Cloud Therapy: EP 008 – Overcoming BareMetal Cloud Hurdles

Internap‘s Director of Solutions Engineering, Paul Herrick, discusses 2 great tips on how to overcome some of the hurdles to deploying BareMetal Cloud vs. traditional Cloud VM’s. These strategies will absolutely help reduce costly deployment mistakes, prior to purchasing. He also tells a great story of why he was raiding the office fridge at midnight, a couple years ago in Atlanta.

Want to hear more Cloud Therapy? Find us on iTunes and Stitcher!

Which Bare Metal Cloud providers have the highest customer ratings?

Click below to see which 3 service providers AeroCom recommends for your company.

MY TOP 3     

See below for a full transcript:

  Mike: Alright. If you’ve ever thought about bare-metal cloud or maybe you don’t quite know what bare-metal cloud is – you, kind of, have an idea that maybe it’s a dedicated server, but you’re not quite sure what’s involved or you’re thinking about deploying bare-metal cloud and you’re a little bit nervous about what you’re getting yourself into – this is a great episode to you. On the phone today, we have Paul Herrick. He’s the Director of Solutions Engineering for Internap. Paul is going to review with you what bare-metal cloud is very quickly and he’s also going to tell you about two specific tips regarding hurdles you might encounter when trying to deploy bare-metal cloud versus traditional cloud technology. These are two things that you can do in advance to make sure that you’re getting yourself into the right solution. Coming from an engineer, I think, these are great tips that everybody can get something from. Then, also, at the very end, Paul talks to us a little bit about Internap, and what’s going on there, what’s he’s excited about. I think they are doing some cool things with their bare-metal cloud, and also just all of their services in general. So, I think there are some good information all around. In addition to that, he tells a great story about when he was working for a previous company and a night that he’ll never forget. I hope you enjoy Paul as much as I did, talking to him. He’s coming to you today from the 17th floor in New York City. As you can hear in the background, there are some fire engines going off. I guess they are just, literally, right around the corner from a fire station. He tried to reserve a private room, but the VP of Sales commandeered it. So, just a little extra personality, some faint fire sirens in the background, but, again, he’s coming to us from the hotspot in New York City. I hope you enjoy it as much as I did, but before we get to the actual interview, I also want to remind you about our free gift we’re giving away. It’s the list of twenty-five or so questions that you want to ask yourself prior to shopping Cloud VMs for your business – not bare-metal cloud, but Cloud VMs. If you’re going to shop this stuff, you definitely want to ask yourself a bunch of questions that will really help you narrow down the providers that will be the best fit for the business. This list of questions took me hours and hours to compile by calling of our service providers that we know that sell Cloud VMs, and asked them a ton of questions about what they do best. From those conversations, I made a list of twenty-five or so questions that you can ask yourself that will help you navigate which of these providers is the best.

I’ll give all these to you for free, no big deal. All you have to do is text the word “CLOUDVM” to the number 44-222. Again, text the word “CLOUDVM” to the number 44-222 and we will send you a free copy of this list of Cloud VM questions that will definitely help you do a much better job shopping Cloud VMs for your company.

Thanks for joining us on the program, Paul.

Paul: Thanks for having me, Mike.

Mike: Can you please tell us a little bit about yourself both personally and professionally?

Paul: Sure. My name is Paul Herrick. I am the Director of Solutions Engineering at Internap based out of Atlanta, Georgia. In my past, I’ve started with a company called Cbeyond. They’ve been acquired and are now titled under Birch. I started with them as a sales rep directly out of college. Found that the best way for me to sell was to sit in a seat and answer people’s questions until they had to say “yes” or “no.” I wasn’t a really pushy sales person, but I can just know my topic well. You had to get around me another way than to throw me a curveball and get me out of your seat.

Mike: Awesome.

Paul: Did that for a while and I moved into a solutions engineering role over at Cbeyond. By the end of my tenure, I was a manager of solutions engineering over there and I had several people reporting to me on the East Coast.

Personally, I’m a single guy. I grew up in the Atlanta area. I have two sisters that are only a year apart from me. In fact, I have some family that’s been adding – my older sister had my second niece last Thursday.

Mike: Oh, fantastic. Congrats.

Paul: Thank you. I have one that’s two and a half and one that’s newborn at this point. It’s, kind of, like having a friend that has a puppy, taking it to the park and then giving it back – it’s the best way for me to go about it right now.

Other than that, I really enjoy being outdoors. You send me to the beach or the mountains, I can find something to do. I’m an avid snow skier. I make it out two or three times a year, believe or not even being from the South. I go fly fishing and hunting on regular occurrences.

Most recent project, I also renovated a house up in North Georgia, so just side projects like that.

Mike: Oh, wow. Yeah, I’ve done some renovation to our house here and there. That’s a lot of work.

Paul: It is a lot of work, but it’s one of those projects that just keeps you busy and on-task. I like projects that have a completion. You get to see your build at the end of it and say, “Look at what I accomplished.”

Mike: Yeah. What are you most proud of with that renovation when you look at it? What makes your eyes sparkle?

Paul: Well, the crazy thing about it is the initial look of it. It was a house built by the National Forest Service back in the 50s that was a cinderblock block structure to begin with. Somebody came along about twenty years later and added a brick extension to it, but they still left the cinderblock on the end. It was a complete gut, and redo, and refacing of the house. Just curb appeal and the way that it looks now is completely different than when it started and the inside is just so much more livable. It’s really rewarding to see something change from something you, kind of, cringe at when you walk in to something that your eyes light up – that’s pretty rewarding.

Mike: Very cool. Now, tell us, the Director of Solutions Engineering – what does that mean on a daily basis for you? What’s your typical day like at Internap?

Paul: Sure. As Director of Solutions Engineering, I’m pretty much a player/coach. Right before I took this role at Internap, I also was the channel solutions for the nation, so I still have national presence with our channel program. In addition to that, I have several SEs that report in to me both in New York, the south east and our east sales team that works nationally, and I still support our channels guys.

I get in the weeds with our deals. Anything that gets complicated, anything that’s large, I’m going to be asking my SEs questions or have my feet on the street to talk about it – talking through the solutions, how to build them, ways to configure things that might work better, be more efficient, and just thinking creatively to get things done. So, my day to day is I’m either… I’m generally in one of our markets, and working with our teams hands-on, and our AEs on the front line.

Mike: Fantastic. So, one thing that came up in your explanation there is you mentioned the word “channel.” I thought it would be a good opportunity… I haven’t done this yet on a podcast – I know that there are a lot of listeners out there who don’t really understand the “channel” and what “channel” means. From talking to a lot of customers over the years, the only really sales avenue they, typically, are familiar with is the direct channel. Can you tell a little bit about your channel program and how that is different than your other sales team?

Paul: Absolutely. Here at Internap, we actually take an approach that channel is really lead generation. What that does for us is it allows our AEs to work directly with our channel partners as a partnership, so there’s not competition per se.

The way that it works in the channel world is, obviously, there are vendors out there that are consulting on services from across the both. They might be bringing two or three different vendors to the table or do the initial qualification and say, “This vendor is going to meet your needs best, let’s talk to them.” They, kind of, steer the conversation, and a lot of that is from experience on their end, on the consulting front, or just historical knowledge. They bring those leads in the companies, and work with the companies to get good opportunities and make the right fit. When those leads come in with us, we have an East Coast and a West Coast channel manager/channel director that then steers that out to a local team, so that we can put feet on the street with that channel partner so that we can be most effective and give that deal a lot of attention.

It works really well for us. There are several different models that people take. Sometimes, there are channel managers that are in competition with direct sales forces, there are some sharing that goes on sometimes. Ours is more of our channel managers/channel directors work the relationships with the managers just to know that our products are out there, deliver training, those types of things. As the leads come in, they put local resources with them to make sure that we’re being as effective as possible.

Mike: Yeah, that was great. Thanks for that explanation. I think it’s something that needs to be said again and again. Like our company, for instance, we have been a broker for telecom and cloud services for thirteen years. Everybody is always asking these tons of questions when they find out we’re a broker. If I’m talking to an IT decision-maker, it’s always like, “Well, how does that work? Do you guys get different pricing?

In our industry, we call it the “partner channel” or the “agent channel.” The “indirect channel” is another word too, but I always hated that word because I feel like the customer is still buying directly from the service provider, so we’re, technically, not indirect. It’s, kind of, a negative word that I’m like, “Ah, I don’t like it.” But, it’s great to talk to people like yourself who are within the service provider word who can explain it a little bit more instead of just us explaining it all the time.

In essence, the way that it works for you guys is that, as a broker for these services, one of our sales people would get approached by a client and say, “Hey, we need to know which cloud service providers would be a good fit for us.” Our sales person would say, you know, “Hey, Internap is one of those providers. Let’s get you on the phone with them,” and then they’d get in touch. What you guys do, you guys team up with your on-staff sales people with our sales person, so they, kind of, team approach the solution, correct?

Paul: That’s correct. We look at teams like yours as a force multiplier. If we’re fighting against you, then… We have some sales people that are being competitive, but if we can help you guys work for us, then it makes us much more efficient. So, yeah, we have our local team work directly with your sales people to bring a solution that works with everybody.

Mike: Fantastic. Great. Okay. I won’t dwell on that too long, but I just want to make sure we clarified it a little bit. I hope that was valuable to everybody listening.

Let’s jump in to your topic. Today, as I mentioned earlier, you are going to talk to us a little bit about the hurdles to deploying bare-metal versus traditional cloud. I’ll just turn it over to you and let you jump in.

Paul: Sure, thanks. There are, really, two topics that come to mind with me. Bare-metal has been something that’s been around, but even the name itself is somewhat misleading as to how it’s performed and the way that it’s deployed.

There are a couple of things to keep in mind. Most people deploy cloud with virtual instances, which means that you come out and you say, “I want X number of CPU. I want X number of gigs of RAM. I want an X amount of storage with each machine.” That works really well when you have a pool of resources that you’re allocating and requesting out of that pool, and you can really specify what that is.

Bare-metal plays a little bit differently. Bare-metal servers, essentially, are public, on-demand instances, but they are physical servers that are already built, racked, and stacked. The configuration doesn’t change because it’s a physical configuration. You’re simply saying, “I want item twenty in the rack,” and you turn it on, and it deploys to your network and everything like that.

You don’t have a whole lot of options to get variation in those, at least, in a timely manner, as in, on-demand deployment. Most providers will, certainly, go in and take those servers out of the rack, break them open, and put in specific amounts of RAM or change out drives, and that’s fine. The difference there is you don’t get a server at the click of a button. It’s the click of the button and then you wait for the order to be provisioned.

That really takes away from the whole advantage of bare-metal. Number one, you’re introducing a human component to it where there’s somebody physically having to do labor. That gives the effect of additional cost to the provider, so, at the end of the day, it actually runs your cost up. What keeps the cost down so far on bare-metal is all the servers have been racked and stacked at the same time. There’s not a lot of hands-on work that has to be done with it, so it’s really low overhead for the provider to give you the services that way.

With that said, there are a couple of key things that I would point out to clients when they are deploying in those environments. The first thing is that they’re standard configurations. You look at somebody’s standard configurations and they may not be exactly as you asked for. You might say, “I need X, X, and X,” and it might come X, Y, and Z. What happens then is you could either deploy what is there or you could put in an order to have something deployed for you manually, which extends your lead time to getting that server provisioned for yourself.

My first suggestion and key to deploying bare-metal is looking at all the configurations that your provider offers and testing your application across multiple ones. When we talk about testing with bare-metal, the nice thing about bare-metal is providers will deploy these on, sometimes, a per second cost or an hourly cost. You don’t have to commit for a long period of time. You can pay $10 and test for a week, as opposed to having to buy a server or sign in to a contract. Once again, these things are racked, and stacked, and ready to be provisioned whenever.

You can take your application and look at the different configurations. Let’s just say, for example, you have the option of an E3-1230 V3 chipset versus a E3-1270V3 chipset – the difference is a couple of cores. The actual RAM or storage might be the same. You might get 32gb of RAM or 64gb of RAM on both systems, and they might both have a single or dual set of 1Tb SATA hard drives. You want to know what the difference is in performance on those two applications. Maybe you prefer to have six cores, but if you test against four cores, you might find that you can also get away with, maybe, a little bit smaller workload.

Customers in these bare-metal deployments are often scaling horizontally as opposed to vertically. How many machines they can add really determines their load versus how much resources on a single box determine how much they can accomplish with a single machine.

Mike: Okay.

Paul: We like to see our customers test across multiple configurations. It might just be a difference between SATA or SSD drives, or it might be the difference between 32 or 64gb with the other components staying consistent. If you test across them, it just gives you the versatility to say, “Okay. The vendor ran out of these types of deployments. What are my other options?” Then, you can quickly make that decision as opposed to having to test while in stream to try to get things that are readily available.

Mike: Yeah, that’s great, especially the point you made – the pricing works almost by the second or by the hour as opposed to having to commit for a long time. I think that a lot of people probably aren’t aware of that and they’re thinking, oh, gosh, you had to really commit and you’ve got to get all in if you want to do that, as opposed to just, kind of, like, throwing something up and testing it a little bit before you go that route to see what your options are should something happen or should you need to deploy something quick.

Paul: Right. That, kind of, leads into to how these things get priced out a little bit. I mean, if you don’t commit and you go in for a hourly term or seconds term, you’re going to pay almost the list rate in almost all cases. If you sign up for a month-long commitment, year commitment, multi-year commitment, you can drive the price of their servers down with many of the providers.

The other portion of that is there’s an area where you can use a combination of both. You can set up what’s persistent, what stays around in your production. Then, you have a variation that you might pay for on that hourly or second basis when needed, but you have a baseline of services that you always pay for and have in a longer term to get those costs down for yourself.

Mike: What thought comes to my mind is if I’m an IT professional or a sysadmin running an IT department right now or an IT manager is: What applications would I have that are sitting there that do very well in a bare-metal environment? Some people out there might be thinking, “Okay, that’s great,” but, like, “I wonder if we should be considering using bare-metal for anything or expanding some of our existing applications and move them to bare-metal. What are some of the things you’ve seen that do very well in that environment?

Paul: Yeah, absolutely. I’d say that the things that do well in that environment are things that really require consistency. The biggest problem people have with using virtualized cloud is that they are either affected by oversubscription, or thin provisioning, or interruption from noisy neighbors, those types of things. Whereas, when you use a bare-metal cloud, there’s no interruption from a hypervisor and your neighbors, the only thing they, technically, could really affect is the network access to the machine. The machine is going to compute at the same level because they’re dedicated resources to you that aren’t being touched by anybody else.

For a lot of people it really, kind of, comes back to knowing what your application could really do. For some, it’s taking multiple web servers and load balancing across those. As your website picks up traffic, adding additional servers horizontally to scale for that. Anybody that has any seasonal or reactionary type business that drives people to their website – maybe you start a campaign from a marketing standpoint, maybe your business is seasonal from a perspective of, like you, you work for Superbowl commercials, or sporting events, or whatever else – your loads get up and down based on that. You probably have a consistent baseline that sticks around, but going up and down beyond that is certainly a thing.

Then, the real area that, really, is the big business here is big data applications. Big data applications such as Hadoop, the unstructured NoSQL databases, they just require resources. They don’t really care about redundancies or how you put them in, you just literally throw all these resources at it and it finds the best way to utilize it.

The configurations for such are generally found in the bare-metal cloud, so I’m not sure how familiar most of the viewership is about what a Hadoop server, generally, looks at. It’s a large multi-core machine, so probably like an E5-2620 / 2630 / 2650 / 2690 even, and a large amount of RAM, 128gb or 256gb of RAM, and then just a bunch of dumb drives for the most part. You load up a machine with 8/10/12 drives and that’s the best configuration, no RAID, you just put it out there, and Hadoop comes and finds those systems, and utilizes that storage, the processor, the RAM in the most efficient manner it can. You’ll commonly see those types of configurations on there that most people will look at and be like, “Is that a large storage machine?” and most of the time, no it’s not. It’s actually based for some of those big data applications and NoSQL databases.

Mike: Awesome. I think that’s great because I think that frames it a little bit for the listeners to, kind of, go, “Okay. Well, if we’re thinking about doing these types of things or have these types of things right now, virtualized in just a public clouds, these are some of the symptoms that we might see that might lead us to thinking about experimenting with bare-metal.” That’s great. Thanks.

What’s the second hurdle?

Paul: So, the second topic is really how bare-metal servers have been pinned to the virtualization of the public cloud. What I mean by that… I mentioned earlier that bare-metal cloud, I’ve always felt, is a little bit misleading term, but they came out with it a long time ago before I had any say. Bare-metal, to me, means a blank machine with no operating system, I can put whatever I’d like to add on to it. But, in most cases, you generally select an operating system in your deployment when you’re going through your online GUI to deploy your machine, or if you’re doing it through APIs you say, “I’d like operating system with…” maybe it’s Redhat, or CentOS, or even Windows 12, something like that. You generally select that from a dropdown and by the time your machine provisions, it already has an operating system. The systems have actually got more advanced and allow you to really orchestrate what that operating system is and, in many cases, it’s allowing you to deploy your own ISO onto those machines, so you
can deploy whatever operating system you’d like in a true, kind of, bare-metal format.

Those orchestration engines by each provider really vary. A lot of it has to do with what they’re using for their virtualization platform on the other side. Here at Internap, we use OpenStack and KVM. If you look at somebody like AWS, they’re going to be using Xen that they’ve really taken apart and made their own, so it’s really proprietary than anything else. You go look at Azure and they’re going to be using Hyper-V.

Each provider is going to have a different type of virtualization platform that sits behind it and that really determines what type of image you would need to have to be able to deploy onto their environments. Keep in mind your image format. Qcow2 is what KVM and Xen generally use. You can also use QED for KVM. Some providers have converters that will allow you to put in Raw file formats. There’s VDI type of image format for VirtualBox, VHDs for Hyper-V, VMDK is what everybody is familiar with when you have VMware.

Keep in mind before you deploy or before you put up your image, understand what your provider’s technology is because certain formats of ISOs will be supported and others will not. Some might be able to be converted from and some will not. Look into that when you’re choosing a provider. If you are VMware shop at home, know that you’re going to, basically, be pulling VMDKs if you’re building those images in-house, and that they might have to convert to something else when you’re deploying them onto bare-metal servers. A lot of that is largely based around what that company had founded in their virtualization side of the house or their orchestration side of the house.

Mike: Have you found that that affects a lot of IT departments’ choice in terms of service providers that they use for bare-metal or is it, kind of, more like they can make it work still? Is that, kind of, a big show-stopper?

Paul: Most of those images are very easily converted from one to the other. I say “easily,” there are tools out there or some systems – generally, your provider will help you with it. It’s more of just an understanding that that hurdle might be there. If you create an image and need to deploy in a short time period, know that you might have to go through some extra work that you weren’t expecting. All these things that, in my opinion, affect how customers bare-metal deployments are really about time to deployment. These are just things that you can, kind of, get out of the way ahead of time so that you have the easiest deployment possible and quickest to deployment possible that you can reach by just crossing your Ts and dotting your Is.

Mike: Those are a couple of good tips and a couple of good pointers. I hope everybody got something good out of that. Definitely, thank you for sharing those, Paul.

Paul: Absolutely.

Mike: Okay. Now, we’re going to get into something fun. As we always like to do, tell us about the funniest or most interesting thing you’ve ever experienced in the workplace.

Paul: You and I both talked previously that I used to work for a company called Cbeyond and I started there right out of college. A lot of good stories about the sales force there, but some not meant for this call. I’ll give a pretty tame one. About just over two years ago, when I was leaving there – our offices were based in Atlanta. I’ve always been based in Atlanta and everybody knows that we don’t handle winter weather very well.

My team came in for the day and there were some weather reports that things could get bad, but they didn’t let out any schools, nobody really called off work. I worked most of the morning pretty diligently, and then around 10:30, they started making announcements that schools were letting out. I told anybody on my team with kids, “Hey, you’ve got to get your kid. Go pick them up from school and go home.” That really happened closer, probably, to 11:00.

Then, about 12:00 rolls around and not a single speck of snow had hit the ground at this point, but 12:00 it starts snowing – stuff starts sticking to the ground slowly. I’m, like, “Alright. Well, maybe I should send the rest of my team home.” Sure enough, I send my team home, and five minutes later, and email comes out and says the offices are closing down.

Within an hour, most businesses in Atlanta had sent everybody home and schools had also required most people leave. So, Atlanta has terrible traffic in the first place, but our traffic period start from about 6:00 in the morning and end about 10:00 for rush hour in the morning, and then they start at around 3:00 and end at about 6:00 or 7:00 in the evening.

Mike: Oh, my gosh.

Paul: So, we stretched it out over a three or four-hour period. Well, everybody hit the roads in Atlanta that day at the same time. On top of that, within about an hour and a half, there was snow on the roads and cars were driving over it, so it was turning directly into a sheet of ice.

Me, personally, I sent my team home and then things got really quiet in the office, and I got a ton of work done in about an hour. Well, I picked my head up from desk and looked outside, and everything was white, and I decided that I also needed to head home.

Our office, at that time, we’re at a place – you could reach the interstates, for me to get home, in about three or four different ways. I got on the road and after about three hours, I tried three or four different routes and I ended up about a hundred yards from the office. Cars were, literally sliding backwards down the hills and whatever else, so I pulled into a parking lot and walked up the hill to my office. When I got there, there were about a hundred other employees that were stranded there for the night.

We had emails from our facility manager who said, “You know what? You guys are all stuck there, raid the fridges. Whatever’s there is fair game.” We went to the gym, we took showers, we took the towels out of the gym and used them as blankets that night. We had our own little Maze Runner that night – all of us just isolated in the offices. We had a sleepover lock-in that night, but several of us got to know each other pretty well. It was a good time.

Mike: That’s pretty cool. I’m sure that was a good team bonding moment. Did corporate enjoy not having to orchestrate…?

Paul: No, it was a good time. The people there were great. That was actually a great company to work for. As you and I mentioned earlier, the sales force there has some good and bad reputations. Good sales training, some less than, you know… things that you probably wouldn’t talk to everybody about on the other end of things.

But, we always did silly things with sales. We’re hitting fifty to sixty orders a day, I’d go out and walk into an office. If nobody’s there, our goal was to walk straight to the back. You would start talking and saying people’s name, or saying, “Is anybody here?” but you’d walk directly down the hall till you saw somebody because the goal was the CEO or President was, generally, the last office on the hallway. If you could walk all the way into the back door and that was the first person that caught you, you were in good shape.

The other thing we did is we used to walk in the lobbies and if nobody’s there, we’d pick up the phone and call 411 – that would tell us who they were using as a provider. You know, “Thanks for calling AT&T, how can I direct your call?”

Mike: Right.

Paul: We used all types of secrets like that, but it was a fun, enjoyable experience that taught me a lot and really toughened my skin for the business world.

Mike: Yeah. Very cool. Like you and I mentioned, the environment there, I think, it was, pretty much, all people graduating right out of college. It seemed like 90% of your sales force was in their first six months right out of college, so I’m sure hanging out – there’s a lot of energy that night when you guys got snowed-in. Everybody’s running around doing crazy stuff and trying to think of interesting ways to keep yourselves entertained.

Paul: Yeah. For me that was year eight at Cbeyond, but in my earlier years, you’re right, the sales force were very entertaining.

Mike: Awesome. Well, great. Thanks for sharing. Okay. Well, I wanted to give you an opportunity to talk a little bit about Internap. We’ve heard some great stories and, obviously, got some good info, but tell us a little bit about where you’re working now over at Internap and what you guys are doing that’s got you excited.

Paul: Sure. For me, I feel like Internap is continuing to push the envelope in the OpenStack world. We went to the conference for OpenStack in April in Austin, Texas. A lot of the questions that we had were around how we accomplish things with the module that’s called “Ironic.”

Ironic is the bare-metal module within OpenStack. What we found is that we have solved a lot of problems that people are running into, some of which we haven’t even published out to OpenStack or committed to OpenStack yet. But, people are running into a lot of issues that we addressed six/twelve months ago to be able to deploy our OpenStack bare-metal cloud.

The consistency of OpenStack across both our virtualized and our bare-metal platform really sets us apart. It means that you can come and consume virtual instances and also consume bare-metal all on the same VLANs, and all within the same deployment, using the same set of tools. If you want to tie in with APIs and whatever else, it all ties into the same orchestration tools, so you’re not working with multiple environments or segmented environments to do that. So, we’ve done that really well, and we feel that we’re really pushing the envelope and helping the OpenStack community really develop that Ironic platform for bare-metal.

The difference between us and some of the other providers is we’re also pushing from our managed hosting products where we rack and stack servers in a custom format for clients… We come in, we stack a bunch of Dell servers, we’ll give them a couple of Cisco firewalls, put a dedicated NetApp SAN behind it – that type of custom build where clients are really giving you a bomb, and you’re building out materials and building everything custom. We’re really pushing those into our OpenStack deployments as well.

What we’re finding is that we’re able to offer some of those managed hosting experience type things in our public cloud offering as opposed to a lot of people that are doing their own home-grown… All that’s home-grown to us as opposed to other providers that are building out and supporting on other people’s clouds. You have a lot of providers out there that are offering management services on AWS or managed services on Azure. They don’t have a whole lot of control over those ecosystems.

For us, we’re trying to push our managed hosting customers into our public cloud and managing it all in-house. We’re working on things like compliance, the flexibility of deployments to offer more VLANs, more networking options, more security options. Those things are really developing rather quickly with us.

On top of that, we just have a really good global span. We have Asia-Pacific regions, we have West, Central, and East Coast deployments, we have a Canadian deployment, we have a European deployment as well for OpenStack, so just a really vast product set that spans the globe. That’s all leveraged on top of our MIRO technology, which is our IP technology that does low latency delivery over the public internet. That was our home base and what we really built ourselves off for years, and we’re just using it to leverage and provide a performance experience for all of our hosting products.

Mike: Fantastic. Now, if there is such a thing, is there, like, an ideal client that you say is like a perfect fit for Internap?

Paul: I would say that the majority of our clients have a performance aspect to what they do. Some of the things that we do unique, obviously, the MIRO product that gives a low latency delivery, but we really built our hosting products around that same theme. We have series in our virtual machines that are not oversubscribed at all, so if you order a core, you get a core, you order a RAM, you get that RAM.

We do have a series that is slightly oversubscribed, but it’s like a 1:3 ratio. Most providers out there, you’re looking at something like 1:8 or 1:6, somewhere in there. We really give dedicated, performance type deployments for our clients. Bare-metal also plays along the same theme – no hypervisor interruption, no noisy neighbor type problems, so all of those things really fall in that performance theme.

We have clients that are smaller offices – twenty to fifty employees – that just have a real dependency on performance IT applications. They don’t really have the time or want to manage those in-house, but then we also have Fortune 500 companies that come and use our services to get the same advantages. So, we’re, kind of, across the board, but I’d say the theme is really around performance. Adtech, gaming, high-frequency trading, anything where people are coming to our data centers to consume services or where content is being served from our data centers, we do a really good job at.

Mike: Very cool. Yeah, I think that gives a great description and roadmap to, you know, just to at least plant a seed as far as when Internap might be a good fit for any of our listeners. If any of you think that Internap might be a good fit for anything that you’re doing, just give us a call over at AerocomInc.com and we’ll get you in touch with someone like Paul or one of the folks that reports to Paul, and get a conference call going.

Well, great. Thanks for joining us on the program, Paul. I think it was very informative and I thought you gave some good information. I hope the listeners got as much value out of it as I did.

Paul: Thank you, Mike. I appreciate you having me on the call today.

Mike: Alright, have a great day.

Paul: Alright, you too.

Mike: Hey, that was Paul. What did you think? Wasn’t he great? I thought that the sirens added some extra personality in there, let us feel like we were right in there in the center of what’s going on in New York City.

Before you go, I just wanted to quickly remind you about the free gift, the twenty-five CloudVM questions that you can ask your IT department about your company’s needs before you shop CloudVM providers that will help you get to the best CloudVM provider a lot faster and let you do a great job shopping. All you have to do to get is text the word “CLOUDVM” to the number 44-222, and we’ll send you that document right away. Again, text the word “CLOUDVM” to the number 44-222 and we’ll send you a soft copy of all those great questions that will help you shop so much faster for a CloudVM service from any provider.

Alright. Well, have a great day. I hope you enjoyed the show and I’ll catch you next time.