We spoke to Juha Holkkola, Co-Founder and Chief Technologist at FusionLayer.
FusionLayer helps service providers and datacentres automate processes that touch networks. This is accomplished with cutting-edge technologies that function as the unified management system for all network-related information. FusionLayer automates workflows and simplifies network management for service providers, datacentres and enterprises that operate at scale. The company’s vendor agnostic technology unleashes network agility by powering interoperability between SDN-enabled, cloud-based and existing networks.
Nine out of 10 of the world’s largest service providers leverage FusionLayer.
Keep reading to get an insight into FusionLayer, their solutions and SDN/NFV
Why did you start FusionLayer?
FusionLayer used to be part of Nixu Group and was originally operating under the Nixu Software brand. At the time, Nixu was divided in two divisions: an information security consultancy and a DDI product operation. Through the Nixu history, our DDI legacy actually goes back all the way to the mid 90s.
A few years ago, Nixu was listed at the Helsinki stock exchange. As the investor story was focused on information security, we carried out a reverse merger before the Nixu IPO. Consequently, the group was split into two entities of which the information security arm currently trades as Nixu at OMX Nasdaq Helsinki, while the DDI arm continued as a private company.
Around the same time, we started seeing the emergence of cloud computing and automation. In an effort to reposition Nixu Software, we started the development of FusionLayer Infinity – our flagship solution – and changed the company name to FusionLayer. The new name aligned better with our technology vision of introducing a new network layer used to fuse automated processes.
What does FusionLayer focus on?
When starting our operations as FusionLayer, we were working with many service providers and MSPs that were building up new cloud operations. Some of these services were to be offered to end-customers and others for in-house activities, but in both cases the automation was usually carried out within a silo. In big organisations like service providers, there tends to be a silo for almost every single IT function.
In my opinion, cloud computing is all about self-service and automation. So, to be able to automate services as part of the digital transformation, you need to make sure that processes across different silos interoperate. And if you think about this in the bigger picture, the one place where all these different silos meet it is the network. Applications are connected to networks; application delivery and network services like security all happen inside the networks; and even setting up the networks themselves is being automated.
To make all this work together, you need you have a unified system in place that allows you to make all the network-related information available to different orchestrators across different silos. So in a nutshell, FusionLayer provides a new network management layer that fuses automated processes on the network level, across the silos.
What is the software-defined IPAM solution – FusionLayer Infinity and what does it do?
FusionLayer Infinity is the unified management system used to manage all network-related data. This includes the logical networks, the VLANs, and the Virtual Routing and Forwarding – or VRFs for short. The idea is, once I’m able to manage all the network-related data in a single location, I can leverage the Application Programming Interface (API) to provide a shared backend for all network-related data.
For example; if someone is running OpenStack and deploys workloads automatically across multiple data centres, they need to make sure that those workloads get an IP address from an appropriate network connected to the right VLAN. Those pieces of information can be can queried from our system by automation, even when the data centre infrastructure is being managed by another team. Right now, people still largely use excel spreadsheets for the job, making automation hard.
Another use case we increasingly see is network management in the multi-cloud use case. Private enterprise networks have traditionally been activated in on-premise equipment, but we are now seeing organizations extend their enterprises networks into public clouds like Azure and AWS. Since segments of these public clouds become a part of the enterprise network, our customers have to be able to manage all segments of their enterprise networks in a unified system – regardless of whether a given network segment is activated in AWS or an on-premise network equipment. This is another area where we can help customers.
What are your thoughts on Hybrid IT?
Originally Hybrid Cloud was relatively simple. For the most part, people were running certain things in their on-premise data centres, and others in public clouds. There was not really a mechanism or technologies in place to dynamically scale out on-premise services to the public cloud – let alone to multiple clouds. It was all pretty static.
What we are seeing now is the proliferation of next-generation hybrid IT solutions. For example, if I am running a business application in my on-premise data centre and I suddenly get a huge increase in the usage level, I am able to burst to the public out in order to increase my service capacity. This changes the management process completely. So, you have lifecycle service orchestrators and micro services and basically what you do is, you start up running your service on premise. But if you need more capacity, you dynamically go to public cloud and get the additional capacity when needed.
So to leverage hybrid IT as part of the digital transformation process, it all boils down to lifecycle orchestration and multi-cloud. That’s where the real upside of hybrid IT is. Without the ability to scale in or out dynamically, hybrid IT really isn’t that special at all. But it looks like a lot is happening in this area right now, so I am confident that Hybrid IT and especially multi-cloud will evolve in leaps and bounds as we move forward.
What do you love most about FusionLayer?
It’s a good culture, a culture of innovation. In many ways, we have been pioneering the industry over the years. For example, we hold a US patent for DNS firewall that we invented back in 2006 although the product category didn’t really materialize until 2012 or so. We also have some issued patents on orchestration and SDN that we invented around 2012, and are now becoming the industry standard.
The most fun part in all of this is the foresight phase during which we analyse the megatrends affecting our industry, and try to envision how the technology landscape will play out, the partners you want to work with. We have actually invested millions based on this kind of analysis, so when you see your foresights becoming the reality it is intellectually extremely rewarding.
How does FusionLayer stand out from the competition?
We are ahead of the curve and customer-driven. Our usual approach has been to combine technologies from different product categories to provide the kind of functionality that our customers find useful in the real world. Take for example our Software Defined IPAM product, Infinity. While we come from the IP Address Management (IPAM) realm, we have essentially merged a traditional IPAM and a Configuration Management Database (CMDB) together.
By doing this, we have created a network-CMDB that is neither 100% IPAM nor CMDB. Regardless, our service providers have found it hugely helpful because they can use it as the backend that powers their automation initiatives on the network side. Traditional IPAMs cannot do this, nor can traditional CMDBs.
More generally, I think that innovation often lies in taking known arts and putting those together in an innovative way. It is actually quite rare for a technology to die completely, what usually happens is that they converge with something else. Mobile phones are a great example of this as they have eaten up at least ten different products that used to have categories of their own.
Who do you see as the biggest innovators in the industry?
That would have to be Red Hat, it’s a really good organisation and they are doing a great job in their field. Their Chief Technology has a great vision for the company.
When did you first hear about SDN/NFV?
I came across SDN for the first time back around 2009 or 2010 through a company called Vyatta. I don’t think they called it SDN back then, but the basic principles were the same. Vyatta then got acquired by Brocade in 2012, and in June 2017 AT&T acquired Vyatta Software Technology from Brocade. Talk about revolution eating its own.
That said, I think it was in 2013 when SDN and NFV became more mainstream. I remember vising TM Forum conference in May 2013 in Nice, and suddenly every vendor had a SDN or a NFV banner at their stand.
What were your first thoughts?
Back in 2009, I thought that SDN/NFV was a very tempting technology. It was interesting that you could actually have the networking operating system as open source software. But as for the concept itself, I don’t think anyone really knew what the use cases were at the time. It was just an inexpensive way of doing what Cisco and others were doing. Over time, it has evolved a lot with some new use cases that are transformational to say the least.
SDN is very helpful for data centres. But it is not for everything. SD-WAN has emerged from this trend and I think that is great. That’s because outside the obvious enterprise-WAN use case, it can also be used for other things like secure Internet of Things networking. Because securing the IoT devices is going to be next to impossible, I believe it is SD-WAN of all technologies that will come to the rescue.
What do you think are the strengths of SDN/NFV?
Traditional networks are rather tedious and it can literally take months to set things up. The agility you get from dynamic networking is going to change the entire IT paradigm, because it enables end-to-end automation from cloud to ground. I cannot even begin to imagine how networks will be used when there will be an endless, instantaneous supply. It is going to change quite a few things.
Think back to the year 2000, could you imagine back then how mobility would be today? I think the same applies to networks and where we are today. The first step is getting the technology working – making it easy to deploy, reliable and easy to run. Once that happens, networks will become a commodity that will be available to practically anyone – at any time. I believe this will unleash business innovation in how networks can be used as part of digital transformation.
It will be huge.
Another strength is that with dynamic networks you can create a huge number of secure isolated networks that connect to the public internet through protected end points. Once this happens you can create new secure networks at will. This will allow service providers to set up hotspots where you can connect all kinds of things securely without actually having to consider the security of the devices at all. When that is possible, many things in the networking industry will change.
When anything is being automated, the biggest threat is often the lack of control and visibility. With SDN and NFV, you still need humans to keep an eye on things and make sure that if something goes wrong they will have a quick access to what the automation has been up to. Relying blindly on software to do the job is rarely the best of ideas.
Where do you see the future go in this industry? What technologies will the focus be on in the upcoming years?
The two key areas for the future will be automation and self-service. These two are related, because you can’t have self-service without automation.
Generally speaking, I think the key focus will be on the silo problem. You can break the silos within certain areas but there is still a lot to do, especially in the hybrid multi-cloud scenarios. But what I think will happen is that all of these different architecture will get weaved together to enable end-to-end automation. That is probably what the industry and service providers will focus on in the next three years.
What emerging technologies other than SDN/NFV are you most impressed by and why?
The one solution I’ve been seeing quite a lot at service providers is Ansible, by Red Hat. While SDN and NFV are great for some use cases, they do not really work in many others. As Ansible offers an agentless model for orchestrating also physical devices, it complements SDN and NFV by enabling automated processes that run all the way from cloud portals into the physical world. This is really important because IT clouds do not run in the sky – you still need the physical infrastructure.
Having talked with dozens of Tier-1 service providers over the last few months, I would say that at least 70% of them have been running Ansible in their testbeds. And much of this is for the physical devices, not just virtualized things that was the norm until late 2017. Although this allows service providers to become more agile, it also allows them to realize huge Operating Expense (OPEX) savings. That is because when you have tons of physical devices to manage, you will have an army of engineers and technicians configuring them manually. So, when you automate this, you get agility, speed and cost savings. This is really the holy grail for service providers.
What is the biggest difficulty you come across when recruiting for FusionLayer?
Finding the right skillset! There are less skilled people then there are jobs right now. If you start looking into automation and orchestration, the guys should actually understand how the processes work, but they also need to have an architect mind set in order to be able to take that knowledge and make it usable in the automation context. In that there is definitely a shortage.
What would be your one piece of advice to a Network Professional who wants to specialise in SDN/NFV?
Focus on the big picture. If you think about how things have been done previously, everything has been manual. In that kind of scenario, one can easily end up putting a huge amount of focus on details. This is perfectly understandable and was actually a virtue in the old world.
But when you have to automate something you need to think through the entire workflow first. You need to have a blueprint in your mind about what you actually want to accomplish. Then you decide how you are going to implement it. Often, people do it the other way around. They start with the details and then you get stuck. I think you need skills like this in the evolving networking industry.