FTTE battles for enterprise/SAN acceptance

Th 315797

by Stephen Hardy

Fiber to the telecom enclosure (FTTE) is often portrayed as an interim step toward fiber to the desktop. However, like other fiber-based architectures, its acceptance in the enterprise faces challenges.

Just as fiber required several interim architectural steps before carriers seriously considered using it to connect homes, the technology has crept closer to enterprise desktops and servers one implementation stage at a time. Only the most data-hungry applications currently require fiber from one end of the LAN or SAN to the other. Network managers often can address escalating bandwidth requirements--including those of many data centers--through the enterprise equivalent of fiber to the node. This architecture, described in standards as fiber to the telecom enclosure (FTTE, which is sometimes called fiber to the zone or FTTZ), brings fiber to a point no more than 100 m away from the worker or data center rack. However, there appears to be some doubt about whether enterprise parallels with the evolution of fiber in the outside plant will continue.

Th 315797

FIGURE 1. The traditional hierarchical star architecture usually includes a telecom room full of crossconnect or interconnect equipment.

If such parallels do founder, it won't be because of a lack of effort by the TIA's Fiber Optic LAN Section (FOLS; www.fols.org). Rodney Casteel, chair of the TIA FOLS, and manager, advanced technology services, CommScope Enterprise Fiber Solutions (www.commscope.com), reports that his organization has worked hard to educate network managers about the benefits of an FTTE approach versus the commonly used hierarchical star network design.

Take a multistory building as an example. In a hierarchical star, a central equipment room houses a core-class Ethernet switch (see Fig. 1). Multiple fibers run out of the equipment room and up the riser to form the network backbone. One fiber terminates in a telecom room (TR) on each floor via, in most instances, a 1-Gbps interface. The TR contains patch panels, workgroup switches (commonly 24-port systems with two 1-Gbps fiber uplink ports), and, if desired, power over Ethernet (PoE) equipment. Unshielded twisted pair (UTP) copper cabling then connects the switches in the TR to each user on the floor.

FTTE architectures take fiber closer to the desktop (see Fig. 2). The hardware in the equipment room remains the same. However, more fibers travel up the riser and run through the TR, because the relatively large switches needed for the hierarchical star have been removed from the TR and replaced by smaller systems located in multiple, smaller enclosures (hence the name FTTE) between the TR and the users. The same UTP connects the enclosures to the desktop.

Th 315798

FIGURE 2. A low-density FTTE architecture has many of the same elements as a hierarchical star. However, the fiber-optic backbone cable passes through the telecom room without the need for splices or other connections. Consequently, the TR on each floor can be smaller than the size recommended in TIA standards.

The FTTE approach has several benefits versus the hierarchical star, according to Casteel. “With the traditional hierarchical star, no matter what media I used, I'm limited to 100 m in the horizontal,” he begins. “So with the FTTE, I'm allowed to go the full 300 m that fiber can do.” This enables users to be located farther away from the TRs without straining the data-carry capabilities of the network, for example.

FTTE also promotes higher throughput to each user, Casteel says. In a hierarchical star, a larger number of users share the same 1-Gbps uplink capacity. The workgroup switches frequently become oversubscribed. FTTE promotes nonblocking architectures and provides users (or servers) a fatter pipe because the 1-Gbps uplink supports fewer user-side ports.

Finally, the TIA FOLS premises cost modeling program (available on its web site) reveals that an FTTE approach usually saves money over the hierarchical star, Casteel adds. The fact that the final connection to the desktop or server remains copper-based means that network managers don't have to replace existing network interface cards, which adds to the architecture's economy.

Love it or hate it

Despite its advantages, Casteel concedes that FTTE hasn't become ubiquitous since the TIA/EIA ratified the standards that govern its use in 2005. The fact that two standards address FTTE--TIA/EIA-568B.1 Addendum 5 covers cabling specifications and TIA/EIA-569 treats the enclosures themselves--is part of the problem. “Because it's not a standalone standard--it's actually composed of two different standards--a lot of people don't recognize that it is a standard,” Casteel explains.

That said, its cost benefits do find appeal. “If you talk to a facilities guy, he is all for the FTTE design--they love it,” Casteel says. “It does save you space and money--it's the lowest-cost architecture that you can use for most environments.” This is particularly true for new construction, he points out.

Unfortunately for fiber proponents, facilities managers aren't the only decision makers. “If it's the IT guy you're talking to, they absolutely hate it, because it means all of their electronics are distributed now instead of in one location,” Casteel points out. IT managers can become particularly possessive of their equipment-filled TRs if they had to fight years to get budgetary approval for them, he says.

Add to the standards confusion and IT manager stubbornness the general bias against fiber in applications where copper has been used comfortably for decades, and you create significant barriers to market penetration. For this reason, “FTTZ will not become mainstream,” offers Frank Murawski, president of market research firm FTM Consulting Inc. (www.ftmconsultinginc.com).

On the plus side, “FTTZ has applications in large venues such as sporting arenas, airports, [and] warehouses that require extended runs of cable that cannot be supported over UTP but need fiber,” Murawski concedes. Casteel adds that he's seen FTTE deployments in education networks and hotels and casinos as well.

Casteel also sees data centers as a prime application for the architecture. “The zone concept is pretty much the best approach for data centers,” he asserts. “I would say most of the data centers that I personally have worked on have applied the zone concept.”

In the data center, the horizontal part of the FTTE architecture would connect to a rack of switches that oversee multiple servers. The use of FTTE provides a modular approach that lessens the amount of required cabling, reduces the number of switches, and improves port utilization dramatically, Casteel explains. It also increasingly requires 10-Gbps connections.

However, even here a large obstacle has appeared. “Cisco is pushing for a different architecture,” Casteel says. “They're going back to a top-of-rack scenario, where you're putting switches in the top of each rack or cabinet. So they're really trying to push that architecture because they get to sell a lot more switches.”

Bridge to nowhere?

Finally, FTTE's market potential must take into account the anticipated increase in the desire among enterprise staffers for mobility. And mobility means wireless LANs.

“We saw wireless as just an overlay,” Casteel says of most players in the enterprise market niche. “Now I'm seeing it going in as a primary [communications medium]. So I think that mindset will continue to change as the wireless capabilities keep increasing.”

The speed with which the popularity of wireless grows clearly will have an impact on the future of FTTE and wired enterprise networks in general. Therefore, FTTE could be a transitional step toward fiber to the desk if bandwidth rates continue to rise rapidly. Or it could potentially prove a bridge to nowhere if wireless takes over the applications space.

Stephen Hardy is editorial director and associate publisher of Lightwave.


Cabling Installation & Maintenance: Bringing Increased Bandwidth Ever Closer to the User

TIA FOLS: Cost Model

Lightwave: Is Power over Ethernet Anti-Fiber?

More in Network Design