Industry Events





Xtera Execs Talk about Technology and the Company’s Future
Submarine Cable NewsFeed - Market Snapshot
Monday, 24 July 2017 17:56

Editor’s Note:

In January, H.I.G. Capital had acquired substantially all the assets of Xtera Communications, Inc.  In a statement made at the time, H.I.G. Capital said that under its ownership, “Xtera’s management and technical team would remain at the helm of the business, focused on successfully executing key existing customer contracts and expanding the business in the rapidly growing markets it serves with a clear roadmap of disruptive product launches.”

I recently had the opportunity to speak with Robert Richardson, Founder and Chief Sales Officer of Xtera, and Tony Frisch, Chief Technology Officer, about the company in light of the new investment.  The following are their views on the company’s outlook and technological advances.

Richardson: The future looks very bright for Xtera.  I think we have a very strong investment partner in H.I.G. Capital.  They’re a $22 billion fund.  They are not a short-term investor.  Their typical investment cycle is about 7 years on average but they have some investments they’ve been involved with for over 10 years.  They certainly understand our business because they spent a lot of time visiting with customers, and they understand that the submarine business is a cyclical one.

Frisch: We originally got into the subsea cable business because we had a lot of experienced and energetic people who wanted to do more than build terminal equipment.  Customers encouraged us to build a subsea repeater. 

In doing so we did a couple of things a little differently.  We decided to build not just a simple EDFA, but to add Raman to it because we had a lot of expertise on that inside the company.  We felt that by putting Raman we could get a better noise figure.  With the Raman, we’re getting some pre-amplification in the 30 kilometers of fiber coming into the repeater, so that’s acting as a low level amplifier rather than acting as a loss.  And we could use that Raman to improve the noise figure or to extend the bandwidth of the amplifier.  We currently have 55 nm, moving up to about 65 nm this year. We also have sold that unit in a reduced 35 nm bandwidth allowing us to get up to 140 km spans without spending a lot of money on expensive fiber. 

With 65 nm, we are capable of more than 40 Tbps per fiber pair this year and in the lab we’ve run in excess of 80 nm.  Depending on the market demand, we could move to production on 80 nm at some point in the future. 

The question of customer demand is a very interesting one.  When we first started making this repeater, of course the first thing we were concerned about was making sure that we tested every aspect of it.  So for the sea trials, we fitted the repeater with a whole set of sensors – pressure sensors (to know if anything was leaking into the repeater under pressure), accelerometers, temperature sensors and so on.  The intention was to strip the sensors out in the production model, but actually we had customers on board the ship looking at all this stuff and they were saying “No, don’t take them out.”  The customers recognized that the sensors were not designed to be super reliable, but as long as they are working, they tell them something useful. 

Particularly, the customers had seen the accelerometers, which are telling you what’s going on with the repeater as it goes down to the seabed.  One of them made the interesting point that if the ship knew that the customer could tell how carefully they were handling the repeater they would be a little bit more inclined to handle it really nicely.  So we kept that accelerometer in there and we also discovered a couple of other interesting things.  A lot of these ships have some really sophisticated software that monitors how fast the train is going out, the angle of the cable in the water, etc.  Using this data, they try to predict where the repeater comes down onto the seabed.  But I think we discovered on the first sea trial that they can be half an hour wrong and that means that you have a fairly poor idea what happens during the lay without the information from the accelerometer.  People we talked to said that’s the stuff they’d really like to have.  It also means they can monitor the repeater on the seabed and just be sure that it really is working. 

Richardson: Something that is really interesting is the idea of data center-to-data center connectivity.  We initially pushed this idea about 11 or 12 years ago at Submarine Network World and it didn’t get a great reception at that point in time, but I think the world has changed.

Frisch: This was a very simple and straightforward idea that seems now to be catching on.  If you really want to get the traffic inland, the smart thing to do is to pick up the submarine terminal and put that inland, rather than having the submarine terminal close to the beach and then a terrestrial terminal at the data center, PoP, or whatever.  That’s now something that I think a lot of people are doing.  We’re hoping, and it’s something that were doing work on, to take some of the electronics used in a subsea repeater and put them into a dry unit that can be used pretty much anywhere. 

The intention is not just a simple extension, but using a consistent piece of technology so that you don’t end up using a subsea amplifier and then a terrestrial amplifier.  It would also have some of the advantages that you get with a submarine amplifier, which has to be very reliable because it is so expensive to repair a subsea amplifier.

I think that one thing we will be looking at, again depending on customer demand, is the notion of putting what looks like a land-adapted version of a subsea repeater that has duplicated pumps and duplicated control units and putting that into a package that can be installed in a rack and powered off the local supply.  If one of the pumps fails, it won’t impact the service at all, it will just raise an alarm back in the data center and someone will realize that at some point they need to go out and replace it.  You can do it during a scheduled maintenance window rather than rushing out immediately because they have an outage. 

And along with that we’ve been looking hard at this whole business of open systems.  I think everyone is saying they’re doing it or willing to do it.  I think long-term this is suits us.  We’ve been involved for more than three years talking to customers about open systems and the fact that our recent commercial strategy has been to focus on the submarine, makes it very much more natural for us to say let’s go really open and not try and fudge the issue.  We don’t have any problem being open.

Richardson: In fact, we have a couple of buyers who are bringing their equipment into our lab for testing to verify that their equipment will work with our undersea technology and our open systems gateway.  We were the pioneers of the whole idea of idea of upgrades and open systems are in our DNA.  I think more importantly that’s where the market shift has been, particularly with the idea of people buying fiber pairs -- now each company owning pairs within the cable has a different supplier that they want to use.  They don’t want to have something unique for their undersea, they want the whole network in totality.  So I think embracing open systems is simply the way the network is evolving. 

One of the big things that came out of our discussions with customers is that they also want to deal with spectrum.  Because of the large amount of bandwidth available from our repeaters it may be better to sell spectrum than fiber pairs.  What we’ve done with our technology is design something that actually embraces selling "virtual" fiber pairs or spectrum.  If you want to sell a fiber pair as a submarine cable owner and you want something to put on there, that’s fine.  If you are a network owner and you want to sell spectrum, you can sell it off yourself because you have such a high amount of bandwidth you can sell and we have techniques to make that easy and managed it like it’s a physical asset as opposed to a virtual asset. 

Frisch: I think that’s one of the key problems that people don’t tend to think about.  You can have customers using products from different vendors.  And if you have to disconnect your terminal because you have to do some maintenance on it and you need to make sure that whatever equipment is facing all these different inputs is capable of making sure that none of them disappear or when something goes wrong it doesn’t impact the other guys.  That’s not fantastically complicated, but when you are taking into account failures and security issues, it’s something cable operators need to think about.  It’s more complicated that sometimes is appreciated. 

Richardson: There will be some interesting news coming out of Xtera in the next few months.  This will show that not only are the customers behind us – voting with their wallets -- but that they also are very supportive of us.  They see that we bring innovation and change to the marketplace.  They invite that.