Cost, port utilization, efficiency, and "green" make fiber to the enclosure and attractive design alternative.
By Rodney Castel, RCDD, NTS
When working with designers or engineers I often hear the phrase "because that is the way I have always designed it" when they discuss their choice of network architecture. Perhaps it is because some network designers are not aware that new standards offer them alternative choices. Or maybe it is a case of "if it ain't broke don't fix it." However, understanding all the available options is critical when looking for a network solution that will support convergence, intelligent building systems, data center reliability, Internet protocol (IP) everywhere, "green" buildings and more, all while saving money and improving efficiency.
This article compares three different architectures addressed in the TIA-568 standards: hierarchical star, centralized cabling and fiber to the enclosure (FTTE), also known as zone cabling. It evaluates how "playing in the zone" can help address many emerging network issues, while simultaneously saving as much as 50 percent or more when compared with the installed first costs of typical hierarchical star deployments.
The First Structured Cabling Architecture
In 1991, the ratification of ANSI/TIA/EIA-568, Commercial Building Telecommunications Cabling Standard, laid the groundwork for a unified way for delivering telecommunications services. Prior to this standard being developed, proprietary systems ruled, and there were no standardized solutions or practices. While the concept of a structured approach for premises cabling was not immediately embraced, it certainly was needed to tackle the growing demand for telecommunications connections for both voice and data. Today, standardized solutions are the norm and have enabled the progression of increasingly complex enterprise networks.
At the heart of the ANSI/TIA/EIA-568 standard was the hierarchical star architecture. In this network topology, all cables "home run" from the work area back to a common space, then known as a "closet," were connected via a backbone cable to the main computer room. Utilizing this approach made it easier to accommodate the myriad networking topologies of the day, including token ring, bus and point-to-point.
The specifications for the hierarchical star were based on the ability of the most common media of the day, unshielded twisted-pair (UTP) copper cable. While other media types were available-including optical fiber and coaxial cable-UTP became the workhorse behind the ANSI/TIA/EIA-568 standard. Its limitations are what set the stage not only for the hierarchical star architecture but also for future topologies.
Offers a Standards-Based Fiber Solution
In 1995, growing demands for alternative ways to deploy networks led to the adoption of the TIA Telecommunications Systems Bulletin (TSB)-72, which outlined the requirements for deploying a centralized network. In 2001, TSB-72 was incorporated into the ANSI/TIA/EIA-568-B.1 standard.
Centralized cabling designs use the high-bandwidth, low-attenuation and extended distance capability of multimode fiber to centralize local area network (LAN) electronics in one telecommunications room (TR) within a building. The cabling media extends from the main computer room all the way to the end user, without the need for an intermediate TR for distances up to 90 meters (m [295 feet (ft)]).
Ironically, while this is the first architecture that leverages the properties of optical fiber cable, it was designed with a copper mindset and was written around the limitations of UTP. While copper cabling is limited to a 90 m (295 ft) link length, fiber could easily support lengths of 300 m (984 ft). As a result, the standard requires the implementation of a splice point or interconnect within a TR for distances greater than 90 m (295 ft). The 90 m (295 ft) limitation was included to ensure backwards compatibility; to protect end users in the event that they chose to change their media from fiber back to UTP; and to guard against potential future issues of restricted distance.
· There are a number of important benefits to centralizing electronics, including:
· Long-term cost savings.
· Improved port and chassis utilization.
· Easier network rearrangements.
· Greater security.
· Centralized management.
· Fewer or smaller TRs.
By reducing the size of the TRs, network managers can reclaim valuable floor space and realize savings in the cost of powering and cooling the TRs.
However, adoption of the architecture has been slow, primarily because computer and telephone manufacturers still do not offer a fiber interface as a standard option on peripheral devices such as laptops, desktops, workstations and voice over IP (VoIP) telephones. This means fiber network interface cards (NICs) must be purchased and deployed separately and deploying power over Ethernet (PoE) becomes more difficult. This factor, along with the cost of electronics, means that centralized cabling often requires the highest initial investment.
Fiber to the Enclosure: Bringing the Benefits of Fiber Closer to the User
In 2004, the FTTE standard was introduced to meet the needs of environments such as airports, education, sports arenas, hotels, convention centers and industrial buildings that have long cable runs and need to frequently reconfigure their work areas.
FTTE was ratified in the ANSI/TIA-568-B.1-5 and TIA-569-B documents. The 568-B.1-5 document explains the cabling aspect of the TE, also known as the mini TR; the 569-B document describes the enclosure and space utilized for this new topology.
With this topology, fiber is deployed from the main computer room out to the work area and terminated inside of a telecommunications enclosure (TE) that can be mounted in the ceiling, on the wall, under the floor or in a rack or cabinet. See Figure 1. Fiber links can extend up to 300 m (984 ft). From the enclosure, a short length of UTP or fiber extends to the user's work area and terminates. The enclosure accommodates one or more small to medium switches, patch panels and power for the equipment. This mini TR functions much like a standard TR, but with less capacity and a few more restrictions.
While the benefits to deploying the FTTE architecture (described in the next section) are significant, awareness and adoption of the standard are still low. Many users simply are not aware that FTTE is supported by standards; others are concerned about locating the TE in the work space, especially when it is mounted in the ceiling. However, given the many benefits that the architecture offers, figuring out where to store the ladder may be a worthwhile exercise.
The Benefits of Deploying FTTE
FTTE is an architecture that offers benefits in performance, flexibility and cost. To begin with, the design frees up valuable real estate by eliminating the need for traditional TRs. Along with the space that can be reallocated, the TEs are less expensive to maintain because they require less power and little, if any, dedicated cooling to maintain. Since the TEs only store a couple of small workgroup switches, the amount of heat buildup is minimal, requiring only vents or a small fan for heat dissipation. Even when utilizing PoE patch panels, many of the TEs can handle the heat from two 48 port PoE switches simply by utilizing a fan. In addition, the architecture allows network managers to shut down zones that are not needed over the weekends or during holidays to help conserve energy.
Serving individual zones from TEs has other advantages too. Smaller zones can be easily customized to accommodate moves, adds or changes (MACs), more closely reflecting the needs of many companies. In addition, using smaller switches helps increase port utilization by more closely matching switch deployment to the actual number of users.
Finally, this design potentially offers the highest throughput. One of the problems associated with the hierarchical design is blocking. This occurs when larger switches (with 24 or 48 ports) are connected back to the main computer room with a single 1 gigabit per second (Gb/s) uplink. This introduces a data bottleneck, where 1 Gb/s is supporting 2.4 or 4.8 Gb of information. In comparison, low-density FTTE configurations are completely nonblocking, and even medium-density systems offer more throughput than a standard hierarchical star architecture.
Minimizing the amount of cabling deployed and reducing cooling and environmental costs also means that FTTE improves the environmental efficiency of a building. In a typical hierarchical star network, the addition of multiple rooms for network equipment and applications invariably leads to inefficiencies in cooling, power, redundancy and materials. With the zone implementation, the amount of cabling being deployed is significantly reduced, which helps to minimize the NEC requirement for the removal of unused or abandoned cable.
While the performance, space and environmental benefits offered by the zone concept are encouraging, the most attractive aspect of the FTTE architecture is the cost savings. According to the TIA Fiber Optics LAN Section (FOLS) cost model, available as a free download at www.fols.org, FTTE implementation can save network designers 50 percent or more over hierarchical star deployments, depending on the number and types of variables being addressed.
Additional Drivers for FTTE
It was mentioned above that FTTE architectures are ideal for office buildings or applications that need long cable runs and frequent reconfigurations. Two other areas where FTTE makes sense are in simplifying a building automation system (BAS) and in the data center.
The ANSI/TIA/EIA-862 standard, which deals with the BAS, recommends a zone approach for integrating all the various building subsystems such as fire-life-safety, heating, ventilating and air conditioning (HVAC) controls, access controls and lighting controls into a common infrastructure. Implementing this concept reduces the amount of cabling and pathways and spaces used, which reduces waste and energy and improves efficiencies.
This section looks at how data centers can be configured. The ANSI/TIA/EIA-942 data center standard, like the 568-B standard, addresses three topologies. The first is distributed, or a hierarchical star approach, where all the server cabinets contain access switches that connect back to the core switches housed in the main distribution area. See Figure 3. In this design, the cabling is limited mainly to intracabinet links with only the backbone cables going back to the main distribution area. This is an efficient use of cabling but a very inefficient use of network switch ports.
When every server cabinet has its own access switch, a lot of switches are spread throughout the data center, making it harder to manage and maintain each switch, especially in large data centers. If the average switch utilizes only 60 to 70 percent of available ports, it leads to a lot of wasted power, extra heat to cool and wasted cost for electronics.
The second approach for the access layer is to use centralized cabling. See Figure 4. In this design, all the switches are located in the computer room, with the cables from each server cabinet home run back to the main distribution area. The result is better port utilization, greater security, less heat generated in the server racks and a reduction in the amount of equipment being used.
The limitation for centralized cabling in the data center is the amount of cables coming from each server cabinet. These bundles of cables (copper, fiber or both) require more pathways and spaces for routing. The centralized approach works well in smaller data centers where the number of server cabinets is limited and only a small amount of growth is expected.
The third architecture is FTTE, or zone. See Figure 5. This design creates a zone of a given size that is replicated over and over throughout the data center.
In the zone deployment, each row of servers is connected to the appropriate number of switch ports. Those switches may be located in a server cabinet or in a stand-alone cabinet at the ends of the rows. This configuration keeps most of the connections within the row and the cabling mostly within the zone, so it does not require a significant amount of pathways and space. The zone switches are connected back to the main distribution area via a backbone link.
Utilizing the zone concept in data centers has the same advantages as in other environments:
· Saves money
· Reduces cabling
· Improves port utilization
· Limits the number of active electronics in the data center space
· Minimizes added heat within the server rows
· Is easier to plan, maintain and grow
With the right planning, zone utilization within the data center can be the most efficient architecture for current and future applications.
FTTE Offers a Solution to Build On
In today's economy, people are challenged every day to look for better ways to live their lives and perform their jobs in ways that reduce cost and help to minimize negative effects on the environment.
The trend is clear: As network speeds continue to increase, electronics will continue to consume more power and produce more heat. This in turn will require more backup power and air conditioning to maintain the expected performance of the network. While the hierarchical star has been a cornerstone for our industry, it was adopted at a time when speeds were slow, power consumption and heat were minimal, the number of connections was low, enterprise data centers were few and small and convergence was just a thought in the minds of visionaries. According to the proverbial saying, "If a hammer is the only tool you have, then every problem will look like a nail," but is this the way to address every network design?
With the standardization of centralized and FTTE architectures comes the responsibility of network designers to evaluate their situation and think outside the box. Progress is an active term. With the passing of time, the technological progress in our industry will require creativity and new solutions for solving the current and future challenges facing our network designers.
The bottom line? No single architecture will address every network's requirements, so the network designer needs to prioritize the issues most important for the environment and choose the architectures that will best meet those needs. However, if cost, port utilization, efficiency and green are the priorities, then FTTE or zone architecture is the best choice.
Reprinted with permission of BICSI News. www.bicsi.org