Content access control aims at ensuring that, in a system with several resources, users can only access resources they are authorized to. Resources are encrypted using cryptographic keys. Generating, distributing and renewing these keys are the challenges faced by key management schemes. While most of the existing key management schemes are typically evaluated by simulation. We propose, for the first time, to use Markovian processes for this purpose. Markovian processes give more accurate evaluation. The key tables-based key management scheme for linear hierarchies (KTLH) is a particularly interesting key management scheme; it was initially proposed for securing group communications, but could easily be adapted to other application such as wireless sensor networks. KTLH requires each user to maintain a set of keys. The keys and size of the key set change dynamically, making the evaluation of the overheads of KTLH a challenging task. Our contribution is threefold, we have (1) modeled KTLH using Markov processes, (2) evaluated KTLH according to its storage, computation and bandwidth overheads and compared it to existing key management schemes and (3) shown how our approach could be generalized to other key management schemes.