IDT and Level 3 highlight new interest in LMDS for mobile backhaul

The LMDS spectrum (28-31GHz in the US) was widely discredited as an asset after the broadband wireless crash at the turn of the century, but in recent years there has been a  significant revival of interest in this band - no longer for wireless consumer access, where it failed at the turn of the century, but for backhaul and point-to-point links. But IDT Spectrum, which bought up the licenses held by bankrupt broadband wireless operator Winstar for $42.5m at the end of 2001, still does not expect to get back the $300m it has spent trying to create a national backhaul network. However, eyeing the interest in backhaul capacity from cellcos and cable operators, it has decided to sell up its portfolio, in a move that highlights the decade’s corkscrew twists in the value and business models for high frequency spectrum.

The LMDS band was originally seen as a vehicle for the first wave of deployments of wireless networks for broadband access and three key operators emerged in the last years of the twentieth century, backed by huge and speculative sums from major players like Microsoft and Craig McCaw. These were NextLink, Winstar and Teligent, but all three failed in the telecoms crash around 2000 and ended up in the bankruptcy courts. This was not just a factor of the economy but because they were using the wrong technology at the wrong time. Only expensive proprietary kit was available for the high frequency, line-of-sight bands, and this was being turned to leased line replacement and consumer access services, for which demand was, at that time, immature and unproven. The business case would not add up until that demand was higher, and low cost standards-based equipment available with mobile and non-line of sight capabilities (in the second broadband wireless wave, driven by WiMAX from 2004).

The three LMDS companies were all acquired at knockdown prices from out of Chapter XI, and gradually the real potential for their spectrum holding was recognized - to support the exploding requirement for mobile backhaul (and enterprise links), created by the expansion of video IP and 3G traffic and denser cellular networks. IDT bought Winstar, whose spectrum covered all 50 states and averaged 615MHz in the top 200 markets. It held 16 LMDS licenses and 931 licenses at 38GHz, for which Winstar had originally paid $200m at auction, and is particularly strong in some top cities like New York and Chicago.

First Avenue Networks bought Teligent with its 350MHz of capacity in 75 markets and used this as the basis of a new backhaul business model, under the company’s newer name FiberTower. And backbone provider XO Communications snapped up NextLink, the largest holder of LMDS licenses, with 101 in total, covering 173m of the population and with about 1GHz of spectrum. It recently set up an LMDS spectrum exchange, another innovative approach.

The new spectrum owners set out to create wireless equivalents of wireline backbone businesses like Level 3, selling capacity and services to US cellcos as they started to follow their European cousins down the microwave backhaul route (a slow process in the US, because of low cost leased lines, but driven by the new capacity and flexibility requirements of 3G). But it is possible they have underestimated the US operators’ desire to control their own backhaul, and the price competition that the big three introduce because of their ownership of so many backhaul links of their own via their wired infrastructure.

This is likely to be driving IDT Spectrum’s sale. Its president Michael Rapaport believes wireless operators could see the licenses as useful assets and said last week that he was already receiving enquiries from large carriers before the decision to sell. “We have known for a long time that there’s a tremendous need for our spectrum,” he said in an interview. “It’s a function of maximizing shareholder value.” IDT Spectrum’s original plan was to build a backhaul network, and it spent about $300m doing that before shifting to a less capex-intensive leasing model. “We were five or six years too early for the marketplace,” Rapaport commented. He plans to sell the spectrum as a single nationwide bundle rather than market-by-market.

As well as cellcos, cablecos could be bidders, since they have been trying to build up their own businesses in selling backhaul for cellcos - mainly over their own infrastructure, but some could look to add wireless. Time Warner Cable, for instance, claimed in 2006 that it wanted to be the “first alternative access provider that can truly address the majority of cell towers” in order to break the telcos’ “virtual monopoly on cellular backhaul”. IDT could also see interest from tower companies, especially those wanting to offer added value backhaul services as FiberTower has pioneered.

So what of the other LMDS holders? NextLink/XO and FiberTower have adopted different strategies to those of IDT to maximize the value of their spectrum and generate new revenue from the mobile backhaul boom. In April 2008, NextLink announced an online marketplace for licensed LMDS spectrum, to allow businesses, government agencies and service providers to lease capacity. LMDSXchange, which grew out of a coalition of members, covers more than 80 major metropolitan markets in 40 states.

“We could potentially lease links to build out entire networks,” NextLink said. “We’ll provide installation, we’ll provide maintenance management of that link and we’ll repair it as needed.” The new business model is part of the company’s strategy to validate and expand the use of LMDS to a variety of applications. As well as leasing, NextLink offers wireless backhaul directly to cellcos and provides business services to large enterprises via resellers such as Covad.

FiberTower, which merged with First Avenue in 2006, had spotted the cellco opportunity even earlier. The combination created a wireless backhaul installed base of more than 1,000 sites and national spectrum holdings in 24GHz, LMDS, 39GHz and common carrier bands, along with customer relationships with five of the top 10 US mobile carriers.  With backing from all three US major cell tower owners - Crown Castle, SpectraSite and American Tower - as well as venture capitalists, FiberTower devised a wireless backhaul service for towers that it says reduces service outages by 50-60% for lower costs than leased T1 lines. It uses 155Mbps point-to-point microwave links to move cellular traffic from the tower to a central aggregation point; multiplexes and grooms it; then sends it to the provider’s main switch via leased fiber.

A fourth player is Level 3 itself, which bought TelCove, together with 300 licenses covering 90% of the US population, in 2006. It is likely to try to augment its wireless capacity for cellcos and could be interested in IDT. Level 3 is the latest to jump, with a new service for cellcos in metro and rural areas, called Level 3 Tower Access.

The company has been offering wireless backhaul services for a decade but is now unleashing new services and more flexible options for carriers, to support the growth in their requirements. In the first phase, it hopes Level 3 Tower Access will tap into the wave of stimulus funded build-outs for underserved areas, which should start in 2010.

It will provide connectivity to its fiber network in unserved and underserved territories, offering cellcos the option to house their towers directly onto existing Level 3 network facilities. It says it already has over 300 ‘tower ready’ cell sites that have the rights of way in place to support new towers. The wholesaler will partner with other providers to build towers on these locations, which will be carrier neutral. These shared infrastructure approaches, often with a wholesaler managing the whole shared system, are becoming increasingly popular in other parts of the world for the savings in cost and effort to the cellco. In the UK, British Telecom is providing a backhaul service over its IP-based 21CN network for most of the cellcos. Level 3 itself is taking on a similar offering from FiberTower.