Cloud Trends and Cross Connects

Mattias

So welcome, Neil. So, for the people that don’t really know you, who are you?

Neil

Currently, I am the Software Engineering Manager for Cologix, but I’ve been a lot of things in my career. I guess I started as a sysadmin. I worked in pen testing and compliance management for a while. I was a virtualization and cloud architect. I was a marketing manager and made my way to product management for a time. Now I run software engineering, automation, customization, and primarily, the network exchange for Cologix.

Mattias

That’s an interesting background. For the people that don’t really know, could you just explain what Cologix does?

Neil

At the simplest, it’s colocation data centers. In a wider swath, Cologix is a network-neutral data center company that focuses on highly interconnected colocation. I’m trying to specialize there a little bit because there are a lot of different types of data centers and a lot of different types of colocation out there. What we specialize in specifically is highly interconnected sites typically built around a carrier hotel, and this whole model of carrier hotel colocation. That does mean it’s a little bit different than maybe some of the standard consumer, mid-level or even some of the data center providers out there that others may know that specialize in maybe just large sites. We do a little bit of everything, but it’s all based around network and network connectivity.

Mattias

How did you end up there?

Neil

I actually got a call from an old colleague who said that there was a wonderful growing colocation company, and they had an opportunity. I was in product management at the time working for OVHcloud, a French cloud-based company. I was one of the product managers and I helped them to launch in the United States and expand over here in North America. I had a wonderful job. It was a wonderful company, there was a lot of good going on there. But I was traveling a good bit between Europe and the US, spending more time away than at home. And they said ‘Hey, we have an opportunity here in Denver where you are based. It would be in North America to focus on the growth cycle of another company and start to get into physical assets rather than the virtual world, where you’re managing cloud resources and cloud assets.’ And it sounded like a really interesting opportunity. I jumped at the chance to spend a little bit more time at home and to get to work with some great people that I’ve worked with before and to focus on real physical locations and being in the data center space, with physical networking being a big part of this. But I was starting to bring my cloud world and what I knew from cloud and that experience back into it.

Mattias

That kind of leads me into the next topic around the cloud. Because I know there are a lot of listeners who wonder, ‘Where is the cloud?’ I guess that’s one of the major questions. How would you define the cloud itself? Where is the cloud?

Neil

It’s everywhere, and it’s almost nowhere at the same time. It gets to be very difficult in some ways because you do have this idea of regions. You can have your services managed over a large region, but you never know exactly where it is running and where it’s processing at one point in time. You may have a market, this idea of a market, where it’s a city, and maybe the surrounding suburbs, where you’re generally in that area. But getting much more specific can be hard in some areas. When you’re connecting into them, and when you’re working with cloud, really, your anchor point is typically what we call the onramp. The idea of this is the network edge, the point where your network stops, and the cloud network begins. And typically, there are not very many of those. That’s why I say it’s not that available at the same time. It’s everywhere, knowing that you could have a Northeast Region – you are somewhere in the northeast, but for privacy and security concerns, are not going to let anyone know. But maybe your onramp is in New York City. And there’s one or two and those are very hard to get to, there may not be many of them around. They are valuable resources, but that’s your physical point that is advertised very well – to be in a specific facility in a specific place. And that’s how you can interact directly with the cloud when you’re not going from a public side. It’s a little bit of everything.

Mattias

Do you feel there is a trend for more and more people to ask where the cloud is? Are people starting to become interested in where their data is stored? Or is it this old feeling that the cloud is in the cloud, and you shouldn’t care where it is?

Neil

I think more people are starting to ask because everyone’s moving toward a distributed model. It’s been decades of going from a hub-and-spoke type of model of keeping everything around one central hub, one central location. That worked well for a really long time for a lot of architecture. But generally, we’re seeing the industry and businesses move closer to customers. This is the whole conversation of edge, which has been around for a while, just distributing and to be distributed. To do that properly, you need to know where you’re putting your systems, where your customers are, how you’re moving, how you’re processing, and moving things closer to your audience, whatever that is. But you’re also trying to keep more control of your data, which has been a conversation now for the last decade too – where’s your data? Are you keeping control of it? Or is it getting out of your hands? Which has been all part of the hybrid cloud architecture – that split of putting what processing you need, what services you need, close to your end users and keeping your data and your highly valuable items in a private area close to that, and then using both of those in combination to reduce latency and reduce your risk. And to keep control of what you need and to let up control of what you don’t, so that you can let someone else manage their specialty, which is cloud. You can use generalizations of that to say, ‘I am generally going to be around the United Kingdom.’ But it would be better to say, ‘I am specifically in a cloud in downtown London or just outside of that.’ And based on that, I can manage my private services and private data that I need to house. I can manage my connectivity and my latency to my end customers, who are going to be generally on the northern side of Europe. You don’t need to know to make those kinds of decisions.

Mattias

What would you say is the biggest challenge facing cloud today?

Neil

I think the biggest challenge right now is all the cloud providers have forged their own paths. They all have their specialty and their differentiation. We moved beyond messaging of one cloud to rule them all – this idea that you can only work with one provider, that “Hotel California,” the vendor lock-in, those kinds of conversations. And everything has moved over the last five years to multicloud and multicloud management. And I think the bigger problem that comes with that is the complexity. AWS is architected and has different value and has different requirements than Azure, which is very different from Google and different from Oracle. All of them have their own methods of connection, their methods of usage and their own security model. They’ve all differentiated in their own space. I think the complexity is actually one of the largest challenges that we have today because there’s not necessarily a uniform standard between them for your usage. And companies are being pushed more and more to use the best tool for the job. That really pushes them to have a multi-tool work bag, which gets to be too complicated, and it requires highly skilled staff and highly skilled, trusted advisors through the process. That starts to complicate your own resource pool as a business as you’re trying to work with this wide swath of cloud providers. I’ve just named the Big Five – there are hundreds of smaller regional providers that you can work with as well that get even more complicated. Finding the right tool, finding the right trusted partner and working with them to get the right solution is probably going to be the biggest challenge as people are sorting through the weeds.

Mattias

Do you feel the trend is that more and more clouds are popping up, challenging the big ones? Or are the big ones really driving away from everyone and will become the dominant factor in the foreseeable future?

Neil

I think they’re absolutely driving away from the smaller regional providers. There’s a consolidation in a way. We see it in the data center world, where a lot of those smaller individual data center providers may just have one facility or two facilities in market – they are starting to roll into the bigger companies. There are a lot of mergers and acquisitions and growth through that way. Now, the larger cloud providers aren’t necessarily buying smaller clouds, but they are offering more value at a more competitive price in more locations, and they’re starting to become a better value than some of the smaller players. There is, in some markets, diminishing choice in some ways where you’re not finding so many hosting groups that are managing really hosted private solutions. There are a few out there, but public cloud has just become so synonymous with five providers this way. You’re really just looking at AWS and Azure very much in the lead. We see them coming up with the majority of the conversations. Google has been accelerating in the last few years. We’re seeing quite a bit more from them than we have in the past, and a lot of development specialty over there. And then we’re starting to see more and more from IBM and Oracle, with IBM coming through the second wave with this 2.0 of their platform since the consolidation of their cloud acquisitions. And then Oracle really launching with their latest cloud platform has been catching up in some ways. But they still have such a specialty in the market that we’re not seeing a whole lot from them quite yet from our customer base.

Mattias

How’s your view of cross connects and their importance? How do you treat that in Cologix?

Neil

Cross connects are the core of the business today. That’s a lot of what I’ve been working to enhance over the last few years and try to move from physical cross connection over to more of a virtual platform – bringing that cloud experience over to interconnection. Interconnection is incredibly important. No business can survive on its own, especially in a virtual world. You’re going to need to connect to other providers, and you’re going to need to do so in a secure way. The best way that you can do that is by coming to one location where you can all meet at the table, and you can share data and services back and forth. And this could be between you and your bank, or you and cloud services. You and your network service providers who are helping to host your website and your other services that your customers connect to, as well as your back end users within your company that your internal IT is serving – all of those have different needs in terms of storage and security and networking. Interconnection is the way that you bring all of them together. The cable, the connection – the one hop between you and another provider – you want to keep that to a limited number of hops to give your business the best performance, the best security and the best availability, to reduce the risk between any of your private IP, your data, your services and your transactions. You want to limit how far that has to go to get between you and the end provider that you’re working with, for whatever the realm of your businesses. And because this is a physical asset, it’s not going to be universal. You’re not going to be able to connect to everyone everywhere at the same time, in the same way. Getting close to as many as possible – to make sure that you’re using something like an interconnection option, fiber interconnection, just a cross connect between you and your other providers – can give you an edge in your business, and can reduce the amount of downtime and other issues you could run into by using longer metro services or bigger connection MPLS rings and other services where you’re going through multiple hops, through multiple areas, to multiple services to get to who you need to get to. The chance of a problem between you and a provider with one cable? Really low. Interconnection is key to a huge number of businesses that need to control for that risk, need to control for downtime and want to be located right next to their provider. We’ve been talking about cloud a good bit and that’s a big part of my background. But if you have your business and two cabinets on a data center floor, your servers, your switches, your storage, your business is running in a core with all of your processing your data, your customer info right there, and Amazon is one cabinet over. If you can connect to AWS with a cable that’s going to be a few feet to get into them with a direct connection, that’s probably going to be the best bet for you and your business and your services, rather than having to jump say 500 Km away to get to them and dealing with that latency and that equipment and all the things that can happen. I hate anxiety-based messaging and I’m not trying to do that here. But being a former sysadmin, being in that position, wanting to manage the risk and the availability and uptime and having that be in IT need, reducing the distance between you and providers is a key way to do that. And interconnection is the best way to manage it. An area like a carrier hotel or an interconnection hub, where everyone can meet to share those connections, is the best bet to do that with the widest number of people possible. It’s a huge value to me.

Mattias

At Cologix, how do you define where your next data center is going to be?

Neil

We focus on that carrier hotel. Typically, within a city, you’re going to have one, maybe two locations where the majority of fiber is going to meet. It’s typically a single building. It’s going to be downtown. It’s typically a skyscraper that most people drive by every day, and they don’t realize it.

We have a facility in Vancouver that’s based around 555 West Hastings, that’s the street address of that common carrier hotel. It’s a skyscraper called the Harbour Centre. It’s been there for ages. It has a ton of businesses in it. But this is where all the fiber in Vancouver comes together in one location. One building meets there and then can go up to a number of suites within the building. We focus on sites like that.

We build our business around that and then start to spread out to help customers get more connectivity to it. Because if we’re using this example of Vancouver, this is one building and all of the fiber is going to this one location, and in one of the most expensive real estate markets in the world. And there’s not a lot of floor space even in this skyscraper for businesses to operate. You can only squeeze so many servers, cabinets, systems and even cloud providers with nodes into this one facility. You have to start to tether that out. And it’s this idea of starting to have concentric circles.

We focus on the inner circle first, which is going to be this carrier hotel. Where’s the fiber? Where are the carriers? The telecommunication? Then we start to build what we would call an annex facility around that. Is there a good availability in that market for us to start to have other sites and other buildings downtown? And then starting to get out to the edge of the suburbs there in the circles. And then can we move on to out in the suburbs and just a little bit further outside of town to where we could fit hyperscale facilities?

That’s where we’re starting to get in the range of 20, 30,40MW for a site. These are massive, massive locations where you can fit tens of thousands of servers. Focusing on that center point where you can have a lot of fiber and a lot of connectivity, and a few customers can fit in there – we’ll do our best within that location. And then really expanding that out to get more people in at a lower price point where you can have that annex. You’re not in core downtown, but you’re just a little outside of it. There’s a little bit more floor space. It’s a little bit more in that mid-range for those customers who need it. And then getting out to the hyperscale facilities for the people that really need tens of thousands of SQFT to run what they need to do. And they need to still connect and get to that carrier in a reliable way and be able to manage from there.

When we’re building our sites, it’s all around that story and all around that message – helping customers with that connectivity and getting them to carriers and other major fiber carriers and providers that are in market, that are regional and local. That draws everyone else in, which builds an ecosystem. That’s really a part of our play – everything rides on network services in our backbone. You have to be somewhere near that.

That brings in all the businesses that rely on it to manage that risk, like we were talking about earlier. That brings in cloud providers to work off of those services and work with those clients, which then brings in more providers that want to work between those and play together. And before you know it, you have a really rich ecosystem of different providers and services that are all interconnected, all working together. And typically, just one or two hops away.

Coming soon!