This thesis presents new developments on the different intrinsic failure mechanisms of silica optical fibres and fibre Bragg gratings with regard to applications in the fields of telecommunications and metrology. These failure mechanisms are mainly due to optical and mechanical degradations. Mechanical degradation is a combined effect of environmental corrosion, or ageing, and stress corrosion, or fatigue. Glass is subject to chemical attack by reagents in the environment, especially water. This ageing phenomenon can be satisfactorily accounted for by an Arrhenius model. High-temperature zero-stress ageing (up to 150 °C) at 85 %rh in air is proposed as a way of reducing the testing time of modern, acrylate-coated telecommunication fibres. Two types of research fibres were tested. The activation energies of 1.3 eV obtained for both types of fibres correspond to the values reported in literature for silica dissolution. Coating degradation and handling problems impose an upper temperature limit of 125 °C for the ageing tests. Assuming a temperature of 20 °C and a fibre handleability limit of 2 GPa, necessary for stripping and splicing, the best time to failure estimate is over 100 years. Environmental corrosion of optical fibres should therefore not be a lifetime limiting factor, when compared to external causes such as accidental cable dig-up. Glass is a brittle material subject to stress corrosion: crack propagation under any applied stress in a non-inert environment results in strength reduction. This fatigue phenomenon can be successfully accounted for by a power-law model. An improved fitting method is proposed and implemented, with which Weibull strength distribution and stress corrosion parameters are obtained simultaneously. Investigations were made to validate the use of the power-law model in the case of sinusoidal cyclic loading. This is of special importance for lifetime estimation of systems experiencing vibrations. The experimental results perfectly agree with the predictions. Time to failure and reliability of commercially available stripped & polyimide (re)coated fibre Bragg gratings operating at high temperature in an inert gas were also investigated. The samples are characterised by bi-modal Weibull strength distributions that probably result from the manufacturing operations. A newly developed model considers that only a fraction p of these samples suffers from strength reduction. The mean obtained values of p = 27 ± 4 % and p = 16 ± 3 %, respectively, can be considered as a figure of merit of the manufacturing process. High-strength proof-testing is proposed as a way of ensuring a minimum strength. Lifetime of the surviving samples in excess of 5 years under 0.5 % static strain are achievable. Optical fibre Bragg gratings are tuneable band-stop filters. Their optical degradation was investigated in terms of time- and temperature-dependent decay. This thermodynamic effect is due to the thermal depopulation of the excited defect sites responsible for refractive index modulation. A new step-wise temperature annealing method is proposed to model wavelength shift and reflectivity decrease with a master curve. This allows to obtain model parameters with one single grating. Proper pre-annealing of both types of grating tested (same as tested above) guarantees a wavelength stability of 1 pm during 5 years at a temperature of 230 °C, with a reflectivity decrease of less than 20 %. Therefore, error-free operation can be assured through wavelength drift reduction to levels compatible with industry requirements. These results demonstrate that optical fibres and fibre Bragg gratings have evolved from an experimental curiosity to a fully mature technique, suitable for long-term applications in both fields of telecommunications and metrology. Given sufficient time and population size to obtain accurate enough information on their mechanical and optical characteristics, their time to failure as strain and temperature sensors or telecommunication components can be predicted with satisfactory accuracy.