The Vehicular Ad-Hoc Network (VANET) paradigm offers the opportunity of extending Intelligent Transport System (ITS) by supporting its applications through vehicle-to-vehicle communications, notably in the areas where the infrastructure is inexistent, in failure, or overloaded. However, the complexity induced by ad hoc network management raises many challenges that have to be solved such as the sharing of bandwidth resources, the limitations on the duration of the connections between the vehicles, and the application-specific quality of service (QoS) requirements. Recently, the Chain-Branch-Leaf clustering scheme (CBL) has been proposed for vehicle-to-vehicle ad hoc routing that combines the information of road configuration, vehicle mobility, and link quality in order to build an efficient clustering connecting the entire VANET through a flexible backbone. This work presents a comparative study between the native Multipoint relaying clustering used in the Optimized Link State Routing (OLSR) and CBL scheme. The results show that CBL reduces significantly the routing traffic overhead compared to native OLSR, thus freeing up more bandwidth for ITS applications and reducing the IP delays for peer-to-peer applications.