The Internet of Things (IoT) intends to interconnect massive amount of heterogeneous, smart devices, with the goal of interweaving the virtual world with the physical world. Smart Cities are typical IoT application domains, comprising networks with large number of sensors that survey environmental data in order to provide different services, such as City Mobility solutions that facilitate on-street parking and traffic flow monitoring. Standards and specifications defined by the Internet Engineering Task Force (IETF), like IPv6, the IPv6 over Low power Wireless Personal Area Networks (6LoWPAN) adaptation layer, and the IPv6 Routing Protocol for Low-Power and Lossy Networks (RPL) are cornerstones of the IoT. In this paper we carry out large-scale experimental evaluations in an IoT testbed of 251 nodes to analyze the performance of the IETF IoT protocol suite in such a large-scale network. We define a City Mobility Solution application, using traces from a commercial Smart City deployment and the commonly employed Contiki implementation of IETF IoT protocol suite. The results show that the out of the box Contiki IoT protocol stack is not capable of delivering a satisfying performance. However, after a thorough analysis of the initial results, a set of improved parameter configurations is derived that allows the network to achieve much higher performance. Among others, improvements of 60.39% in PDR and 63.67% in delay are achieved. Furthermore, the paper presents and discusses the technical solutions and best-practice guidelines for a specific City Mobility solution being developed by Urbiotica, a company with ample expertise in Smart City deployments.