendered come to the side selection.
Raritan's sustainability and renewal of the energy come they are at the court right after even tax incentives. So if you look at the top three, two of the top three influences or side selection happened to be financial. So that brings us to the question of how do we make sure then a new data center goes out it's gonna be cost to least to bring it up and about and also it costly run of an operational base and that's there like in the case of the eBay presentations that I will talk about in a little bit. The knowledge of what is happening in the data centers on a day-to-day or you're in an hourly basis or even in an instantaneously basis, how tremendously for the data center operators and also for the people that are in charge of the budgets of either the data center or the entire organization. So even though I'd looked like sustainability takes a back seat if you can tie this back to gaining financial and operational efficiency sustainability looks good everybody wants to get into it. One of the things that we see the data becomes the driving factor to show that there is ROI or even the sustainability initiatives and of the efficiency initiatives. So where is this data come from. This is one of the major contributions from the Green Grid for the rendered community to say, if you have data there is always user and if you had more data there is a lot more useful in how your clients invite the data and one of the things that Raritan has done right from the beginning that all of products is to incorporate the velocity, took a wide-end user with as next data possible and then it's up to end-user to use the data in an intelligent way and we can enable that with the different tools and like give us mentioning software and such, but the emphasis is to provide the data in the first place. And so seeing that data like in the site of eBay, makes it come together, where now sustainability can co-exist with all of these financial incentives and the operational efficiencies that you can gain.
I think that's a great answer, Jawahar. I think that leads right into the next question which is this guy name Taylor that who puts up videos on the Green Grid website in which some people mentioned that data centers are now working with the right metrics and that sort of what you are touching on I think Jawahar and Yee do you agree with that? Our data centers working with the right metrics finally?
Oh okay. Yes okay. Jay Taylor, he is from electric in his company in __20:17__ and he is quite active in that the Green Grid activities. He is a wonderful guy I mean his video as you mentioned many people are saying the data center operation has enough working with the right metrics. Yeah, from the Green Grid point of view they have main theme of resource efficiency IT that is what we are looking so and the reason is, there is folk it's economic is social environmental, okay. Touching on three important aspect of operating a data center and how we do that is through introducing metrics with the metrics you can measure it you can analyze it and then improve upon it. So during past several years as we all know that the number one metric is most popular is PUE, now it's vaguely recognized every data centers shall must do this type of measurement and then base on the benchmark baseline you analyze it why my PUE is so high where can I improve it and continuously you improve the efficiency of how you use the power, okay? So this is number one metrics. On top of that they have metrics for measuring the carbon usage called CUE, carbon usage efficiency.
And on top of that is the on site generation efficiency and energy carbon intensity. For example, let's say they try to encourage data center to taking the green energy as much as possible from solar power or wind power instead of carbon footprint power source so they have a metric for you to compute how many percent of your power is the green power, okay. And also there is another metrics called waste utilization. After all, the data center will generate about the heat, heat is a good source, good power source, good resource. How do you reuse the heat for other purpose, they have metric for you to do that, okay. So, all these metrics are wonderful, they defined it and they help you, teach you, how to measure it and how to analyze it and then continue to improve it so that's continue to achieve the data center resource efficiency operation, okay. So, yes, the answer is definitely yes, we have lots of good metrics to work with
That's a good thing. I have always noticed and it's my opinion not a fact that one seems lag the other one seems to lead the other -- all these data is flying in, we're talking about big data earlier. All this data is flying in and people are like "Okay, we got it now, what do we do with it?" And now it's like the software exist and Raritan is a big part of that with Power IQ about the solutions that allow companies to actually make the data intelligible, help data centers manage all that data correct?
Yes, yes, okay. Well, no one know Steve when we're speaking of PDU, actually, this is a three-level accuracy and to achieve the level three, you do need an important product or intelligent PDU that's why the Raritan has produced and market it. Through that level of measurement, you can achieve the accuracy and to gain the maximum improvement.
Okay. So, Jawahar what's the next step in getting more metrics, getting even more granular data because we're never gonna quite get this perfect and more data and more understanding of that data will lead to more power and energy efficiency, right?
Yeah, that's a good question so the industry is knowing -- it's so interesting to look at how the end users are using that data and how the vendor communities providing that data so the vendor community keeps innovating for instance, Raritan tries to __25.21__ was saying tries to provide as accurate data as possible as grand as possible to the inducive community and on the other side, the inducive community is looking at it and posing the same question that you have which is how all this data, how I'm gonna use this. I know it's valuable dealer, it's a great dealer to have, but if I can't find an effective ways to present it and tweak it and make use of it then it's just data, it's just big data residing in some cloud or some constructor. Now, that's one of the challenges that eBay took upon themselves and to find out how am I gonna get this data to make sure I can use it to tune how my business operates. So they devise a new efficiency metric that's for their own organization which they called digital service efficiency. What it is, it's just a, it's more like a miles per gallon or eBay's technical infrastructure to put it simply so if you know that your car often running as well and it's not getting as much as MPG then you tune it. Now, if you don't have a data for your business, how are you gonna tune it? The best take made a goal and they took three years to bring this to a state of realization where now they have a single dashboard with different people in the organization with their different roles.
And look at the exact data that they want to know and make some intelligent position using that data. For instance, they have taken a model of, there is a state of role that people play and there is a currency they are interested in and how that currency impact the business and what are the nods that they can use to create it and if they do so effectively, what are the end results gonna be. So as an example, we had laid this out, but in the case of eBay like STIO or STFO as a role will look at transactions -- number of transaction at their currency. So the more the number of transactions, they feel good about it, but at the same time, they all something to look at the impact that it has on their business than it is. It cost us each transaction or the environmental impact that each transaction has or the performance of each transaction so they can make improvements on it although the revenue that each transaction drinks so now that's the impact, those are the business impact. Now, how do you click it, now you know what you're looking for and how it impacts the business, now, how am I gonna click it. So those are the nods that they're talking about and the nods would be business deficient based on the data that you see that impact the business and once they addressed that, then they get the result which is -- if you look at the -- okay, I have a performance so every transaction taking longer or shorter as __29.02__ and if they see that it's taken longer then that's need to be tweaked.
So that they can accommodate more transaction within the same number of minutes in an hour. At the same time, if they see the transactions are improving, but the revenue approach transaction they get from their website is not as high as last year then they need to do something to address that and those are the business decisions that they can take once they have this level of data then they can set goals for the next year. Let's say this year, my average revenue per transaction is for a thousand transaction is what they measure is 87 cents. Is that sustainable or should we go higher and it's still what it is? So you can look at the historical data and say you know we're here, my goals would be 10% so that could be a good goal for 2013 or 2014. So without the data and the definition of what that currency really is and the impact that it go higher on your business, there is no way you can tune the performance of your organization and business to achieve the results that you need and you can fill it down to the data in their operative level. This is great for a CIO to know how many transactions are happening and how I can tweak it, but at the data center manager level, the number of transactions is not evident for us because she is operating one data center, so now, you can use same model and change the role which is from a CIO. Now, we are looking at a data center manager, the currency now could become the overhead follow that the number of watts that you want to consume or conserve then the business impact could be how many more data centers do you need to built to keep up with the demand that you have for a followup.
And how do you tweak it? You tweak it by looking at the POE, you're looking at the cooling efficiency and the same which you get from intelligent design from the follow distribution site and if you look at the results as to how do I maximize the output. Steve, are we good?
Steve, are we still good?
I guess we can continue.
But do we lose Steve?
Yeah, it's a... Well, why don't I ask you the rest of the questions if you wouldn't mind so we can continue this?
Okay, no problem.
Alright, cool. So the rest of the things that you wanted to discuss were the Data Center Maturity Model was one of your entries that you had in Green Grid and you had been involved in it and so we wanted to find out what you found new the Data Center Maturity Model and how that's going right now?
Oh, yes, yes. I think the Data Center Maturity Model, we call it DCMM, is a very comprehensive model to gauge the efficiency of the data center operation with respect to company executive, okay. When you're running a company, you have the right to know how well my data center is running. Do they making money for me or they're just not running efficiently, what can I improve and what kind of capital investment I had to do to improve the data center operation and what kind of expense I can reduce, okay, to maintain a reasonable cost. All this question can be measure and analyze to DCMM model, okay. It's a very comprehensive model consist of six level of maturity from 0 to 5 and if you use a tool or training and use a tool, you can find out how well you operate your data center in the way you can do the improvement from energy point of view, from the capital investment point of view, from the operational expenses point of view, okay, so this is a clay model at the Green Grid, put in lots of effort to develop this model and then develop the training and the tool to help the member of company to use a tool. So it's quite valuable tool to the data center operation.
It create a clear road map for a data center operation to significantly improve energy efficiency and sustainability through the bench marking and other data center facility of performance and clearly show where improving can be made and DCMM also online could invest practice in the five-year road map for data center to integrate every aspect of data center operation including power cooling, computer storage and network and all full guideline based on the data center area such that you can benchmark the current performance using Data Center Maturity Model Equalizer, there's another sub tool.
Guys, I'm back on now, sorry about that. Yee, the website greengrid.org has a lot on the Data Center Maturity Model, it looks very robust, they've got lots of papers on this so I think people could check that out at thegreengrid.org, sorry go ahead.
Yeah, okay. I have a question for you, Steve. We're talking about Data Center Maturity Model a little bit with Jawahar. Should I start from beginning?
No, no keep going.
Okay. So continuing to the value to the data center operations, finally that the DCMM will help you to design a new and support decision making the data center upgrade, okay? So if you're planning an upgrade or building a new data center that provide the design guideline, sort of, to do the proper decision making every step of the way from planning, site selections, building it and how to deploy that and finally to the operation, okay? So it is a very comprehensive tool, a good way to gauge how well you operate a data center. Thank you, Steve.
Absolutely. The another thing why you're asked that, another thing that Raritan offers that's getting a lot attraction speaking of software is DCIM solution. What was the awareness of DCIM and software as a solution for power management and power monitoring at the Green Grid show.
Yeah, okay. Thank you, Steve for the good questions. Well, the software as a tool helping data center operation, I believe the member of company of Green Grid begin to pay attention to it, right, not until this year, really we talked about what kind of management tool we shall include in the data center operation, okay. So it's a good news for Raritan since we have two products power IQ and DC track and both sort of slightly touch on in this particular conference. Power IQ for example, it does address the energy management including the computation of PUE, okay, so it is a great product, okay, we will continue develop and enhance so that we have proper market share and support our customer to improve the energy efficiency. On the other hand, more sophisticated infrastructure management tool, the DC track, it begin to talk about it among the participant and some session to talk about it and become one session does a survey they felt currently that the software product, the DC track is more like a early adaptive phase. Only less than 10% of company had purchased it and deploying them.
However, many, many participant had indicated that within next two to five years, they would purchase and deploy DCIM solution for operating the data center so definitely this is a good news in product maturity level, it get into a massive acceptance phase, okay. Once it happened, I think the DC track will gain not more attention and hopefully our leadership, the technical leadership of the DC Track we can gain a fair share of the market through this so I think that's my take in this conference, Steve.
Very good, very good. Jawahar, any final thoughts that you have about the show?
Yeah, I'm looking forward to the next one and it's always good to see what the world is doing in terms of data center efficiency and the Green Grid forum especially is very good in capturing what's happening across the globe whether it's in Japan where they constantly have to design their buildings and data centers to accommodate seismic activities or in Europe where they have plenty of free cooling, where they can leverage that as efficiently as possible and if this is a great forum for us to learn what's happening across the world not just in the United States, so looking forward to it and I enjoyed this interview.
Well, this has been a great way for us to learn and for the audience to learn, those who couldn't make it to the show, so I really appreciate the detail you guys have brought to the show in particular and in your accounting of everything you saw and framing it forwards, it has been a terrific show and Yee and Jawahar, I hope to have you back on here very soon.
Oh, thank you very much, Steve.
Thank you very much.
Bye everybody. In the meantime, our listening audience, please learn more at raritan.com and you can keep up with the blog there and keep up with the future shows that the folks of Raritan will be attending. After which, we'll be interviewing them back on our Blog Talk Radio show, right here. Thanks everyone, again, that's raritan.com. Bye-bye.
Sorry we couldn't complete your registration. Please try again.
Please enter your email to finish creating your account.
old-style code for hosted blogs
300 x 295
400 x 370
640 x 550
It's good to talk.