When deploying a fiber network, one of the key factors used to calculate ROI is how long it will operate. After all, some copper networks have lasted 140 years in the UK, even if they can’t necessarily meet today’s high-speed broadband needs in their current form.
Optical fiber is inherently more fragile than copper. It is a particular type of glass (fused silica), with a typical tensile strength that is less than half that of copper. However, even though fused silica looks, and can feel, fragile and brittle, if correctly processed, tested and used it has proven to be immensely durable.
To assess the durability of any material it’s useful to consider certain attributes:
- Initial strength
- Rate of degradation
- Any flaws that can weaken it
- Reagents that can weaken it
- Its optical lifetime - as the silica must still be able to function satisfactorily
With this is mind, there are essentially four factors that will affect the longevity of your fiber network:
1. Surface flaws
Pristine silica glass that is free of defects is immensely resistant to degradation. However, all commercially produced optical fibers have surface flaws (small micro-cracks) that reduce the material’s longevity under certain conditions. The critical factor here is the stress intensity factor "K", related to the applied stress and square root of the crack depth. It means that "real" fibers - those with small flaws - suffer at first a slow growth in those flaws followed by rapid growth to failure.
To overcome this, reputable fiber suppliers carry out "proof testing", which stretches the fiber to a pre-set level (normally 1 per cent) for a specified duration to deliberately break the larger flaws.
2. Reducing degradation
The user is then left with a fiber containing fewer, smaller flaws that need to be protected from unnecessary degradation. This means primarily stopping the creation of new flaws by coating the fiber with a protective and durable material for its primary coating.
3. Installation strains
We know that stress is a major enemy of fiber longevity, so the protection task is passed to the cabler, who will ensure that the use of suitable strength elements limits the stress applied to the cable to much less than the 1 per cent proof test level. The installer then needs to ensure that the deployment process does not overstrain the cable.
Of the three techniques commonly used - pulling, pushing and blowing - only pulling creates undesirable stretching (tensile stress). Unlike metal, glass does not suffer fatigue by being compressed, and so the mild compression caused during pushing causes no harm to the fiber.
4. Environmental factors
Once deployed, the local environment has a big impact on fiber life. Elevated temperatures can accelerate crack growth, but it is the presence of water that has been historically of most concern. The growth of cracks under stress is facilitated by water leading to "stress corrosion".
You can check what the tendency of a fiber to suffer stress corrosion is by reviewing its "stress corrosion susceptibility parameter", much more conveniently referred to as "n". A high n value (around 20) suggests a durable fiber and coating.
Calculating how long your network will last
Bearing in mind the four factors above, how can you calculate the lifetime of your fiber network? To do this, a network planner needs to take into account two inputs:
- How flaws are distributed among optical fibers, using the Weibull Distribution
- The results of fiber strength testing, normally carried out by manufacturers
Combining the (predicted) flaw distribution with crack growth theory has generated several optical fiber lifetime models, which can be seen here.
In general, these models give a probability of failure for any given fiber km over a chosen lifetime, of somewhere between 20 and 40 years. For correctly installed tier 1 fiber, the failure probability over such a timeframe is of the order of 1 in 100,000.
In comparison, the chances of a fiber being damaged by manual intervention, such as digging, over the same timeframe is about 1 in 1,000. Quality fiber, installed by benign techniques and by careful installers in acceptable conditions should, therefore, be extremely reliable - provided it is not disturbed.
It is also worth pointing out that cable lengths themselves have rarely failed "intrinsically", but there have been failures at joints where the cable and joint type are not well matched, allowing the fibers to move - for example, due to temperature changes. This leads to over stress of the fiber and eventual fracture.
The evidence from the field
Given that the first large scale fiber systems were deployed in the early 1980s, how have they fared? The good news is that during the last 35 years there have been no large-scale outbreaks of fiber failure in correctly installed systems using tier 1 components. Breakages have been seen, but these are predominantly where fibers have been removed from cables and bent below their permitted bend radius.
In fact, provided fibers are stored and coiled correctly it is quite possible that they turn out to be stronger than we at first thought and perhaps the original flaws begin to heal with time and exposure to water under low stress levels.
As we’ve seen in other blogs, the biggest enemies to the carefully engineered reliability of optical fiber can be either humans or animals, rather than the fused silica itself. Leaving these non-technical issues aside, it is quite possible that fiber networks themselves could continue to operate, with upgraded optoelectronics, for many years to come - perhaps even for as long as their copper predecessors!