Countries have laws and policies around motor vehicles in order to shield citizens from harm, protect important infrastructure and preserve the environment. However, enforcing those laws is difficult because of the scale of such a task as well as the presence of corruption and forged documents. One of the key ways governments have tackled this issue is by creating driver and vehicle registration systems where information is collected and accessed by relevant participants such as police, vehicle insurers, and government agencies. These systems often use centralized architectures that rely on a single party. Because this approach has limited transparency it can undermine the trust in the system. This is the case of Colombia. A country with a system that handles about 1.5 million procedures every month and is classified as critical national infrastructure. This system is nearing its end of life and the requirements for its replacement have been made public. According to these, the new system will use blockchain, a technology that can decentralize responsibility between several actors thus removing single points of failure, increasing transparency and trust. This would make it one of the first blockchain solutions in Colombia in the public sector, and globally one of the first uses of blockchain by a transit organization. In this paper, an architecture for the new driver and vehicle registration system is proposed. First the context of the system is described, followed by Architecturally Significant Requirements (ASR), constraints, design decisions, architecture views and validations.