Handicapping the small-form-factor connector race
Now that the market has been left to decide among the new small-form connectors, where do six competitors stand in their battle for acceptance?
Stephen Hardy Editor in Chief
Few aspects of the fiber-optics industry have generated as much competition and controversy as the new generation of small-form-factor (SFF) connectors. Touted as a significant step toward breaking copper`s stranglehold on horizontal premises applications, these components were the subject of a fierce battle within the Telecommunications Industry Association (TIA) standards community earlier this year. That internecine conflict produced no clear winner, and the vendors behind the different SFF offerings have now taken their fighting to the streets. The hope is that the standard that could not be established within the TIA could be created in de facto form by the marketplace.
But if standards bodies can`t choose among these connectors, how can the vendors expect the market to be any more decisive--particularly when a new SFF variant was introduced this past September?
Not surprisingly, each connector camp has a different answer to this question. As the beating of marketing drums may obscure these viewpoints, we asked representatives from each connector group to tell us where they stood on the questions of standards compliance, component availability, and market acceptance. We also asked them to handicap the race to the desktop. The responses to our questions showed clear leaders in certain market segments--but also that the contest is far from over.
Let the market decide
The positions of the different SFF proponents are best appreciated within the context of the evolution of connector technology and its role in bringing fiber closer to the desk. As discussed elsewhere in this issue (see "Taking fiber to the desktop may be only a question of when" on page 45), the premises market represents a gigantic and relatively untapped opportunity for fiber optics. Unfortunately, while carriers can absorb the cost premiums necessary to enjoy the benefits of fiber optics, the same can not be said as readily about premises applications. Thus, the extra money and closet space fiber-to-the-desk offerings traditionally required have slowed fiber`s march to the desk.
The appearance of SFF connectors about two years ago seemed at least a partial solution. Half the size of the standard SC connector, the new SFF component was roughly the same size as the 8-pin modular jack (commonly referred to as the RJ-45) used in copper applications. That meant cabling densities for fiber could now equal those of copper offerings; less rack and closet space would be necessary for horizontal fiber networks, and the cost of installing fiber would shrink accordingly.
The SFF connector seemed such a good idea that the TIA was urged in May 1997 to consider it for the new TIA/EIA-568B Commercial Building Telecommunications Cabling Standard. The recommendation was accepted, and five SFF connectors available at the time were put through their paces before one of the TIA`s working groups, which was expected to whittle down the five to one or two through several rounds of evaluation. The process seemed to work as intended, with one connector, the MT-RJ, left standing at the end to oppose the incumbent SC in a final polling this past February that many viewed as a foregone conclusion. However, for reasons that the industry may never sort out completely (see Lightwave, May 1998, page 44), instead of deciding on a single SFF connector or rejecting the SFF concept entirely, the working group elected to enshrine the existing SC as the standard wall-outlet connector and open up the closet to any connector that could be certified to meet certain performance criteria.
Thus, the TIA did not establish a "standard" SFF connector. Any of the SFF connectors that gain the necessary Fiber Optic Connector Intermateability Standards (FOCIS) documentation and meet the TIA/EIA-568A/B performance and reliability parameters will become "standards-compliant" and can be used behind the outlet. In other words, the TIA has left it up to individual users to decide which connectors will survive and which will disappear.
Naturally, the vendors behind each of the connectors (there are at least six SFF variants offered by U.S. companies, with the E2000 from Diamond S.A. and the MU developed by NTT also in the wings) stand ready, willing, and able to help the market make these decisions. The plan of attack encompasses two potentially synergistic fronts: familiarizing end users with the product and convincing subsystem and premises-equipment manufacturers to adopt their connector for use on transceivers, switches, and routers.
First out of the gate
The first SFF connector to make a significant impression on the premises market belongs to Panduit Corp. (Tinley Park, IL). The Fiber Jack (marketed as the Opti-Jack) made its debut in December 1996 and is FOCIS 6 on your documentation scorecard. The connector completed its FOCIS balloting in September, according to Rick Akins, fiber-optic product manager within Panduit`s Network Systems Division. The connector documentation is now being reviewed by ANSI and should be published at the beginning of 1999.
The Fiber Jack is a duplex connector with two 2.5-mm outer-diameter ferrules configured in a housing similar to the 8-pin modular jack. Akins says that comparatively wider fiber spacing makes the connector particularly good for transceiver applications. It is designed for use with singlemode and multimode fiber, and is now available as part of cable assemblies using both fiber types. Panduit has emphasized the cabling angle in approaching the market.
"Since we`re a premises-cabling company first, our goal had always been to market our product in the cabling system," Akins explains. "And then once the installed base was there, then we would go back to the transceiver manufacturers and hub manufacturers and be able to show them an actual installed base of customers."
Akins estimates that there are more than 200,000 lines of fiber in the field that are terminated with an Opti-Jack. Sprint is one of several major premises customers using cabling systems based on the connector, he says.
This number of cables represents an adequate installed base to move to the next phase of the marketing strategy, which is to approach equipment manufacturers with the idea of incorporating the connector in their products. Berg Electronics is the first licensee of Fiber Jack technology, and the two companies are discussing the development of a transceiver that Panduit would market. However, the company`s greatest success has been with MRV Corp., which released singlemode transceivers this fall that use the Opti-Jack (see Lightwave, October 1998, page 100).
More agreements can be expected, Akins predicts. "Because of the market acceptance that we`ve had, we`re starting to see a lot of interest from the transceiver and hub manufacturers," he reports. The fact that the connector is already being applied to singlemode applications in the field represents a strong selling point for these manufacturers, he adds.
"We`re trying to get across that there`s been a lot of talk about the other connectors because the transceiver people have been spending a lot of money on it, but there`s really no installed base. Whereas we`re doing the opposite--there`s a big installed base, and now transceiver and hardware manufacturers are starting to come out with products to address that base."
Who needs ferrules?
While the small size of the SFF connectors represents a technological advance in its own right, the VF-45 connector from 3M (St. Paul, MN) takes technological evolution one step further by offering ferrule-less V-groove mating (see Lightwave, July 1997, page 34; the connector was called the VG-45 at the time). The result is a connector generally agreed to be the lowest in cost among the SFF offerings. Unfortunately, a side effect is a reliance on a connection technology with which the user community has no experience, a factor frequently cited by 3M`s competitors.
The company is confident that the advantages of the technology will overcome users` fear of the unknown. "3M`s objective from the get-go was to look at all of the fiber interfaces that exist on active electronics today and be able to convert them over time to being a single interface, much as RJ-45 is the ubiquitous interface for copper connections on products," says Fred Hammond, manager of business development for 3M`s Telecom Systems Div. "We wanted VF-45 to be the same for fiber, so that basically you only had to have two kinds of patch cords in your facility, a fiber version and a copper version."
To support this concept, 3M made the VF-45 one of the lynchpins of its Volition fiber-to-the-desk cabling offering. Introduced in January of this year, Volition includes both cabling and electronics (such as media converters and switches), all of which feature the VF-45. Hammond predicts that approximately 75,000 lines of Volition cabling will be installed by the end of the year.
Meanwhile, 3M has been active with transceiver vendors, particularly Honeywell, Siemens, and Sumitomo. As yet, the transceivers that have reached the market from these vendors are all multimode, operating at between 850 and 1300 nm. Siemens is shipping modules for 100Base-FX applications, according to Hammonds; the transceiver manufacturer says it will have samples of Gigabit Ethernet and Fibre Channel versions available this month. (The VF-45 was selected as the standard interconnect by the Fibre Channel Association.) Besides moving to singlemode applications and higher speeds, the next phase of the transceivers` evolution will encompass designs that use the SFF package announced by such vendors as AMP, Hewlett-Packard, Lucent, Nortel, Siemens, and Sumitomo early this year (see Lightwave, April 1998, page 1). Current packaging options include 1ٷ, 1ٻ, 2ٷ, simplex, and quad-pack configurations.
With transceivers in place, the next step will be to convince equipment manufacturers to use these components and the connector itself with their products. 3M has already enjoyed success with Cisco, which used the VF-45 on its Catalyst 5000 Token Ring Module. These premises-equipment manufacturers are ready to embrace the SFF concept, Hammond believes.
"All of the top-tier OEMs, I would even say six months ago, did not have a small-form-factor, high-density fiber development program in place for 10 or 100 Mbits," he says. "The last six months of efforts from companies like 3M to really make available lower-cost fiber-to-the-desktop solutions has spurred on an end-user demand back to them. And all of them have development efforts ongoing right now. I`m talking about six companies: 3Com, Cisco, Bay Networks, Intel, Cabletron/Digital, and then IBM. You will see high-density fiber modules come out of all of them over the next six, in some cases nine months, but no more than that."
To capture the attention of these top-tier vendors--as well as other equipment manufacturers--3M has formed the VF-45 Action Group. The group consists in large measure of companies that have already adopted the connector for their equipment. In addition to 3M, charter members include BATM Advanced Communications, Canary Communications Inc., Corning Inc., Davicom Semiconductor, Gemflex Networks, Honeywell, Microcosm, Phobos, Racore Technology, Siemens Microelectronics, and Sumitomo Electric Lightwave. Other firms known to be working with the ferrule-less connector are ACT Research and Allied Telesyn. (The latter`s AT-MC30x family of media converters will support the VF-45, LC, and MT-RJ.)
Hammond says the best weapons in the group`s arsenal are the connector`s low cost and wide availability. The Volition family of products is priced 5% to 7% more than comparable enhanced Category 5 copper cabling offerings and 7% to 10% less than those of Category 6. Meanwhile, 13 distributors have VF-45-based products in stock and more than 100 systems integrators are promoting and selling the connector, Hammond adds. Meanwhile, its FOCIS documentation (the connector will be covered by FOCIS 7) is in the latter stages of balloting.
Feeling single
Every company attempts to play on its strengths. In the case of a company such as Lucent Technologies, that strength comes in the area of high-speed products for high-bandwidth applications, particularly for carrier networks. Thus, it should come as no surprise that the LC connector offered by the company`s Microelectronics Group (Norcross, GA) is considered among the sources consulted for this article to be the connector to beat for singlemode telecommunications applications.
Certainly Lucent brings a wide-angle perspective to the LC, the subject of FOCIS 10, which recently passed its final round of balloting. The worlds of telecommunications and data communications are merging into a single high-speed entity, says Dan Hendrickson, LC product manager at Lucent. Thus singlemode networks will eventually be the answer for networks throughout the enterprise, including horizontal applications. Lucent expects its LC will be in a position to be the end-to-end network choice when this eventuality becomes reality.
Lucent therefore has emphasized performance at high speeds in marketing the LC. The company has its own series of high-speed transceivers (the gigabit-rate SpeedBlaster) and has touted the connector`s support of 1000Base-SX over distances of 300 m as part of its OptiSPEED premises product line. This distance is 36% longer than competing product lines, according to John George, systimax SCS fiber offer manager at Lucent`s facilities in Atlanta. The distances are the result of the connector`s low (0.1-dB) channel loss, he says.
The LC has enjoyed success in the field. Premises users of the LC include Emory University, Ark Asset Management Inc., Baylor University Medical Center, Georgetown University Medical Center, and the Army College of Engineering. The company announced this summer that it has shipped 500,000 LC connectors.
Meanwhile, Lucent has enticed other vendors into adopting the LC for their products. Methode Electronics, Molex Fiber Optics, and Sumitomo Electric Lightwave will offer SFF transceivers with the LC; these units will feature light-emitting diodes, vertical-cavity surface-emitting lasers, Fabry-Perot lasers, and distributed feedback lasers. The three companies are also licensed to manufacture the connector, as is Senko Advanced Components. Polycore Technologies also will develop LC transceivers. At the systems level, Allied Telesyn has announced it will support the LC in its media-converter products; Transition Networks and LANart are said to be interested in offering similar support.
The target
No SFF connector has garnered as much attention--for better or worse--than the MT-RJ. Based on the MT connector developed in Japan, the MT-RJ is the product of a coalition that includes AMP (Harrisburg, PA), Hewlett-Packard (Palo Alto, CA), USConec (Hickory, NC), Siecor (Hickory, NC), and Fujikura (Tokyo).
The development of the product got a late start, according to Ken Hall, manager of network connect systems marketing at AMP, because the MT-RJ was originally based on another connector type, the mini-MPO. However, the team was dissatisfied with the latching technology, so they took the single-ferrule technology from the mini-MPO (which enabled them to retain the MPO`s performance levels and singlemode/multimode capabilities) and changed the latching mechanism to something similar to the MT. As a result of this switch, the MT-RJ trails the other original five SFF connectors in the pursuit of FOCIS documentation.
The switch also affected product availability--a factor the MT-RJ`s competitors have raised frequently within standards bodies and in the field. However, Hall says that cable assemblies have been shipping since this past April and transceivers had been in beta sites even earlier. The availability of product for field termination in jacks and patch panels has been the major shortcoming, Hall says, but these products are now rolling off the assembly lines as well. "At this time, availability is not a problem," he concludes.
Despite the initial hiccups, the MT-RJ Alliance (as the group calls itself) has created something of a juggernaut. The alliance boasts that more than 30 companies have agreed to support the connector. This support ranges from local area network equipment to integrated circuits (see Table on page 56).
As the list in the table indicates, the MT-RJ Alliance didn`t wait to build an installed base of users before it went after supporters. Its success in luring these companies onto the MT-RJ bandwagon can be attributed to the strength of the individual team members in their respective markets (particularly Hewlett-Packard in transceivers) and the momentum generated by the near miss within the TIA`s standards deliberations. In fact, competitors complain that MT-RJ salespeople speak of the connector as if it had indeed attained standard status.
Hall downplays the internecine aspect of the SFF landscape such charges suggest. "Everybody`s looking at replacing the SC connector, but technically the battle is between copper and fiber networking. So we took that on as more of the challenge," he says.
While much of their effort has gone toward acquiring adopters, the alliance hasn`t ignored the necessity of building a customer base. AMP has announced a multimode fiber-to-the-desk cabling offering, called netconnect Solarum, which features the MT-RJ.
Waiting in the wings
With all the attention that Siecor has showered on the MT-RJ, it might be easy to forget that the company has an SFF connector of its own that was developed in cooperation with IBM. The scdc/scqc connector (the "dual-contact" DC accommodates two fibers, while the "quarto contact" QC can handle four) actually preceded the MT-RJ on the FOCIS treadmill (like the MT-RJ, it is still in the preliminary stages of acceptance). However, the connector has remained in the shadows of the MT-RJ--a position it has assumed by design, according to Markus Giebel, product manager for both connectors at Siecor.
As Siecor views the situation, the MT-RJ is clearing a space for the SCDC, as the two connectors are more or less interchangeable. "We engineered the SCDC and the MT-RJ exactly the same," explains Giebel. "The only difference is one uses guide pins and has holes, and one uses rails--and these rails have exactly the same center spacing as the guide pins. Also, the fibers [in the SCDC] have exactly the same spacing as the fibers of the MT-RJ."
Thus, what`s good for the MT-RJ is good for the SCDC. "Our strategy is that people have two system solutions; one is symmetrical [the SCDC], one is asymmetrical [the MT-RJ]. Both of them are compatible--you can plug an SCDC into an MT-RJ," Giebel says. "The key point is that the SCDC and the MT-RJ are intermateable. People never understood the closeness of those two products because we never made it available to the public."
That said, the standard SCDC has a push-pull latching mechanism that is different from the mechanism of the MT-RJ. Thus, Siecor hopes to use the MT-RJ to convince users of the superiority of the single-ferrule/multiple-fiber design (versus the ferrule-per-fiber designs of the Opti-Jack and LC and the ferrule-less design of the VF-45), then let customers choose between the MT-RJ and the SCDC based on latching preference and the requirement for symmetrical connectors.
That choice will not be offered right away. "The reason we don`t do this [immediately] is we don`t want to confuse the market right now," says Giebel. "We want to make a push with the MT-RJ, and after that has been established, we`ll follow up with an SCDC transceiver, and then give the customer the ultimate flexibility of choosing."
To increase that flexibility, Siecor plans to offer new versions of both connectors. The company will apply its UniCam prepolished technology to release UniCam MT-RJ and UniCam SCDC connectors by the end of this year. In 1999, the company will introduce the QuickNet and DeskNet concepts. Both technologies will offer field installation without tools. The DeskNet will be targeted specifically at fiber-to-the-desk applications. Besides offering ease of installation, Giebel claims the DeskNet MT-RJ and SCDC will be price-competitive with the VF-45.
Who`s the new kid?
Given the amount of competition fostered by the five major SFF connectors, it might be hard to believe there would be room for additional offerings. But that possibility hasn`t stopped ADC Telecommunications (Minneapolis, MN) from unveiling the LX.5.
The connector appeared as a result of the company`s work in the area of fiber-management products such as frames. According to Sandra McWilliams, a product manager within ADC`s Fiber Optic Div., space concerns and cable densities have been ongoing problems for the company`s customers. A next-generation frame product introduced a few years ago addressed these concerns to some degree, but true savings awaited the development of a connector smaller than the SC. Hence, the LX.5 seemed a natural extension of the company`s product line, despite the presence of other SFF connectors.
"It is definitely targeted right in the same area as the others--well somewhat in the same area as other small-form-factor connectors," McWilliams explains. "We still believe that single-circuit access is of importance to our customers, so it is a single-fiber connector. And it can be duplexed very easily, so it does manage a pair. But for those customers who still want to test a single port or make changes to a single port, it does allow them to do that."
The company emphasized the use of familiar technology in developing the 1.25-mm ferruled product. "It takes a lot of today`s technology and just shrinks it down," says McWilliams. "So any customers who are concerned about epoxy studies and making too big of a change in what they`re doing, we think the LX.5 answers those questions with staying with the epoxy-based connector and the smaller ferrule just really making it a much smaller form-factor."
ADC hopes that the comfort of familiarity will combine with a proprietary angled-polish technique to adequately differentiate the LX.5 from the other connectors. The polishing technique provides the low return loss customary with angled polishing as well as the insertion-loss performance common to spherical polishing. The latching mechanism also is designed to minimize eye-safety issues. ADC has just started to assemble the paperwork necessary to acquire FOCIS documentation for the connector.
Despite relying on the tried-and-true in developing the LX.5, McWilliams reveals that the connector will not always follow traditional lines. "We think the future of the technology does call for a significant change in connectors," McWilliams says. "I don`t think people are ready to accept ferrule-less connectors today. But we do think that they are going to be a key feature in this marketplace in the future. And that will be the second version of the product family."
Thus, expect to see a second ferrule-less connector on the market. "That is a product we are still continuing to work on, and it will be some period of time before we release that, but it would look very similar to the ferruled version. It would just use a different alignment mechanism and no ferrule," McWilliams predicts.
Room for all?
Ask marketing managers about the future prospects for their product and you won`t be surprised to hear predictions of profits and success. What may be surprising in the case of the SFF connector competitors is that while establishing their offerings as the undisputed industry (and, eventually, TIA) standard is their ultimate goal, few think that goal is achievable.
"I`d love to say that there`s only going to be one [survivor]," says AMP`s Hall. "I suspect there`s going to probably be two, possibly a third. Everybody will play in their niche. The LC has got some strength in the telecom marketplace, just because of the Lucent and AT&T background. I`m sure they`re going to maintain a position there, so they`re not going to necessarily back down from that."
"I think you`re definitely going to see more than one," agrees McWilliams. "I really do think that the customers` applications are different, and they`re going to want to drive the requirements of the connector for their own particular use. A lot of customers, while they look to the standards organizations, they`ll make their own decisions anyway."
The continued existence of several SFF connectors is easy to foresee in the context of the overall connector marketplace, says Panduit`s Akins. "You still have FC connectors, ST connectors, SC connectors--even the quote-unquote standard SC still has not achieved the majority of the market share," he explains. "So I think, down the road, the MT-RJ certainly had a lot of early support because of the transceiver people. Conversely, we started real strong in the cabling side; now we`re getting more transceiver people onboard, but it`s still too early to tell whether that will go through. So those two are probably in the strongest position right now. And obviously Volition has a unique solution, so if the market really takes to that, they might also survive."
With so many choices to make, it`s probably not surprising that users are not expected to flock to a single connector. One question that remains unanswered, however, is whether the different connector offerings will add an unwelcome layer of complexity to fiber-to-the-desk offerings. As Hall pointed out, the first battle is not among the connectors, but between fiber and copper in horizontal applications. And as any general will tell you, dissention among the ranks is one sure way to subvert the effectiveness of any army before it goes into battle. u