GLG News by Samuel Greenholtz

Implications: 1.      Some wireless providers, apparently including Vodafone, are looking for an integrated Ethernet over Copper (EoC) backhaul solution that includes both TDM and packet traffic. 2.      Actelis argues that Hatteras offers just an overlay network with its (EoC) device. 3.      Actelis admits that not all carriers will be attracted to its way of providing backhaul.

Analysis:  Actelis is testing out the mobile backhaul application in its Ethernet over Copper (EoC) solution in a number of customer labs (and may even be field trialing in some cases).   At first glance, it appears that the vendor has just added IEEE 1588 sync capabilities to its box. The pseudowire implementation is fully standards based.  However, both its clock synchronization and its FPGA implementation were developed in-house and a patent is pending.  

Actelis asserts that for the most part, Hatteras has been pushing EoC for backhaul as  only delivering data to base stations.  Actelis acknowledges this is a key application and that it has had success itself with this overlay approach.  But the supplier believes that it has limited applicability.  Actelis says it is converging the TDM circuits, which is already on the copper with packet traffic, with pseudowire.   While the supplier knows that Hatteras has been talking about a circuit emulation solution, Actelis expects it will have challenges in dealing with synchronization and holdover problems.  

Actelis asserts that unlike alternative solutions it will continue to deliver very accurate synchronization, if all but one of its copper pairs go down, and there will be no loss of traffic connectivity with the base station.  The supplier say its holdover implementation, such as the case of all of the cooper pairs being cut, is advantageous because its clock will be maintained on a remote ML650 to keep that base station from going down.  In effect, the base station comes online as soon as the data is restored.  Actelis states that one of the significant reasons that carriers do not want to use other pseudowire solutions is the long amount of time (as much as an hour) it takes to resynchronize and bring the network back up.  

It would be reasonable for skeptics to think back on other technologies in which all of a sudden the mobile backhaul card is played.  Actelis knows that it has to prove its claims in the actual marketplace.  And the supplier does not see its product as revolutionizing the mobile backhaul space – just helping it along.  In addition, it concedes that there will be some carriers that will still want an overlay network because they are not copper-pair constrained and they are fine with keeping legacy TDM circuits up and using EoCjust to do the data.  

Implications: 1.      Tellabs is a technology leader in WDM. 2.      In marrying these capabilities along with its PON experience, it could result in a high-margin solution. 3.      Tellabs’ recently announced plan is another indication to the telcos that it is going back to spending the resources on designing innovative products that can realistically make a difference in the market.

Analysis:  Tellabs’ move to WDM-PON gives it a fresh start in the passive optical network equipment space.  While other competitors are battling it out on price for GPON, the vendor will look to develop a product, in which no other solution will be close to being competitive.  Tellabs is hopeful that it will result in something similar to what occurred in moving beyond 1/0 cross-connects to the 5500.  It was able to capture the bulk of the wideband cross-connect market – and milk that opportunity for a good number of years.  

Tellabs is refreshingly being candid in saying that it “believes carriers are five years away  from needing WDM-PON.”  But it is also communicating to the telcos that it intends to get back to being a leading incumbent vendor in telecom for a very long time.  While it seems that Tellabs’ major strength in the future will be more on transport and switching gear beyond the last mile, at a bare minimum, it does not want to leave the impression that it will precipitously abandon the PON space.  

While Alcatel-Lucent and Calix offer high-quality GPON products, the track record of the former has been significantly less than stellar in WDM, while the latter really has no experience with the technology.  And even though Fujitsu has excelled at WDM solutions, at least in North America, it has not been as successful in attempting to penetrate the optical access market.  Therefore, Tellabs, at least on paper, should have a substantial edge over the rest of the market.

Implications: 1.      Sprint already knows the answers to “how people will use the [WiMAX] network.” 2.      It really does not need any “more insight into ‘data usage’ and traffic trends.” 3.      It has plenty of data from its EVDO Rev A experience to adequately project that the current lack of applications will hardly be sufficient for all of that bandwidth.  

Analysis:  If there were new applications generated with higher rate wireless services with EVDO Rev A, Sprint would probably be touting plans to go to the next revisions.  “The sudden ascension of Baltimore to leader of the WiMAX pack,” as pointed out in the initial Unstrung article on this announcement, is revealing.  Originally, Sprint talked about the Baltimore-Washington corridor as a whole.   And if anything, Washington itself was viewed as more favorable because of it being close to its North Virginia office as well as the attractiveness of potential government customers.  Although Baltimore is an NFL city, the first commercial deployment will garner less attention than it would in Washington or in Chicago.  

No time frames are given for when Sprint expects to finish its data gathering from its “ideal ‘model’” (potentially portending a less than aggressive roll-out) and no dates are provided for the next two cities.   As the service provider apparently meanders in Baltimore, it will be killing time until operational control can be handed over to Clearwire.   

Sprint has certainly come a long way from the initial rhetoric of covering a minimum of 100 million people with WiMAX by the end of 2008.   

Implications: 1.      Several interexchange carriers (IXCs) can barely get 40G turned up. 2.      The high level of sensitivity involved with its operation means that technicians need to be very meticulous. 3.      Given the inherent problems at 40G, expectations are that 100G for long haul will be a tremendous challenge.

Analysis:  If craftspeople in IXC networks do not clean the connectors properly, it cannot get turned up.  A lot of times network managers have discovered that these technicians are not disciplined enough to do it right the first time.  Nevertheless, there are a number of engineers, including at AT&T, who are apparently adamant that 40G will be realized.StrataLight’s scientists are also working very hard to help smooth the way for the higher rate capability.  

The very tough obstacles with 40G make confronting 100G that much more daunting.  At present, there is the belief by certain long-haul planners that they will have to fusion splice everything in order to achieve this level of capacity – making its prospects questionable.  In addition, the lasers at 100G get incredibly hot.  So, it will require developing a laser, which is cost-effective, that will not burn up.       

Implications: 1.      Selling dark fiber IRUs should be a last resort for most wholesale providers. 2.      It offers a one-time, upfront big dollar amount, but then it is no longer a continuing revenue stream for the length of the contract. 3.      Obviously selling, let’s say, 15 OC-48 wavelengths is going to bring in a lot more revenue based on the same period of time of the IRU.  

Analysis:  A good rule of thumb for a metro provider is to keep the mix of business at about 70% for lit services and the rest for dark fiber.  Ideally, this kind of service provider should have enough revenue growth and network expansion to keep that model working to do both fiber sales as well as offer wavelength and other lit services (DS-3s, etc.).   

Level 3 is more of a backbone model with a tremendous amount of assets in the ground.  It is still operating in the first fiber bundle in the first conduit of what it deployed.  And Level 3 installed six or seven conduits throughout its network.  So, the service provider will easily have a much higher ratio of dark fiber assets for sale than a typical metro carrier.Still, the perception is that Level 3 at least to some extent might be cutting its own throat when it comes to selling fiber – in effect, killing revenue streams.   

Every route at Level 3 is different and in some cases it will have fiber in more than one conduit.  Some routes are much more highly utilized than others including those that go from Dallas to Chicago,New York to Washington, DC, San Jose to San Francisco – and to a lesser extent, – San Jose to Los Angeles.  In those situations in which there are huge amounts of traffic, Level 3 is not going to be able to do a lot of dark fiber sales.

Implications: 1.      Level 3 is expected to close several contracts in the next few months that will generate a huge amount of revenue.  2.      The agreements appear to be for both enterprise and wholesale services.   3.      The wholesale business is a lot easier to manage than retail.  

Analysis:  The new contracts mentioned above along with other opportunities could get Level 3 to a point of being cash flow positive in the not too distant future. It is also interesting to note that Level 3 has historically priced its services much higher for enterprise services than its smaller competitors.  Therefore, it does seem to have a cushion, as it gets at least somewhat involved in a pricing war on the wholesale side.  

Level 3 has been able to keep closer to the RBOCs on pricing than other carriers because it is only one of three companies in the U.S. to offer the service support of a major telco provider to big enterprise customers.  Sprint is pretty much dead in a lot of these areas.  Level 3’s footprint looks favorably compared with AT&T and Verizon, especially in Tier 1 and Tier 2 cities.   

Level 3’s wholesale business is much simpler to deal with than its retail operation.  There are just so many more end customers that can call at a drop of a bucket when they see anything happening on the network.  And most of the problems they have are with their own applications – not Level 3’s infrastructure.  So, they consume a lot of time to solve their problems on the carrier’s network.  Level 3 needs to extend and add more cost into the NOC and for customer support in working with these types of customers.    

Implications: 1.      PAETEC is based on the old and classic CLEC model.  2.      It is almost exclusively a reseller without any fiber assets in the last mile. 3.      It needs to get increasingly bigger through acquisitions to keep its revenue up.  

Analysis:  Any CLEC that is still around tends to have pretty decent management by definition. PAETEC is not going away anytime soon.  Its sales momentum for the rest of the year looks fairly good.  In later years, it will be a question of how long it can milk its voice revenue in a tough business.   

As voice opportunities continue to decline, PAETEC needs to keep expanding to show revenue growth.  However, it is like a squirrel in the cage – going around faster and faster, but not getting anywhere.  With an inability to ever get over the hump, its long-term value is questionable.  The best outcome for the company will be to eventually merge with another firm.    

Of course, voice will never go away – the implementation will just be different.  It is like music, which will not vanish either – in fact, there are people with more of it than they ever had because of the new tools to access it. And at least with voice, a business can actually put a dollar figure around it and extract value.  Voice always had a pay as you use model and nobody every questioned the concept.  The other models are struggling because customers can eat as much as they want for a monthly fee, while there are only X number of subscribers.  Service providers are attempting to start going to a fee per usage basis, but it will be tough to pull off because obviously nobody wants to pay for that extra cost.     

Implications: 1.      McLeod was a poorly run company. 2.      It operated in a tough service territory. 3.      But PAETEC did not pay much for it anyway

Analysis:  McLeod does provide PAETEC with impressive interexchange assets.  However, it appears to be little more than a marginal acquisition.  McLeod operated a lot in kind of the square states in the Midwest, where it was not easy to get revenue.  

With the superior management team at PAETEC, the results from McLeod will surely get more significant.But better supervision can only do so much to change a poor business. McLeod had to declare bankruptcy twice.  The integration effort will be a little tricky, but it will get done successfully.  PAETEC’sexperience with bringing in US LEC will be helpful in the process.  

Because McLeod needed cash, it sold some fiber IRUs.  Unlike Level 3, which has taken the same actions, McLeod never had that much fiber.  So, its inability to make any more money off these assets hit it much harder.    

Implications: 1.      The overall difficulty in obtaining funds in general in the current environment certainly makes it that much harder. 2.      Obviously, unless a FTTP CLEC project is financeable, the whole idea is toast. 3.       The simple math does not make it very attractive to VCs.

Analysis:  For the sake of argument, let’s say the cost is $2,500 to wire up a sub.  Several years down the road, perhaps it could be sold for $6,000 to $7,000.  No VC is going to get hot with that kind of return.  

So, the party involved needs to take on debt and turn it into a big leverage play.  Perhaps they hang in until it reaches two to three times what it costs.  But they only put 10 or 20 percent equity in the place, so then they can potentially get a big multiple.  

The CLECs that tend to be successful on FTTP have already been in the telecom business for five to ten years.  And as pointed out in a previous article, there can be no dawdling once there is a commitment.  But the individuals have to be very picky in terms of where they want to go -- and typically the development gets phased in.  The smart move is to only do one community at a time, start the cash flow and then it is easier to get financing.  In addition, being conservative in terms of the number of subscribers served is important.  It could be argued that a few thousand homes and a few hundred small retail businesses are more that enough.  One would be hard pressed to find any telecom provider, other than Verizon, that has done more than 5,000 homes passed a year in the U.S.    

Implications: 1.      It appears that the rationale of Cogent Communications “more strongly differentiat[ing] its pricing from its competitors” is the major reason for its aggressive actions. 2.      “Stimulat[ing] more growth in Internet traffic,” especially given its size relative to other carriers appears to be somewhat of a stretch. 3.      It is difficult to find any other carriers that “have seen Internet traffic growth slow over the past year.”

Analysis:  Google and other prominent companies appear to still be driving huge levels of demand for Internet traffic and carriers like Level 3 are picking up that business.  It is not out of the question that Cogent’s apparent low usage of its network is an isolated situation.  

While Cogent may argue that it is an “advantage…[to have] only eight products,” it can often be a weakness in that it pretty much just sells Ethernet.  Other wholesale guys are selling additional services such as OC-3s, DS-3s, OC-48s, wavelengths, etc. – and they are in a position to offer a fuller package at an attractive price – while not necessarily having to drop their Ethernet prices to Cogent levels.  

Rock-bottom pricing can have other potential negative ramifications.  There is the    problem with customers erroneously associating cheap with a low level of quality.  Also, in the past, Cogent in some cases had a hard time even at $10 a megabit – in terms of justifying deployment based on that price point and not having enough customers to pop a building and turn up service.  

In the final analysis, making long-term agreements for lower and lower prices is not the best way to go.  Ideally, a carrier needs to be able to keep it at medium-level pricing as long as it is under the tariffs that the LECs can charge for those types of services.  In the short term, the potential key benefit of Cogent’s actions “are balance sheet related.”      

Implications: 1.      Cogent Communications does not have a lot of fiber anywhere in its network. 2.      Its coverage tends to be concentrated in NFL cities. 3.      These limitations can result in customers that are looking to expand -- leaving Cogent to go to a service provider with a bigger footprint.

Analysis:  Cogent pretty much has a couple of strands of fiber across the country -- using DWDM and CWDM to manage and support all of the different customer traffic.  In certain applications, like going from a POP location into a building, it uses a single fiber – using a wavelength for transmit and another one for receive (such as 1550 nm in one direction and 1310 nm in the other).   

Cogent’s backbone has used Cisco Systems’ONS 15800 series DWDM gear.  The carrier also has a lot of Ethernet switches from Cisco, but may have avoided using the supplier’s high-end routers.  Cogent also made use of CDWDM as well as Cisco’s 15454 systems.  

Based on how much fiber it has and how its wavelengths are being consumed for IP  traffic, Cogent lacks the capacity to offer OC-3s, DS-3s, etc.  Obviously, other wholesale firms are selling such services.  And while Cogent covers all of the major cities, subscribers often want to get into secondary markets.  There will be a tendency to lose that customer who wants to use a one-stop shop – taking advantage of a larger infrastructure as well as the opportunity for better discounts.   

Yipes, which has a similar business model to Cogent, appears to be in better shape because it has a more extensive footprint that gets it into cities with less competition.  Of course, on the big routes, all of the wholesale providers are on them -- making it harder for Cogent to get a piece of the action.  In second and third tier situations, there is the LEC and perhaps a second competitor – and so Yipes is a position to more likely win the sale.        

Implications: 1.      A lot of people at the independent telcosare apprehensive that the USF and RUS funds could go away. 2.      It has resulted in some surprises in terms of FTTH equipment purchases. 3.      Suppliers such as Calix and Alcatel-Lucent might benefit if these seemingly irrational fears remain.

Analysis:  Some telephone companies get panic-stricken because they are no longer monopolies. And smaller telcos in particular are highly dependent on money from the USF and the RUS.  However, worries about the elimination of these funds appear unwarranted.  If anything, the way the politics in the U.S. is going, they will probably have access to even more cash in the next several years.  

Nevertheless, there are examples of carriers coming to the conclusion – why don’t we just fiberize everything while we can.  And in some cases, they do not care whether it is one farm for every five miles – they are deploying fiber.  Some of the justification for what appears to be an impetuous action is that they are probably not in the video business today.  So, they can get at least a little revenue (with apparently a pretty low hurdle rate).  Moreover, they will be potentially future proof for at least the next 100 years.  In addition, their cost to operating the network will be relatively low.  In short, they can only go so wrong in rushing to install fiber.  

These telcos tend to be part of non-profit cooperatives that really view FTTH as a benefit to all of the members of the community.  So, if it takes 20 or more years to pay it off, it is no big deal.  A couple of companies that were part of co-ops in west Texas installed fiber despite it costing them $30,000 a home served.

Implications: 1.      There are some inherent differences between Ethernet over Copper (EoC) and HDSL. 2.      The latter is much simpler. 3.      No supplier is expected to capture the entire EoCmarket.  

Analysis:  Adtran dominated the HDSL space in North America.  With EoC, dealing with all of the complexities of the technology itself, will probably not allow any one vendor to keep everybody else out.  

An HDSL could be put into anyone’s chassis.  In the front, one would get HDSL.   In the back, there would be a T1 or DS-x port.  With EoC, the supplier is not going to take a blade and just plop it into an Alcatel-Lucent 7300 or a Calix C7.   This is one of the ways vendors were able to be so successful.  It was very easy to develop a blade for the half dozen or so channel banks that were out in the market.  It is not so easy anymore.  There is now an aggregated GigE uplink that provides some significant challenges.   

Implications: 1.      Nortel has seemingly been quite proactive in making sure that it does not lose its high-end SONET interexchange business. 2.      It’s attractive, next-generation SONET gear is being offered at very competitive prices. 3.      Nortel’s proposals to service providers for its new ROADM are also appealing.  

Analysis:  Some telco engineers view Nortel’s OME 6500 as being very similar to the very successful FLASHWAVE 4500 from Fujitsu.  IXCs can do a direct launch of the 6500s into their base of Nortel 1600 MOR Plus gear.  Nortel’s amplifiers use a proprietary AM1/AM2 overhead byte in the SONET payload.  Even if the carrier uses the FLASHWAVE 4500, it would still have to buy Nortel transponders before it can launch the OC-192 into the DWDM.  So, it is cheaper and simpler in some cases to go with the 6500 instead.  Perhaps Nortel should be given credit.  It is about the only vendor that ever came up with a single fiber, high-density, DWDM to ensure its transponders are bought along with the SONET gear to use it.  Of course that is also the reason most of the large players dislike the box, since it requires them to buy from Nortel only.

Nortel’s ROADM is unique in that the carrier does not have to worry about Dispersion Compensation Modules (DCMs).  Its system has an electronic DCM.  Once the device is turned on, it just talks across the hops and automatically does the dispersion compensation equalization.  So, itdoes not matter whether, for example one adds a 2-mile piece of cable -- or is cutting a new node in -- it will automatically adjust. 

While Nortel's approach seems to be working, they are still struggling to get the RBOC's interested in their products again. This is one time that incumbency doesn't mean an automatic win-back.      

Implications: 1.      Fujitsu’s 7500 ROADM continues to sell very well. 2.      Those service providers that have maxed out on the FLASHWAVE 4500 capacity-wise can move to the 9500. 3.      Some engineers see the 9500 as being like an optical cross-connect.  

Analysis:  When Cerent (now part of Cisco Systems) came out with its low-cost SONET box, it took Fujitsu a while to respond and beat the newcomer pricewise. 

Also, when the FLASHWAVE 4500 was introduced, its backplanes were already beefed up for future applications, which had not even been developed yet.  

The advantage of the 9500 is that it has a 480-gig backplane today (next year, it will be up to 960-gig).  It offers much higher density on its -192s and -48s.  The 9500 also comes with a layer 2, Ethernet switch already in the box.  

Interestingly, some engineers view the 9500 as an enhanced optical switch.  In fact, they see it as being like Ciena’s CoreDirector.  

Fujitsu has developed a reputation at the service providers that for any product, which becomes GA, it works right at the outset without any hiccups.       

Implications: 1.      After spending several billion dollars on Cerent some years ago, Cisco Systems did not fully take advantage of the SONET product. 2.      Unfortunately, the large vendor has a habit of telling telcos what they need instead of listening to what they want. 3.      If Cisco had paid better attention, it probably could have derived substantially more revenue from what became the ONS 15454.

Analysis:  As the source article points out, Cisco Systems is still getting new sales for its 454.  Yet, even some minor refinements would have made a difference in growing its installed base of the product.  

The really negative experience came for the first buyers of the 454.  Network planners are fully cognizant of the future repercussions with going with any product with serial numbers that are in the first 100.  Certainly, at the time, the initial 454 boxes were a third of the price of other SONET gear.  At such a low cost, it was a good, enabling solution.  However, the initial purchases would have perhaps just software release 1.1 or 1.3 on it.  There would certainly be penalties to be paid for the first few years of using the equipment.  

Compounding this experience was the irritation felt by engineers because their  complaints were not heard.  One big issue was the relatively simple request of taking it from a 19-inch chassis and putting in into a 23-inch one because in the former, the carrier could not get all of the bandwidth out of it that the system could carry.   

One service provider, which had been an early customer, and had become 100 percent Cerent on the SONET side for a while, has close to 200 nodes of the product.  Instead of bothering to upgrade them, it actually plans to forklift them next year because it will wind up being cheaper to do so.  The carrier cannot get the benefit from the latest software from Cisco because of the limitations of the backplane.        

Implications: 1.      Level 3 burned $160 million in cash last quarter. 2.      It has also been selling off assets. 3.      More importantly, it seems to be turning its attention away from its retail business.  

Analysis:  Level 3’s CEO is apologizing to shareholders.  The carrier has in effect been mortgaging some of its future by continuing to sell fiber IRUs.  It sold off the advertising distribution business of Vyvx despite about a year and a half ago touting the uniqueness of its HD offering.  AT&T is increasingly looking like a major competitor in the CDN space.  Plans by both of the large RBOCs to build extensive IP networks could decrease their reliance on Level 3.   The carrier has been making a lot of noise about the relatively small opportunities in going after  “independent cable operators” for wholesale.  In essence, Level 3 is really struggling.  

Statements by the Level 3 CEO are illuminating such as “we should have stuck to some  of our knitting” along with the old rhetoric of “we’re the low-cost provider of a technical commodity.”  Despite the carrier having “fixed most, if not all, of the operation and provisioning problems that plagued it last year,” why is the focus apparently on “increased aggression in the wholesale space” instead of executing more on its enterprise strategy?  It is possible that Level 3 has lost some confidence in its ability to get itself out of the hole with its primary emphasis on retail?  Wholesale still represents the majority of its revenue and there are still way too many competitors in this market.  By applying tremendous pressure on price, it seems evident that it is trying to get more competitors to go out of business or get acquired.  Revealingly, among the future acquisition type of candidates mentioned by Level 3 is in the long-distance space.  

Level 3 built a good, low-cost network.  The problem is that its balance sheet does not fit the network.    

Implications: 1.      The acquisition of World Wide Packets is making a difference in unexpected ways. 2.      It gives Ciena another means to gain greater leverage in general at carriers. 3.      Operational cost assumptions in connection with WWP tend to be overblown.

Analysis:  CenturyTel as a whole uses a lot of WWP for Layer 2 switching.  The large independent telco has found it to be cheap and that it works.  CenturyTel engineers have also discovered that as long as they follow a few rules, they will not stub their toes too hard with the WWP equipment.  

Ciena had been aware for a long while that Century Tel’s long distance division, LightCore, is in the middle of a 2-year program of upgrading and replacing all of its core DWDM gear to a next-generation platform.  The supplier also apparently knows that LightCore is forklifting all of the Ciena DWDM out of its network and will sell it on the grey market.  

As a result, the vendor was initially informed that there was no intention to renew the maintenance contract for next year with LightCore. However, a little while later, after Ciena bought WWP, it was able to say to the long-haul group that it will still be doing business with it.  

Ciena has tried to sell the 4200 to LightCore, but despite feedback that it is a nice, little product, it was not what it wanted to deploy.  The manufacturer also could not make the case where the CoreDirector will pay for itself at LightCore.  

The “sugges[tion]…that operating expenses have increased” and that it is “likely from World Wide Packets…” is an exaggeration.  Ciena is in the enviable position that it needs to spend money on development because of impressive customer demand.  The idea that the supplier should be criticized because of great success is shortsighted.  One often needs to spend money to make money.

Implications: 1.      Timing is everything. 2.      It appears that the Ethernet over copper market is finally moving beyond early adoption. 3.      Even with the most positive view of the technology potential, six or more vendors in the market probably cannot be sustained. 

Analysis:  The optimistic outlook for Ethernet over Copper (EoC) is that at the end of the day, it will kind of be a redo of the T1/HDSL business.  A decent guess of the number of links deployed in that mature market might be around 30 million.  While it is not known whether EoC will reach that number, the argument is that people do not want T1s anymore.  They only desire a slug of Ethernet bandwidth.  With all of the copper wires still out there, EoC ought to be used where fiber is too expensive, “where companies don’t want to pay for a construction build,” or they are in a hurry for the service.  There is the acknowledgement that “the only limit to doing that…is the availability of copper loops capable of supporting the higher speed Ethernet offering.”   

The pro-EoC crowd says the underlying market trends are good and that most of the business is in front of them.  They point to the success of startup, Actelis.  But it is hard to believe that the space can support all of the other vendors as well that includes Adtran, Alcatel-Lucent, Aktino, Hatteras, Zhone and others.  Consolidation and exiting from the market are inevitable.  

One compelling application of EoC for ILECs is DSLAM backhaul.  With an asymmetrical configuration on the EoC device, it could be helpful to a DSLAM supplier in getting a lot of bandwidth downstream – especially appropriate for residential broadband.   

Even Time Warner Telecom (soon to be tw telecom), which has lots of fiber assets to deliver Ethernet services, takes advantage of a small number of EoC systems.  

There is also chip development going on that could significantly lower the cost of EoC.  And there are people studying the idea of getting even greater capacity with Ethernet over copper.     

Implications: 1.      When it comes to cities getting involved with fiber projects, no two stories are alike. 2.      It is not always easy to come up with a model that protects munis so they do not have to stumble over themselves trying to get into the telephone business. 3.      In bigger cities, it is an even more of a challenge with large competitors already existing.

Analysis:  Frequently, municipalities already run a utility, such as the water department.  They put into place pipes 100 years ago and they will be sending water down those same conduits 100 years from now.  This kind of monopolistic control by the city is not going to be the case when it comes to telephony.  While fiberizing the area can be done relatively easily, companies such as Comcast and Qwest will go crazy, especially in the larger urban areas – the city will be taking the food out of their kid’s mouths.  These government entities simply do not know how to run a competitive business.   

Often in talking to telephone people, munis finally start to get into their heads that they do not know anything about this space at all.  And so they start asking themselves do they really want to be doing this themselves.  

Some of the municipalities are doing an adequate job with their fiber programs.  There is a huge area in the middle between the two extremes of the city running the network and open access.  The key to success is a public/private partnership.  Team up with a service provider – an over-builder that is truly enthusiastic about getting penetration into the community.  Sign an exclusive agreement for triple play (the mantra for carriers looking to work with cities) with the provider and if it does not work out, the carrier can always be kicked out after a certain amount of time.  With the partnership, the cities can kind of ease into the technology with a knowledgeable company showing it the way.  

The advantage with cities is that for better or worse they do not mind taking on debt.  So, the fiber build-outs do not necessarily have to prove in within a short period of time.  

The use of non-recourse bonds instead of general obligation bonds eliminates the risk to the citizens.