David DiGiovanni, CTO, OFSRole of the CTO
Thinking about technology in general and Software-Defined Networks (SDN) in particular, a CTO’s role is to determine how we do things internally, the product’s usability, and its evolution towards SDN.
"As bandwidth starts to become scarce and precious, you need to find more efficient ways. That’s one of the profound changes that SDN allows"
We’ve got plant managers in the factories—developing new products, adopting lean manufacturing techniques, and constantly improving to lower costs and improve efficiency— and are usually not up to speed on the role of Artificial Intelligence (AI) and big data in our factories. As a CTO, having the luxury of staying up-to-date with current trends and thinking ways to implement big data technology, my responsibility is to be the forward thinker and alert the plant managers on unexplored trends. The executive team realized a lot of technological changes, so they needed somebody to be their eyes and ears into the new areas. My role has become critical in the organization as specialization and technology development continues to accelerate. As a CTO, you’ve got to have your fingers in a lot more pies.
As a supplier of hardware that goes into these networks, we make fiber cables, racks, and connectivity that are going in to the data centers. The flexibility awarded by software defined platforms started 10-15 years ago. From a cable perspective, it doesn’t matter what the data looks like or how it’s controlled, so it hasn’t been necessary to change or adapt at the fiber level. We’re only starting to see the impact of SDN on the network management and our product space, and I think it will have profound implications.
SDN is becoming important because of efficiencies. A lot of things come down to cost. It’s the capital cost of acquiring and deploying the hardware, and the operating cost of managing the equipment and the network. The efficiency of both these things is improved quite dramatically with SDN, but on an operational side, there’s an exponentially increasing demand for bandwidth. This has started to put a strain on the overall network. So, as bandwidth becomes scarce and precious, you need to find more efficient ways, and that’s one of the profound changes that SDN allows.
The answer comes back to cost of ownership, the operating cost of having to provision new customers, bring new services online, and adapt things in a flexible fashion. How to make a simple change in software in order to provision new customers or new services—that’s always been a Holy Grail and SDN has just started impacting it.
It’s becoming an issue at the fiber cabling connectivity layer because you need to abandon proprietary standards and adopt some standardization. Standards typically seem to be lagging where the industry wants to incorporate higher level of standardization and get all the competing vendors on the same page. That’s a slow process. Seeing bandwidth demand growing exponentially and the pace of tech continuing to accelerate, arises a contradiction. It’s interesting to watch what’s going to happen with standardization.
You can’t do self driving cars today because ‘latency’ just isn’t there yet. In order to widely adopt these cars, we need a different type of infrastructure with a higher reliability, much lower latency— craving for bandwidth and driving the demand for fiber, cable, and hardware. As the volume goes up, scalability needs to be there in terms of cost, installation cost being the dominant factor. But you need to be able to install it at a much faster rate with a much lower cost. So the big challenge in the optical world is how to get that scalability. You can scale software quite readily, and although computing power is scaling very quickly, installation of hardware is not.
In a schizophrenic environment, we need to execute efficiently and quickly on the current portfolio of programs. We can’t delay it because the customer demand is very high, but at the same time, we need to pay attention to these trends on the horizon. People need to quickly understand when to abandon the current technology and get on the next new thing. While implementing AI kind of process, they should be able to switch and adapt to it quickly. In terms of data and moving data, we’re waiting within our company, looking at the tip of the iceberg and at new management approaches to manage some of these things.
Our machines do a lot of chemical engineering— making glass, making cables and other components that use of them. With a big machine running, uptime is very important. Can you do things like scheduling maintenance, reduce scrap, improve yield, and find correlations among processes that in the past were difficult for an engineer. Staring at pie charts and analyzing the data on a laptop aren’t going to show insights that some AI scheme could.
Talking about innovation, a need of bridging physically distributed engineering and R&D people from various facilities as a team is hard. Our teams from Japan, US, and Europe needed a simple way for brainstorming. In the past we suggested to do it during lunch time, we can’t do that anymore.
It depends on the team. In my case, the management team is tech savvy and future oriented, not framing business decisions on numbers, but more in a strategic sense. I can say—I don’t know what the ROI is going to be but this is an important trend. We can’t decide a hard numerical value now but these things could be leveraged into new space— that kind of argument is acceptable. This researched approach doesn’t work sometimes, but that’s our culture.
The most critical thing is the environment and the surroundings that you’ve got. You need to surround yourself with good people, because they will leverage each other. Once you’ve got people, you need the luxury of being able to explore and a bit of freedom to think about interesting problems. Having some slack to do some more interesting stuff often is very fruitful. You don’t necessarily know what it'll lead to, but unexpected things may happen. In terms of environment, it’s no longer sufficient to be an engineering working in aerospace, you must learn to adapt. Having good people around and a little freedom makes a world of difference.
