In this paper, we examine the tradeoff between fault tolerance capacity and protective (spare) wavelength utilization on a two-dimensional torus, which can be applied as a topological choice for the next-generation mission-critical Local Area Networks (LANs). The fault tolerance is enabled via applying a Four-way OptimaL Disjoint routing (FOLD) scheme to establish alternative (redundant) lightpaths. Along the spare wavelength utilization spectrum, four different Wavelength Allocation (WA) schemes, with varied spare resource demands, are studied and their wavelength efficiency in tolerating network failures is evaluated. The first WA scheme allocates no protective wavelength and hence provides zero fault tolerance capacity. The second and third WA schemes incorporate fault tolerance capability by enabling and allocating spare wavelengths, in a shared fashion, to redundant lightpaths at different levels of wavelength utilization. The fourth WA scheme allocates dedicated wavelengths to all alternative lightpaths such that no two redundant lightpaths need to share wavelength resources. A Monte-Carlo sampling assisted simulation on a 4×4 torus exhibits the performance-cost tradeoff and it is found that shared WA schemes achieve an appropriate balance between network reliability and spare resource utilization.