From the January, 2014 Issue of Cabling Installation & Maintenance Magazine
The emergence of passive optical LANs has added a new wrinkle to the debate over copper-versus-fiber lifetime system costs.
By Patrick McLaughlin
Debates about whether an all-fiber (or mostly all-fiber) enterprise cabling infrastructure is more cost-effective than an infrastructure with a fiber backbone and copper horizontal have taken several forms. Advocates for either side have been able to make bona fide arguments based on any number of real-life cabling scenarios. For years a tool that has helped system designers, specifiers, installers and managers navigate the possibilities has been an interactive model developed by what was then called the Fiber Optics LAN Section of the Telecommunications Industry Association (TIA; FOLS). The group recently was renamed the Fiber Optic Technology Consortium (FOTC; www.tiafotc.org).
The Network Architecture Model, frequently referred to as a "cost model" for the practical purpose it serves, most recently was updated in 2005 and remains a guiding tool for parties interested in sizing up their architecture options. In an executive summary provided to users of the model, the FOTC explains, "Installed first costs are one of the main factors that influence which cabling medium end users and designers deploy in customer-owned local area networks.
"This model evaluates costs on an ‘installed-first-cost' basis--the cost of the entire network as installed," the summary continues. "In many cases, users consider the cabling costs separately from the electronic costs; this can lead to problems in the cost-effective management of the installation since using low-performing cable can mean a significant increase in electronics costs. Only by comparing first installed costs can one get a true, apples-to-apples comparison of a fiber installed to a copper UTP installation."
Though written more than eight years ago, that executive summary sums up quite well the crux of the ongoing copper-and-fiber comparisons, and is prescient of the comparisons' most recent iterations. Specifically, the emergence of passive optical LANs over the past few years has added a new wrinkle to the cost debate. This time, the fiber system used for comparative purposes is singlemode and the optical network equipment supported by this singlemode fiber is the historically high-priced long-wavelength variety.
A group called the Association for Passive Optical LAN (APOLAN; www.apoloanglobal.org) includes several members with a business interest in advancing passive optical LAN technology. While the group continues to build out its website, it does have a page titled "Why Passive Optical LAN?" that literally illustrates some of the technology's advantages. That page also lists both technical and economic advantages. It says, "Passive optical LAN provides substantial savings in capex and opex compared to legacy LAN designs." Among these advantages, it says, are the ability to eliminate wiring closets; the elimination of midspan electronics, power and cooling infrastructure; and other advantages.
In March 2013, The Fiber Optic Association (www.thefoa.org) president Jim Hayes presented a web seminar on the topic of optical LANs--OLANs. (Exact nomenclature for these technologies is evolving. Hayes uses the term OLANs to refer to passive optical LANs as well as fiber-to-the-office architectures.) Hayes has written and presented on this and similar topics many times over the past few years; the March 2013 event was in cooperation with Cabling Installation & Maintenance. As most often is the case, Hayes's presentation prompted many questions from audience members; Hayes responded to several of those questions during the seminar and to others offline.
It probably was summed up best when an attendee asked, "Do you have cost-saving analysis comparing copper and PON somehow, like per-user? Per OLT [optical line terminal]?" Hayes replied, "Each system is unique and needs a careful analysis. I disagree with suggestions that one can produce a cost comparison that is generic. However, various papers have detailed the cost analysis for specific systems and OLANs are usually cheaper, especially when facility and energy costs are properly accounted for."
There also was this interesting exchange. Comment from attendee: "Even if singlemode is very efficient (bandwidth) with 10G, 40G, 100G or more, singlemode is cost-unaffordable (active components) with those speeds in a LAN!" Hayes's response: "If you look at ‘traditional' LAN prices, the big IT suppliers charged approximately 10 times more for a multimode fiber port than a UTP port, and approximately 100 times more for singlemode. The FTTH [fiber to the home] suppliers are in a different market, where a FTTH connection costs a couple times more than a typical LAN UTP drop. OLANs are a disruptive technology because the companies that dominated the LAN equipment market are being displaced by companies that have been selling equipment to the FTTH market with a different market dynamic. Costs are significantly lower. … A box the size of an old desktop PC can serve 8,000 users and is fed by up to 10 Gbits/sec from the outside world ... The times they are a-changin'."
And the analysis of which network architecture is most cost-effective is changing with them. We will go into greater detail on passive optical LANs in our next issue.
Patrick McLaughlin is our chief editor.
Archived CIM Issues