6 Mistakes to FTTx Architecture Analysis

A 2013 study on Australia’s National Broadband Network, 3 years into the rollout, found that capex savings of over 25% on the FTTP rollout could have been easily achieved ­ with no flow on negative effect ­ simply through changes to the architecture. This national effort aims to bring robust broadband to all of Australia, and has extensive urban and rural components. A further 15% of savings were identified but not pursued as it may have caused an increase to the operational expenses. Below are a number of items that, if well planned, can make a tangible difference in the success of any FTTx network rollout.

Mistake 1: Only looking at homes

Many FTTx networks begin with a goal to serve homes. But getting to those homes means passing a whole range of other locations with revenue potential. Maybe they won’t fit into the business model, but not considering a full range of opportunities puts an FTTx network at a disadvantage from the start. Some customers want high bandwidth services, some want multiple services, and others want just a basic broadband service, but a well designed architecture will be able to cater to any and all of them without adding much or any complexity.

Mistake 2: Assuming all Multi Premises Sites are much the same

When we first think about sites with multiple premises we think of apartment buildings, but that is just one type. If you try to service a business park using an architecture made for apartments, you are likely to have a network that simply doesn’t match the requirement. Recognizing the variations that exist will help you develop a flexible architecture.

Mistake 3: Simplistic Modelling of network and demand

Draw an “average” street and you can come up with an architecture that achieves perfect utilization. Put that same architecture against a real suburb however and you may find it consistently fails to deliver what it is supposed to. Any proposed architecture needs to be run against a simulator to see how conforming designs respond to different densities, take up, existence or otherwise of duct and aerial infrastructure.

Mistake 4: Underestimating the cost of complexity

“It’s just one more tier in the network”; “It’s just one more technology in the mix”; or “It’s just one more service offering”. They may seem innocuous, but the cumulative effect can multiply. Not all complexity is bad, but recognizing the cost is important in determining whether to proceed.

Mistake 5: Assuming staff will follow instructions

Human behavior can be unpredictable, but one thing is certain, people will sometimes fail to read or understand or follow instructions. The result is delays, rework, and poor quality, which can happen at each stage of the process. But what does this have to do with architecture? There are a number of things that can be done in defining an architecture that minimizes the frequency or impact of these instances all the way through the chain.
Before looking at specifics, it is useful to understand the causes of this behavior. It’s not that people are malicious or careless — usually.

Mistake 6: Not planning for change

In our industry, the only constant is change. Our challenge is that fiber equipment such as cables and splice joints have an expected life of 20+ years, so how do we ensure we can get 20 years of valuable use from it? We can’t predict what will happen in one, five, or ten years, let alone 20 years. Nonetheless, ignoring the potential for change is a recipe for disaster.

How can we overcome these challenges?

Old methods of planning and designing FTTx networks can no longer support the complexities of FTTx deployments. We need to adapt with new technologies, compare the real underlying costs from plan, design and construction through to maintenance and we need to be able to be certain, to the best of our abilities that we’re making the right decisions.

1 reply
  1. Steven Wolszczak
    Steven Wolszczak says:

    I thought this white paper was quite comprehensive. Really important and relevant content, kudos. I’d like to make one general comment, and two specific comments. These are intended to amplify and reinforce the author’s content, respectfully.

    General: One of the beautiful things about standard PON FTTx architecture is that it IS, by design, flexible. After initial installation, there is generally good scalability, with OLT’s and ONT’s (Capex) only provisioned when services are taken. There is some promise of further flexibility in the future via XG-PON (10G) and NG-PON (Next Generation), the former offering additional upstream and downstream spectrum (not unlike the DSL, ADSL, ADSL2, VDSL evolution), and the latter combining WDM technology with PON architecture. Paradoxically, this perception of flexibility may cause one to over-simplify the requirement for detailed architecture planning prior to implementation, as the author points out in detail.

    Re: Mistake 4 – Complexity As a splice trainer and general OSP guy, it’s always staggered me the extent to which the planners have no or little knowledge of OSP processes. If you have 12 fiber ribbons and 6 fiber drops, you create some complexity, but if you have 12 fiber ribbons and 8 fiber drops, you create TREMENDOUS complexity. The first 8 fibers are simply a 8/4 split on ribbon 1. It could be done improperly (i.e. mistakenly) by splitting 4/8 instead of 8/4 but that’s all. On the next 8 fibers though, you need 9-12 from ribbon 1 and 1-4 of ribbon 2, so you have to get not only the 8/4 and 4/8 right, but to splice to an existing 8f, you want 9-12L/1-4R (left/right) in your splicer. You need to make sure you don’t have 1-4L/9-12R, 12-9L/1-4R (if you “flip” either partial ribbon), 9-12L/4-1R, or 1-4 or 4-1L/9-12 or 12/9R. I.e., there are far more ways to get this splice wrong than get it right. This can be solved by very good tech training, AND ADDITIONALLY by training of the PLANNING STAFF to understand and even observe live in the field splicing procedures, so as to 1) minimize the complexity of the plans and 2) maximize the “understandability” (my term) of the splicing instructions. As far as instructions, colors and graphics are much prefered to numbers and letters in black and white.

    RE: Mistake 5 Assuming staff will follow instructions. I would simply add two more bullets. 1) Staff simply doesn’t understand the instructions, and 2) The instructions are clear, but staff has never been trained in the process to implement the instructions.

    Summary: These are just a few examples of mistakes or problems, where the solution is PROPER TRAINING. At the end of the day, after all the modeling and planning, somebody simply has to cobble this thing together. I see this point overlooked all too often. Granted, I’m a trainer. I keep trying to tell people, “fiber techs aren’t born, they are trained”. This stuff isn’t just picked up by “osmosis”. Novices don’t go from, “would you like fries with that”, to, “we need to do a ribbon into an 8f and 4f fanout at splice point 13”, overnight. I could go into pages, and I do on my website as to how to solve the “training problem”, but i won’t here. Training is an investment, and simply the “cost of doing business” if you want to do this stuff, however, it doesn’t have to cost tons. A good “train the trainer”, partnering novices with veterans, keeping training records and checking off individual Skills can help. Coupling this with a minimum amount of professional training for the planners, splicers, testers, and managers helps. Also, training, like fiber architecture as the author points out, is NOT one size fits all or cookie-cutter, but must be assessed and customized. steve@midwestligthtwave.com CFCE, FOA Certified Trainer


Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *