Getting to the zone

Oct. 1, 2001
SPECIAL REPORTS: Installing Networks

New installation trends prepare local-area networks for a fiber future.


Fiber is on the move, and new installation practices and higher-density products are making the move more affordable for everyone. IT managers will not argue with the migration since new network-management software allows them to place active equipment closer to the end users without the old concern for checking individual ports. This progression of the active components, along with increasing bandwidth demand at each workstation, will require fiber to be pushed out of the telecommunications room (TR) and placed closer to the end users-and fiber product manufacturers are paving the road for installers to get there.
Photo 1. Multimedia workstation outlets support the logical progression from copper to fiber at the workstation.

Today's typical structured cabling architecture finds copper in the horizontal and fiber in the backbone. That hierarchy clearly makes sense since most LANs run some form of Ethernet and capacity tends to be an order of magnitude higher in the backbone than in the horizontal for Ethernet systems. Lightweight fiber cable is also more installer-friendly for the higher count runs from the main crossconnect (MC) to the TR. Since IT managers are just beginning to make the wholesale conversion to 100Base-T at the workstation, it is safe to say that fiber backbones and copper links in the horizontal meet today's networking requirements.

But what about future bandwidth requirements? It would be nice to be able to predict the future and thus choose the least expensive solution that satisfies many generations of network upgrades. Lacking an infallible crystal ball, it seems logical to assume that as 1 Gbit/sec to the workstation becomes more widespread, 10-Gbit/sec fiber backbones will become more of a requirement. To meet this need, it is imperative that new-building backbones include no less than 50-micron multimode fiber, the cabling medium currently recommended by the Gigabit Ethernet Alliance in their draft standard for 10-Gigabit Ethernet.

Hybrid fiber-optic cables solidify the network's future. Cabling backbones that include six singlemode fibers for every 12 50-micron multimode fibers will be optimized for the future migration that will undoubtedly prove necessary to satisfy growing bandwidth requirements. Extending fiber from the TR toward the workstation is imminent, and it makes sense to start planning for it today.

In the past, the primary LAN architecture was a copper homerun cabling system. Every workstation in the network had its own dedicated twisted-pair copper cable back to the MC. Bandwidth was limited by the cabling medium, and so were the distances from the MC to the workstation. Monitoring links required IT personnel to physically check ports in the MC. At its inception, this concept was the most logical network solution because extending fiber's reach was cost-prohibitive.

Over time, networks migrated to a more traditional star topology, utilizing active electronics in intermediate crossconnects (IRs) or TRs. Each workstation still had its own dedicated four-pair copper cable, but the horizontal cables now terminated in the TR. That allowed for mixed media in the network, with fiber dominating the building backbone. Bandwidth and overall distances improved by using IRs to regenerate signals, but this architecture still required significant space in the TRs and a tremendous amount of cable density compared to today's proposal.

An alternative approach is to extend fiber deployment to particular zones (or departments) within the building, then use copper from the zone to the workstation. A key factor in this new structure is the ability to remotely manage switched hubs (switches). No longer does IT personnel need to physically check each port in each TR.

The "fiber-to-zone" configuration consists of horizontal fiber cable (usually placed in ceiling plenum areas) terminated in the center of a workstation "zone." From this zone consolidation point, longer copper patch cords run to each station. The key to this architecture is the hardware required to house the consolidation point. It must provide termination points for both cabling media and support active electronics, most likely in the ceiling.

And where there are active electronics, there are air conditioning and backup power concerns. Certainly, this zone box will be both a challenge and an opportunity for product manufacturers. Although the cost of each zone box may be significant, the savings from purchasing an optical switch port for each user are significant as well-approximately 40% or more, according to our research.

The fiber-to-zone strategy works especially well for large cubicle groups or for very large work areas that have poorly spaced TRs. Fiber-to-zone also is relatively cost-effective since traffic back to the MCs may be aggregated over just two fibers. In fact, using just two fibers from the zone consolidation point would eliminate running an individual copper link from each workstation to the TR.

Furthermore, if a com pany should de cide to implement fiber-to-the-workstation in the future, it would already have a fiber backbone out to the workstation zone. Re placing the short runs to the zone en closure would be far easier and more cost-effective than tearing out entire copper links back to the TR to implement fiber-to-the-workstation. Another way to reduce overall costs is to install multimedia workstation outlets today (see Photo 1). These outlets will serve the copper patch cords in fiber-to-zone architectures today and support the fiber-optic connectors of fiber-to-the-workstation tomorrow.

The deployment of these fiber-extending strategies is anticipated to be sporadic due to interesting cost tradeoffs. An all-fiber network (fiber-to-the-workstation) is obviously the most expensive alternative because of the higher cost of optical-switch ports and network interface cards (NICs). Fiber-to-zone implementation reaps both the benefits of cost savings today (fewer optical switches and NICs) and perfect positioning for continued upgrades in the future (since the fiber backbone to the work area already will be in place).

These fiber-extending architectures are increasingly implemented due to the reduced cost of the cable itself. Fiber-optic cable has recently become very cost-competitive with the latest in copper-cable technology, but the perceived difficulty of fiber-cable and -connector installation continue to cause some installers to shy away from extensive fiber projects.
Photo 2. Two fibers are polished at once in the LC duplex polishing fixture.

Recent strides in this area have provided the market with no-epoxy/no-polish connectors. These connectors reduce labor costs tremendously since they take just minutes to install. Also, training is inexpensive because most students can learn how to install these connectors in just a few minutes of instruction.

In addition to the savings afforded by quick-install connectors, traditional epoxy and polish terminations are more economical than ever with the advent of small-form-factor (SFF) connectors. For example, thanks to a duplex polishing fixture, LC connectors can be polished in half the time that larger, traditional connectors require (see Photo 2).

Fiber-optic terminations are also using economies of scale within hardware and the connectors themselves. The new SFF connectors effectively double hardware capacity and reduce termination costs per fiber. Higher-density ribbon connectors of the near future will further close the remaining cost gap between fiber and copper terminations.

Other key factors that will affect the progressive installation of fiber are facility ownership (versus leasing), system up grade timeline, and the size of the network/facility.

Fiber will be a prime consideration if a company constructs a new facility to own, rather than leasing the building or office space. Cabling, unlike the active electronics, does not move when the lease runs out and the company moves to a larger facility. A fiber-to-zone solution also lends itself to leasing operations as the multitude of twisted-pair copper cables are replaced with an easily installable fiber-optic cable to each zone. New fiber-cable designs that offer smaller cable diameters are lighter and easier to route than their predecessors (see Photo 3).
Photo 3. New cable product lines are now available that are half the density of standard workstation cabling.

One of the most interesting reasons that fiber has not yet fully been pushed from the TR is the cost of a full electronic component upgrade (NICs and switch ports). Most of today's LANs are running some version of twisted-pair Ethernet. Companies that wish to upgrade some users while maintaining current capabilities for others will be best served by a copper-based system that uses a 10/100 switch.

This same system upgrade logic also holds true for all-fiber Ethernet, but upgrading to a fiber-based system like fiber-to-zone or even fiber-to-the-workstation from a primarily copper-based system requires the purchase of a new horizontal cabling system and new electronic components. That's the best way of truly futureproofing the network as the requirement for 10 Gbits/sec in the backbone approaches. Fiber-to-zone architecture reduces some of these costs, especially in optical-switch ports. Since Ethernet over twisted-pair has become the standard configuration for most applications, millions of copper Ether net ports have been sold, and a lot of them are still being used. So when it comes to upgrading, companies must decide when to upgrade to fiber and how cost-effectively they will be able to do it.

With the advent of the 10-Gbit/sec backbone in the enterprise, a full system upgrade is becoming necessary for companies that need to meet their bandwidth and infrastructure requirements in an Ethernet system. Those who previously converted to a fiber backbone and are already looking at how to push fiber further downstream in their networks will be the first to take advantage of this latest Ethernet advancement. Fiber-to-zone puts them in an advantageous position today.

Finally, the larger the network, the greater the need for fiber. Fiber is a far better choice when the physical distances between TRs are significant. Previously overcrowded TRs may even be omitted entirely. Fiber, with its large bandwidth over longer distances, allows that kind of flexibility.

The future is certainly bright for fiber-optic networks. Fiber is fully accepted and used as the backbone product, whether the network is operating 100Base-T or 10 Gbits/sec. Fiber-to-zone shows great promise for LAN architectures, since it eliminates an entire copper cabling run-from the zone consolidation point to the TR-utilizing as little as two fibers in its place. Installers will not mind pulling more fiber cable as they reap the benefits of lighter cable designs and easy-to-install, higher-density connectors.

The fiber-to-zone architecture is enabled by new network-management software and advances in product technologies, making installations much more affordable. Additionally, fiber-to-zone moves a company one step closer to the possibility of a future upgrade: fiber-to-the-workstation.

Mike Hoppe is fiber applications engineer, Jay Kochman is fiber sales support engineer, and Kam Patel is fiber marketing manager at KRONE Inc. (Englewood, CO, and North Bennington, VT). They can be reached at 800-775-KRONE.