Call in to speak with the host
Hi everyone. Welcome back to Raritan Radio. My name is Steve St. Clair. With me today is Khaled Nassoura, the PE General Manager of DCIM Software with Raritan. How are you Khaled?
Good morning Steve. Good to be with you.
Yeah. Thanks for joining us. So you're just back from Data Center Dynamics which was October 16, where you presented a paper called Data Center Infrastructure Management, a real world first-hand experience and I just love the title because from what I understand, it's quite a bit of hype out there about DCIM and there are many examples of success stories to go along with it. Is that correct?
Yeah. So in this show, you're going to be telling us about an important financial institute in Canada where you're working with currently, is that right?
Yes indeed. They are of customer for hours who had the chance to explore DCIM and used it and recognized the benefits of DCIM.
But first, I mean you brought your presentation at Data Center Dynamics into two parts. So I want to start with point one, which was kind of an overview that you did. So what are some of the common pinpoints in data center management?
Well, you know Steve as I travel and visit with customers and I do that around the world and I get to see customers of all sizes from small, midsize to large enterprises and in many cases, these are Fortune 50 companies and you will be surprised about the commonalities of the pinpoints that they all have in managing their data centers. It could be sometimes something as basic as knowing what assets you have in your data center and then not only knowing what assets you have, but where are these assets located, what part of the data center, what rack therein and what are you positioned therein, but I admit to you that that could be sometimes a basic problem. They use an asset management system to solve and keep track of these assets, but the problems in the data center go beyond that. It's really knowing the physical or logical relationships that tie all these assets together as well as how these assets are tied to the infrastructure that supports them underneath. Right, so I will give you some example of this. You would like to know how perhaps an infrastructure element such as a circuit breaker or a network switch port may have an impact upstream on the servers that are connected to that breaker or to that network switch port and furthermore, what applications are running on these servers and what departments or what business units are being supported by these applications and these problems become very difficult to solve especially solving them on the fly when you have a problem and you're trying to understand those dependency mappings.
I imagine data centers. These must have real financial implications in terms of uptime and service level agreements, right?
Absolutely. So, many of these data centers especially in the financial sector, they cannot afford to have any downtime and when they do have a downtime, they need to be able to resolve these problems very quickly so they need to have a good visibility into their systems and further more into the infrastructure that support this system so it's understanding the relationships between the assets and the infrastructure extremely important that you need to be able to do the dependency mapping.
Now, knowing what's connected to what in the data centers so to understand how a power supply on a server is connected to the power strip and how the power strip is connected upstream to the power chain or perhaps how a network port or a SAN port on the server is connected to a patch panel and how that patch panel is connected upstream to the network's switch port and how that switch ports may be connected upstream to a course which are the router. It's important, but if you don't know what's connected to what, that means you do not know the resources that are being consumed in your data center. So to me all these infrastructure are resources that you live into and you use everyday and that really ties in to a very important topic in data centers today which is capacity management and capacity planning and it becomes almost impossible to figure out what capacity you have if you don't know what resources you are using at any point in time. So to me, capacity planning is a simple arithmetic, it's to know the total provision of infrastructure in the data center and that's usually a very static information that you get from your architects and engineers who have designed and built your data center. That's the total sizes of data center, how much power you have in there, how many network switches and ports you have, how many power strips you have, how much powers you have available in every rack, that's pretty static, that's what you supply into your data center. What's very difficult to get a handle on is the number of resources that are in use at any point in time and that's a very dynamic set of information and you can only get a handle on that if you have a system that can look at your data center from a holistic point of view and keeps track of what goes in and what goes out, what's connected and what's disconnected.
So going back to that arithmetic of a capacity of planning, it's really the total provision of infrastructure which is free static minus what's in use at any point in time and that's the difficult part, that equals your capacity. So without a DCIM solution, it becomes very difficult to get a handle on your changes in the data center and therefore, you will not be able to calculate your available capacity accurately when you need it.
Right. So that must make it pretty difficult for these guys to comply with regulatory audits, correct?
Absolutely, because they're doing things ad hoc, every time that they need to know what they have in the data center and many times they're being asked to do that. But for audit purposes whether these are internal audit requirements or external regulatory audit requirements, they resort to days and weeks of manual surveys to figure out what they have so they can simply give you a snapshot of what they have in their data center whereas customers that use DCIM -- DCIM is an operational system that will tell you at any point in time, not only what's in use, but will tell you what you have so you'll be able to do that. Literally with a click of the button, you will be able to comply with these requirements, but really what becomes important then, once you deploy DCIM is how do you keep this information organized and how do you keep it accurate and that's really when the tool needs to tie in very nicely and very tightly to your change management process even the change control that you need to institute and then enforce into your data center. So without that, you'll end up having garbage into the system and you'll get garbage out, so excuse this -- excuse me for this terminology, but you need to put a good process in place so to make sure that any changes that are taking place in your data center are being -- you're following best practices to do them and then you're putting that information into your system using a rigid and structured change control processes.
So, there's a lot of buzz about DCIM and obviously, a lot of data centers are starting to take it on and look at different options. So how is the data center operations model evolving at this point?
Well, that's really an excellent question because the way data center are organized might be one of the reasons as to why data centers don't have a handle on what they have in their data center. As you know and as many of your listeners had -- you know that there are multiple groups or domains that exist within a data center. You have the facility's group that is responsible for the physical space in the data center, the cooling plant, the power plant and then you have the IT infrastructure groups that are responsible for the networking resources, everything from the structured cabling, to the networks, to the SANs and LANs and __09.16__ and then you have the IT system's group that's responsible for the whole system, the servers, the storage systems and these are the people that are essentially the drivers of what goes into the data center. But what I have seen over the past two years, that as these data centers have become their own stand-alone facilities, they are no longer attached to the corporate headquarters building. They really have become their own stand-alone facility. You have a new group that is emerging that's called the data center ops and this group is in charge with managing the day-to-day operations in the data center. They're in charge of receiving request from the various groups especially the IT systems groups. They are in charge of racking and stacking the servers, putting them in the racks, connecting them to the power chain, connecting them to the network infrastructure and this group now needs to work seamlessly with all these other domains that I mentioned, the facilities, the IT infrastructure, the IT systems groups and they lack the tools to be able to communicate properly so they all end up resorting to manual processes.
So every time, I'll give you some example of how things happen in the real world and some of these data centers that I have visited, so every time that perhaps the IT system's group put their request to add a server or better yet add a dozen servers, the data center ops groups tend to react to that and they react manually. They dispatch people into the data center and many times they dispatch multiple people based on the size of that organization. They may send a technician to figure out the rack spaces available in order to support that request that they have. They may send a second person to figure out the cabling and networking resources that are available in that cabinet to support their request and they may send a third one to figure out what power is available to them in those cabinets and this is really a very manual process, very laborious and some of our customers have told me that this process takes some several weeks from beginning to end and by the time they've reached the end of that process and they figured out what rack space they have, by the time they reset in to that process, the rack space is gone because there is no good way to communicate preferably across all these departments and domains, so it's really almost as a sad state of affairs and our customers recognize it and they are really, they're desperate and they're searching for better ways of doing things. So, this manual work not only takes time as I mentioned, it could be weeks, which slows down your response to your internal customers and some cases, if you're external customers, if you're in the __12:16__ hosting of business, but also that time as up to a lot of man hours, and those man hours cost money so you could easily shorten your SLAs.
You could certainly reduce your cost by highly automating these manual processes in the data center.
So I imagine based on what you just said that companies who were looking for and starting data centers and comparing them, if not already, certainly, those who were looking are going to start to look for DCIM as a major factor in who they choose, would you guess?
I know (crosstalk) I mean it sounds like a mess out there. How does Data Center Infrastructure Management Systems, how do they address these problems you've defined?
Well, it's an excellent question. So DCIM followed therearound for now, we can say going back to 2007 where many of the vendors have emerged, but I will still consider this industry to an emerging industry, but the big part about the DCIM that it's a holistic system that is trying not to address one of these silos or one of these groups that I've mentioned whether it's the facilities or the IT infrastructure, IT systems or data center ops, it's a holistic system that's meant to serve all these groups together and regard as where people sit within each of those organizations whether they are the technicians, the administrators, the managers or even the VPs that run each of these groups. So DCIM, a true DCIM needs to have multiple components to serve these groups. At its most basic level, DCIM needs to be able to understand the configurations of all the assets that you have, okay, but as I mentioned before, tracking the assets really does not solve the complete problem. So what you need to have is a strong connectivity management so you can understand how all these assets are tied together physically and logically. So the first component is the asset and configuration management, the second component is the connectivity management and then that would allow you now to do better capacity planning so you need a good capacity management component, but without strong change management and change control processes that are built in into the DCIM tool then the capacity planning will not be accurate and...
If it's not active, people will not use the system and then you'll end up going to this downward spiral because the less people use the system, the less is going to be updated, the less rival that the data is going to be in it and the less people will believe in it and then again, they'll use it. So, these four components need to be highly integrated and they need to work holistically, okay, and then to make it easy to use and to increase the adoption of this systems, you need to have multiple layers of visualization tools within DCIM so you want to allow the data center operators and managers to visualize the data center from the top view and then drill down to what they're looking for to drill down to a rack, so they can see what's in the rack then drill down further to every device and server in the rack and look at its front and back so they can understand what ports are configured on that device. And then from there, they need to be able to visualize the circuit that is connected to these ports so they can do the dependency mapping to do it in a visual way. And then finally, you need a real time component to DCIM because without it, you are relying on static data that's being entered by the users. So you need to be able to collect real data in real time from the data center from various parts of the data center and feed it back into your database to make sure that whatever you have in there is accurate. So you need to be able to collect data from power equipments from panel boards, from UPSs, from -- certainly from intelligent rack PDUs in the cabinet. You need to be able to connect that from environmental equipment whether it's the air conditioning units, whether it's the air handling units, the temperature sensors, humidity sensors, air flow sensors.
You need to be able to collect all of that data and then feed it into the system in a meaningful way so you can now take -- make smarter decisions based on informed data. You can make informed decisions based on good data rather.
Well, Khaled, that sounds like quite a tall order for a software to manage and I know you wouldn't be telling us that unless Raritan had a solution that does it, but we're going to get to that in a little bit with your real world example that you're working on. So this clearly outlines or starts to outline where the industries had it and can you tell us a little more about that like where you see the industry going towards DCIM.
Well, I'll probably characterize that -- many of the vendors today in this phase are still in the first generation of DCIM and I see that Raritan were working hard to enter at the second generation of DCIM and I will suspect that some of the vendors are also doing the same, but we are trying to take a leading position in terms of how we can help our customers adopt DCIM by making it simpler, simpler to implement and then simpler to use and that's really critical, right? You don't want to complicate it because some people will walk away from it. It's not going to -- it's not going to help them. So we need to make it simple to deploy simple -- to operate. Now, how do you that? It's very easy to say what I just said, but how do you make it simple, it's by increasing the intelligence that's built into these tools and increasing the level of automation that you have in these tools and finally, increase the level of integration that you have between the DCIM system and all of the other software tools that may already exist in the data center because DCIM is not necessarily going to replace all the other software tools and point solutions that someone may have or some of the enterprise asset management and enterprise change management. You need DCIM to be integrated and federated with these systems. So we are working hard on increasing those, that sophistication of automation and integration into our software.
Great! And you got a real world story that you want to tell us about which I think gives your discussion a lot more credibility because it's already working for a customer of Raritan. Can you tell us a little bit about that?
Like for instance, why did your customer choose to use -- why do they need DCIM?
Excellent question, Steve. So our customer is probably -- there was no difference than many of the customers I've referred to before that I've visited around the world where they were doing things the old fashion way. They've had multiple computer rooms. They never understood what they have in these rooms and because they don't know they have, it was very difficult to understand the impact analysis that maybe an infrastructure on it would have on their servers, on their applications, on their users. Likely, they were in a good position that they've been charged with building a new primary and secondary data centers to consolidate older computer rooms into these primary and secondary data centers. But because of that, their executives understood that in the past, things did not work very well because of the silos that existed. So they had a big -- they had a mandate to think out of the box and to say, "We need to have better integration between IT and facilities." So that was one of their major drivers that they need that integration at a high level between the different groups, IT and facilities. The second driver is the fact that they were building new data centers. That was the Greenfield opportunity. Anytime that you have a Greenfield opportunity, you can think big and you can say, "We're going to do things better" and that really helped them in saying, "Okay, DCIM would help us get there." And finally because of those new buildings that they were constructing and now they are using, they are no longer attached to the corporate headquarters, to the facility buildings. They are stand-alone off-site facilities.
So they could not afford to do things manually and dispatch people into the data centers every time they need a question answered so they have truly wanted to operate their data centers with lights out. So they needed smart tools to be able to not only tell them what they have, but they needed the tool to manage the data center to orchestrate the physical changes that takes place in the data center. So these three big drivers got them thinking about DCIM and they said, "Well, what can -- how can we benefit from it?" So they really listed a number of benefits where they said "At the end of the day, this is how things should work." So they wanted to automate their manual processes so they no longer want to be dependent on these people going in and out of the data center to figure what to do. They wanted to do things literally with a click of the button. They wanted to -- when they make changes, they didn't want to do it ad hoc and say, "Okay, let's try to do this and put that server there and oops!" and then a circuit breaker is going to trip maybe a few weeks or few months down the road because they did not do better planning. So they wanted to do that planning process and to do the "what if" analysis in the software and only once the analysis was done and the work was planned, then they would generate the work orders and the tickets that will help them execute the work. So this way, they can do better utilization of the new space and the new power infrastructure that they have in the data center. And ultimately, all of these will go towards their mandate of __23:48__ mining the communication between the various stakeholders across the IT and the facility domain.
Right. So as these guys went through the process in their search for DCIM, tell us a little bit about what they were looking for and extended even beyond that anyone else considering DCIM, what should they be looking for?
Alright. So I can share with you, for example, what this particular customer went through because it was a very extensive process, it was -- they formed a cross-sectional, cross functional rather theme to go out and look for DCIM based on the mandates that they were given. So as I mentioned before, bridging the gap between IT and facilities was important for them so they really listed that. One of their criteria is to have an equal balance in the tool that they are looking for and there should be an equal balance between IT and facility features and functions within that software. Now, why is that important? Because many of the tools out there today, they can probably -- you can trace their history to either coming from the IT services world so they have a bias towards IT feature set and other vendors are coming at it from the facilities worlds. Again, there is a bias of facility features. So they wanted a software that could have an equal balance in the data center between IT and facility features. The second criteria they set, they recognize that DCIM is not going to be the only software tool that they will have in the data center. They do have other enterprise systems in place and these enterprise tools go beyond the data center. For example, the enterprise asset and configuration management is concerned with all the IT assets within the enterprise whether it could be servers -- I'm sorry, it could be PC, the laptops and printers. So these are things that go beyond the data center and DCIM will not necessarily replace that. It needs to work with it. Also, they have enterprise change management.
An enterprise change management exists not just for the data center but for the sake of the entire IT enterprise that every time someone needs to make a change or they're requesting a new application loaded on their desktop or requesting a new laptop, that goes through the IT change management system. So those systems were not going to be replaced so they needed DCIM to work and integrate with this enterprise system. So they wanted the system to be scanner-based. They want it to be open. They want it to have APIs that will enable them to do that kind of integration. They also recognize...
It really is a tall order you're setting up here and the natural answer to this is Raritan has a software solution that solves this, don't they?
You bet and we do have a product that has been around since 2004 that addresses all these issues. The Raritan product is called dcTrack and dcTrack was designed with facilities and IT in mind right from square one and so they were able to see that dcTrack did not have a bias towards one area or the other. It had a complete functional and proven feature set. The feature set that we have in our software today also came from people that understood how data centers were designed and operated. So we didn't just simply hire a bunch of software developers and we threw some market requirements at them and said, "Okay, go ahead and build the product." It was done really in reverse as compared to most other software tools in the industry. It was really built by data center professionals who have the domain expertise who have witnessed these pinpoints of firsthand and said, "Let's really build a tool that can solve that." So having that viewpoint on how data centers are designed, managed and operated allowed us to create a product that can address all these points in a very simple and intuitive way. And keeping the software simple and intuitive is critical to the success of any tool that is being deployed in the data center because without it, people will avoid using it and unless they use it, as I mentioned before, it's going to go into this downwards spiral and it will end up being self-ware.
Right. And I mean if you were going to have a used case as company that makes this kind of software, you guys could not have done any better than the one you've talked about, a real world case where this is working and achieving ROI for client, a very important financial institution in Canada. It just -- it sounds like a slam dunk to me, Khaled.
Indeed and we were very happy to work with our customer in understanding their used cases and their requirements. And more importantly, we do not claim that we will have 100% of the features that every customer wants, but what's more important is to say to the customers that we're going to work with you and understand how you do your business and incorporate their requirements into our road map and our immediate next releases and that's exactly what the customer saw in Raritan and they are actually currently experiencing that, not only they saw it, but now, they are actually -- they're benefiting from that and how we can work with them and continuously enhancing our software because really we knew investing software is a journey and you need to have that journey in partnership with your customers.
So there's a very strong service component behind dcTrack software.
It is and also that's true about older DCIM software, but having a good service around your product is what's going to make you successful because DCIM, as you mentioned before, is doing a lot, is doing a lot for good reason because we're trying to solve real problems in the data center. But without having good services organization behind you, you will not be able to implement it properly. You will not be able to implement it quickly so that's definitely a key ingredient to the success of our customers as how they were able to rely on our services organizations.
So to wrap this up Khaled, those listening to the show can learn more about Raritan's Data Center Infrastructure Management solutions at raritan.com. When you get there, look on the left at the top under the products link and you'll see Data Center Infrastructure Management. Click in there, there's a contact link. They can read case studies and read about the complete offering. Khaled, thank you very much, very, very informative. You guys are all over this DCIM thing. I appreciate your time today so thanks a lot!
Well, thank you so much. Thank you for giving me the opportunity to speak with you and to your listeners.
Alright, thanks everybody! Check out Raritan Radio next time and we'll talk to you soon. Thanks Khaled!
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.