Xtera Communications has a single facility in Allen, Texas, where 150 people are employed, engaged primarily in research and development and engineering roles. Production activities are outsourced to undisclosed contract manufacturers, one each for optical and electronic components. Intellectual property relating to Xtera Communications' products is held by the company, which has 10 patents and 29 pending to date.

Xtera Communications raised funding of $110 million in January 2001, in addition to $23 million raised at launch. Commenting on the financial position of the company, Mr Harrison said that present funding was projected to be sufficient to carry Xtera Communications to the third quarter of 2003 at the current cash-burn rate. He remarked that funding issues were not currently a matter of concern for the company, however if the need for success-based capital were required it would be made available.

Adapting to the market

Xtera Communications has announced a revision of its business model in response to the dramatic changes over the past eighteen months in the optical telecommunications market. Essentially, the company is shifting its position from sub-system supplier to system vendor - it will now be competing with what were, until recently, potential customers.

A concrete decision to refocus was reached in July 2001. Xtera was originally formed with a view to exploiting Raman technology to address escalating demand for bandwidth. As such, the company developed a sub-system for integration into DWDM systems, based upon discrete Raman amplification technology. This solution was developed to open up the S-band, boost network capacity, and thus augment the capabilities of systems vendors with solutions operating in the C and L bands. At the time Xtera Communications opted to regroup, its product had reached the stage of trials with operators.

As a result of changes in the market since the original product concept was formulated, the company reached the conclusion in late 2000 that such a solution would not be commercially viable in the present climate, and would likely not be so for some years in the targeted long-haul and ultra long-haul sectors.

Paul Harrison commented, "Market requirements have changed from demand for bandwidth at any cost to demand for adequate bandwidth at low cost. Bandwidth capacity is not currently the prime concern for service providers."

Mr Harrison noted that, despite the market downturn, demand still exists, and is estimated to be growing at between 60% and 120% annually. Driving this growth is demand for broadband access from the small business and residential sectors, combined with a recovery in IT expenditure generally. He added, "Companies realise that investment in IT - moving to web-based applications - remains the way to go, and that it will deliver substantial cost savings."

Xtera Communications has developed a solution based upon discrete Raman amplification. Paul Harrison commented that Corvis and Nortel Networks (through acquisition of Qtera) have promoted Raman amplification as a means to address issues associated with distance, primarily in ultra long-haul networks. These companies utilise distributed Raman amplification - pumping of the line fibre, causing the fibre to amplify the signal at the far end of the span. This technique improves signal to noise ratio of the system, allowing increased cascading of EDFAs, using Raman for the reach.

Discrete Raman amplification holds the advantage that, unlike EDFAs, which are restricted to bandwidths above 1525 nm, amplification is possible anywhere on the fibre - from 1300 nm to 1700 nm. The original Xtera Communications product used this technique to extend capacity to the S-band - 1490 nm to 1520 nm. The product uses specialised fibre, on a spool within the circuit pack, to achieve the Raman amplification effect. Overlying this sub-system is software that controls factors including transience and power fluctuations. This complete sub-system is contained in a shelf of equipment - hence discrete.

Nu-Wave

Paul Harrison explained that the 'secret sauce' differentiating Xtera's redeveloped product - to be named Nu-Wave - is the combining of both discrete and distributed Raman amplification into a single amplifier product. The integration of these two techniques gives a solution delivering amplification over any bandwidth, through use of discrete Raman amplification, together with long reach, through use of distributed Raman amplification. Defining the operating parameters of the system, Mr Harrison stated that the product offers 100 nm of continuous amplified bandwidth over distances up to 1500 km.

Xtera Communications' Nu-Wave system has been developed and prepared for launch extremely rapidly, due both to its derivation from the companies existing technology, and to the expertise and experience of the staff. The timescale spanned slightly less than one year from the initial decision to change direction, taken in May 2001, to planned product launch at OFC in March 2002.

With the 100 nm of bandwidth offered by the Nu-Wave product, Paul Harrison noted that Xtera is bringing to the marketplace a three-fold increase in capacity, above that of a C-band amplifier. This bandwidth is continuous, usable capacity, across the C, L and S bands. Mr Harrison believes Xtera Communications' approach to the issue of network capacity contrasts with that of most other systems vendors. In general, vendors pack in added capacity by increasing bit-rates.

Expanding on this theme, he said, "Companies such as Ciena are going with utilisation of a 25 GHz or 12.5 GHz space to increase capacity of their systems. As an analogy, if you have a jar full of toothpicks, how do you get more toothpicks into that jar? One way is to reduce the size of the toothpicks, and so fit more into the jar. This is effectively the approach that most systems manufacturers are taking. We decided that it was simpler to get a bigger jar! Xtera Communications has basically increased the size of the pipe, rather than try to pack more capacity into the existing pipe."

In this way, Xtera has discarded the concept of 'bands' in favour of continuous spectrum, which, available anywhere, will optimise the available bandwidth dependent on fibre types.

The company's strategy brings a number of advantages from a technological aspect, arising from the ability to use tried and tested, second and third generation technologies - for example, the use of 50 GHz filters, standard FEC and Lithium Niobate modulators. In addition, such components offer the benefits of standardised footprint and pin-outs, through pacts such as the Multiple Sourcing Agreement (MSA) entered into by many components vendors supplying the transponder market. Combined with current depressed prices for components, Xtera Communications has been able to leverage these advantages to produce a densely packed and low cost product.

Xtera Communications claims to have developed the highest density product available on the market. Paul Harrison stated that the Nu-Wave system offers, over 100 nm of bandwidth with 50 GHz spacing, 240 channels at 10 Gbit/s, contained in three 26" X 7' X 12" bays of equipment - a total of 2.4 Tera-bits of capacity. This capacity equates to 90 10 Gbit/s channels per industry-standard rack. Comparing this with the competition, Mr Harrison said that Nortel, through Qtera, achieves 16 channels per rack; OptiMight Communications offers 24 channels (closed down -ed); Fujitsu, with the highest density product presently available, delivers 44 channels.

Low-cost

The Nu-Wave system also promises to surpass the competition in terms of value for money by providing three times the bandwidth of a single C-band solution for the same initial expenditure.

Commenting on this, Mr Harrison noted that a customer could therefore justify buying the Nu-Wave solution rather than a lower capacity system, even if the additional capacity were not needed immediately. As demand on the network climbs, the carrier would be able to increase capacity on up to 2.4 Tera-bits by the simple addition of transponders at either end of the network. Buying a similarly priced C-band system, a carrier would need to install additional components twice over to achieve the same 2.4 Tera-bits capacity. Furthermore, the system is fully automated, able to calculate channel and power balancing as capacity is increased.

Strategy

Xtera Communications' strategy for positioning its Nu-Wave 'Wide Reach' (wide bandwidth plus long reach) system is based in part upon an effort to reclassify a portion of the long-haul/ultra long-haul WDM market. The company has identified a niche from 600 km up to 1500 km in the U.S. market. This segment can be defined as the NFL (National Football League) cities. The majority of service providers serving these cities require networks of between 600 km and 1500 km in length - further, most of the traffic growth is between these cities.

Commenting on the market topology, Paul Harrison said, "Until recently distance was key, long distance capabilities of network equipment was considered vital, the longer the reach the better, with talk of 3,000 km and 4,000 km systems. This perception is now totally unfounded. From our discussions with service providers, we have found that the majority of traffic is generated between certain points - the NFL cities - that are not going to change. These points usually fall into the 600 km to 1500 km range. These are the 'sweet spots' in the market."

Add-drop capability

Xtera Communications has incorporated optical add-drop capability into its systems, traditionally an expensive or inflexible approach. Vendors such as Marconi and Sycamore have integrated add-drop functionality with expensive MEMS-based dynamic functionality. Ciena and Nortel provide inflexible systems that suffer from problems with noise and cross-talk as the fixed filters are cascaded.

With Xtera's solution, add-drop multiplexers can be sited at 80 km intervals along a 1500 km network, with no link distance penalty. Given that the system delivers 100 nm of bandwidth, large portions of capacity - up to fifty-four channels - may be taken off to serve intermediate points along a route. Filters allowing takeoff of four or thirteen channels are also available. Customers will be offered the option of pre-positioning filters at any points along a link in their network, at no extra cost.

Commenting on this strategy, Mr Harrison explained that it brings ultra long-haul functionality into any network. "A carrier could install a long-distance network in a ring topology to serve the major population centres in a state or region."

With respect to global markets, Mr Harrison noted that regions such as Asia are far more difficult logistically, and much more expensive, to address - "for a start-up company, higher costs shorten the window for success." He regards Europe as less suitable due to the general absence of ultra long-haul networks. Moreover, the pan-European operators that would be potential customers are not investing in equipment at present. These companies' investments are success-based, and, as a result, they are unlikely to be spending before 2003 at the earliest. Regional applications, however, particularly for incumbent PTT's, lend themselves to the Nu-Wave architecture as multiple OADM's can be folded into a continuous network, picking up feeder traffic and bringing it back to the regional hub.

Market size

Paul Harrison cited the total value of WDM and transport equipment markets as having declined from around $6 billion in 2000 to $3 billion in 2001. This market is expected to remain flat, or possibly fall even further, during 2002 with recovery anticipated in 2003 or 2004.

Commenting on these figures, Mr Harrison remarked that the key fact to bear in mind was that this remains a $3 billion market, and that it is driven by growth in traffic. He offered the view that 2002 will be a year when the manufacturers seek to plug gaps in their existing systems. From discussions with major players in the industry, Xtera Communications expects to see growth return to the market in early to mid-2003, with service providers beginning to install new routes. Mr Harrison pointed out that even conservative estimates concur with this view.

In terms of Xtera Communications' prospects in this climate, he added, "Carriers are still investing in equipment and the market remains healthy, although absolute spending levels have fallen drastically. For a small company such as Xtera Communications, success will come down to having the right product, at the right time, at the right price. Timing is key here, hence our change in direction. Any small start-up that has launched product over the past six months is likely to fail. Making any impact in the current market is practically impossible."

Mr Harrison noted that RHK has recently begun to split the long-haul market into 600 km and less, and 600 km to 1500 km segments. This is viewed by Xtera to be logical market segmentation, allying with analysis that the company has conducted internally. Paul Harrison said, "Even though the WDM market as a whole is projected to be flat, or almost flat, over the next two to three years, separating out the 600 to 1500 km segment reveals growth estimates of 46% CAGR, from 2001 to 2005."

Customers

Xtera is confident that it can establish a position in the market, despite competition from the major vendors. The company's key strength will be pricing - offering capabilities at a price that the competition is unable to match. Whilst the major carriers have established relationships with specific large systems vendors, increasingly, these carriers are realising that such vendors do not always offer the best technology or the lowest prices. Mr Harrison noted that these factors are becoming fundamental in the selection of equipment suppliers. In addressing these issues, carriers are tending to sign up a small system vendor in addition to their core supplier. This policy has the effect of keeping the established, major equipment vendors focused.

Commented on the market shakeout, Paul Harrison noted, "Whilst carriers fail, their customers, and the traffic generated, remain - this business will be taken over by surviving service providers. The end result is fewer customers for equipment vendors, although the overall demand for equipment will stay more or less the same."

Xtera Communications is now concentrating on delivering the Nu-Wave product to market. General availability is planned for July 2002, in time for carriers to complete product testing in readiness for live installations early to mid-2003. Mr Harrison stated that, to date, the company has not missed any product development milestones.

The company is targeting the IXCs in the U.S., and plans to sign two major IXCs, although, Paul Harrison added, one major contract with one of these customers would suffice to launch the company. Xtera is focusing its attention on those carriers with an intention of making investments in network equipment over the coming years. Referring to the previously mentioned market value figure of $3 billion, Paul Harrison noted that he would be happy to take a $20 million portion of that total in the first quarter of 2003, adding that this figure did not represent projected earnings for Xtera Communications.

At the time Cerent launched its product onto the market, there were nine major vendors profitably selling products into the OC-48 space. Cerent then appeared on the scene with a product offering only the basic features required by carriers, at a price - initially below cost - that these players could not equal. Cerent's strategy was to gain a foothold in the market before the established vendors could react to produce a competitive product - a cycle typically taking two years for a large vendor.

Comparing Xtera Communications solution with that of others in the ultra long haul space, Mr Harrison stated that they fall into two categories:

1. Integrated transport and switching solutions

2. Transport only solutions

"Integrated product offerings are by definition proprietary closed systems, comprising a WDM and a switch that eliminates a carrier's choice to adopt "best in breed" policies as well as a multi-vendor solution. Also any perceived economic benefit is only obtained when a network of large nodes is interconnected - a very expensive first application - rather than placing capacity where and when you need it a minimal cost."

"Many who compete in the transport only section have chosen a combination of obsolete and cutting edge technologies that are right at the beginning of their respective cost curves. Multiple EDFA bands mean multiple truck rolls and manual tuning to alleviate band interactions. 40 Gbit/s upgrades mean expensive amplifier and terminal designs even if the product is brought out in its initial release as a 10 Gbit/s product. Similarly 25,12.5 and 6.25 GHz channel spacing means new technology filters and modulation techniques that create real technological and cost challenges for component manufacturers."