Industry Spotlight: DE-CIX’s Ivo Ivanov on the Future of Interconnection

January 30th, 2018 by · Leave a Comment

The world of interconnection has become far more complex than the simple peering and transit relationships that arose in the days of the early internet.  Looking ahead, given the rise of the cloud and the expected flood of data from IoT and 5G data in the years ahead, there will be a lot of moving parts.  With us today to help figure out how it will all shake out is Ivo Ivanov, Chief Strategic and Corporate Development officer at DE-CIX.  Over the last few years, DE-CIX has expanded aggressively outward from its giant Frankfurt IX into markets around the world, and is rapidly transforming its own model in preparation for the dynamic future ahead.

TR: How do you see the interconnection business changing over the next few years?

II: The key thing for us as an interconnection service operator, is that enterprises are now confirming that they see interconnection needs.  The trend is that with just one operator as service vendor they cannot meet their needs for their global communication, for new services, or for the digital demands of their business models. They need to be directly involved with their networks choosing cloud solutions that are directly interconnected with suppliers and business partners. So that means that the existing interconnection ecosystems of today can be dramatically extended to a much broader group of participants. Today, we see mainly two types: hypo-access and content, and content delivery is also there. While we do see a limited number of enterprises, this will change in the near future dramatically. We already have requests and active discussions with big companies from the automotive industry, from the finance industry, and from IT system integration side.

TR: How do different enterprise verticals approach the interconnection business? 

II: The finance guys are more focused on security and on better aggregation worldwide with their providers of finance-related data, where they do have already interconnections with us today but they are not aggregated enough so this can be improved.  Providers like Thomson Reuters or Bloomberg, for instance, have a very strong interest in serving their clients in an optimized, interconnected way. The automotive guys are more interested in interconnection services very close to the edge.  For instance, interconnecting cars in live traffic situations, you need a significantly better resilience and the lowest latency possible, actually. It only makes sense if the delay is about one millisecond or two milliseconds.

TR: How does DE-CIX plan to meet the interconnection needs from enterprises?

II: DE-CIX started operations in 1995, so we do have 22 years of experience in the interconnection business. We collected a lot of lessons learned, not just in Germany but also in markets like Dubai, Madrid, Marseilles, Palermo, New York, and Dallas -- all different markets, regulatory frameworks, business and market structures, and cultures. All those lessons learned have given us a better chance to analyze properly what the key drivers are for interconnection or creating digital hubs.  We believe that the ecosystems created for peering are natural fundament for further interconnection services because they are a natural biotope for the interconnection business. Even if you look at it from the pure infrastructure perspective, the one interface network operators have or the bandwidth they use for peering can be used also for additional interconnection services in a more flexible way so the participants in those ecosystems can decide how to use the infrastructure. This leads to more flexibility on their side to a better utilization of infrastructure in general and to more efficiency and effectiveness. We'll turn the simple peering port from the past into a multi-talented, multi-service access to an interconnection fabric.

TR: What types of services will we see evolve?

II:  Peering is still very important and it will stay there, but is already implemented of course. Next there is cloud exchange functionality, for which operators can just literally slice the interface connecting them to the platform. In different VLANs they need to use a variety of different services.  One is cloud connectivity. We have already certified 22 cloud service providers including all big guys but also a variety of different specific cloud service providers. All participants on the platform can easily, seamlessly create a link to those cloud service providers and get the cloud connectivity they need for their business. Or they can sell this cloud connectivity to their end customers if they have this need. Another we have in mind is a marketplace for transit services also.  If there are parties who want to buy or sell a transit, they can do this on the same port.  A further area of products is so called security area where we offer a black holing service on the same port and a virtual PNI service.  You can also establish closed user groups, which is very interesting for those guys who have specific type of traffic exchange needs. Take the example again of the automotive business or the finance business. They can create their own closed group within the DE-CIX interconnection world, and this closed user group will be handled under the specific quality standard they need. The variety of different services we can add to the portfolio is unlimited because you have the platform, and everything else is a matter of bandwidth, the size of the physical interface, and the number of participants. The bigger the ecosystem is, the bigger the benefits are for all participants. But of course this has to be done over time.

TR: Was that something you had to do with new hardware or via software upgrades?  Is it something you handled internally or through partners?

II: It's done internally, and it’s based on the DE-CIX Apollon platform. The hardware piece of the platform on the router side is Nokia. And of course, we do have management systems other different gear. But the software part is done internally and, of course, we also use features which are on board of Nokia’s operating system.

TR: What limits platforms like this the most today?

II: I would say the power needs are always a problem. If we have to handle tons of traffic, because of more data in the future, we'll need a different level of power efficiency. This is a physical challenge. From the size perspective we will need a different port density on the smallest possible size of gear. Nokia has already made great progress on this. However, there is always room for improvement in technology, of course.

TR: What do you think is on the horizon that has the potential to push those limits?

II: We see the mountains of new trans- and inter-connection service world if we talk about the Internet of Things. The IOT topic is an early one for our business, but we have to start developing now. The question we ask ourselves is, "How can we serve different types of inter-connection needs if it comes to an IoT inter-clutch?" There will be two general types of inter-connection demands. One is based on the aggregation model as we have it today in inter-connection hubs. For instance, consider a connected refrigerator exchanging traffic with Amazon Fresh. This model doesn't depend on the lowest latency possible and can be easily handled this way. The other model, for which a good example is self-driving cars, is a completely different one. You will have things which have to talk to each other live and over very low latency and therefore very close to the edge. In this case, we need a new type of inter-connection service. We have started our plans of creation of a small platform which can be even implemented on mobile towers or even at highway crossroads. You obviously need a very solid 5G infrastructure, but at the same time, a very solid interconnection of this infrastructure with the content guys.

TR: How long do you think it will be before we see something like that in practice?

II: I believe everything beyond five years is unrealistic. We are already having discussions with companies from the automotive industry, they see the future there. When the the CEO of Audi introduced the new A8 version, he didn't talk about great wheels or how powerful and efficient the engine is it.  He talked about AI. He talked about different data services the car can deliver or provide. He talked about a digital car company, where the car becomes just a platform for new type of products. This will not happen tomorrow, not for sure, but it has already started. We have to be prepared, it's just a matter of adoption.

TR: How is your expansion to Mumbai going?  What differences do you see in the exchange business in India?

II: There is definitely a difference. You have a limited number of relay carrying mutual data centers, dispersed. Whole the size of the country, limited size of proper data centers, states, right? But this is changing. Companies like GPX are expanding in Mumbai, and we could see investments by companies like Netmagic, Vodafone and Tata. A further difference is of course the price for local loop. Sometimes, it costs 10 times more to get from the cable landing station into the data center than for transatlantic or transpacific.  But we see this changing also. The price reduction is there. When we started five years ago in Dubai, a similarly very regulated market with extremely high pricing on the local, a 10G link from the cable landing station into the data center was $170K, but this has changed five years later to under $8K.  I'm not saying we are the only one reason but we believe our contribution helped to make the market more open, introduce new interconnection services and motivate competition. I believe this is what we'll see in India. We plan to run the same type of service we offer New York, Frankfurt, Madrid, etc. That means really distributed access points in all relevant data centers like GPX, Netmagic, Tata and so on. So we'll see within the next months, expansion of Mumbai-IX into further data centers, and I believe this will help the market to grow, and especially to create more dynamic in terms of pricing for the operators. And we do have definitely a momentum there. We see all big international guys, like Apple, Netflix, Amazon, you name it, they have already arrived or will arrive very soon. From the regulatory perspective, the regime was changed so it's easier now to operate without a lot of restrictions. I'm not saying that they opened up everything and it's perfect, but it's getting better. It's not a walled garden like we see in China, for instance. It's different. It's a long way, will not happen immediately.

TR: Are there any new geographies you see on the rise in the interconnection world?

II: Madrid is the fastest-growing interconnection market on the planet. We have more than 100 networks already connected. We had a trade show in November, and the attendance suggested that Madrid is becoming the next big telecommunications capital in Europe. The big content guys have already arrived in Spain, the market itself is growing fast due to digitalization, and the Iberian Peninsula is a perfect base station for a new regime of traffic and connection for West Africa. In other words, for traffic from West Africa, Madrid will become the next London because then it's much cheaper and closer to the edge to interconnect in Madrid. Of course, the best would be to do it in Africa directly, but it will take time because they don't have political and business stability there.

TR: Thank you for talking with Telecom Ramblings!

Categories: Industry Spotlight · Interconnection

Discuss this Post


Leave a Comment

You may Log In to post a comment, or fill in the form to post anonymously.





  • Ramblings’ Jobs

    Post a Job - Just $99/30days
  • Event Calendar