At HUAWEI CONNECT 2021, Huawei launched a novel L3.5 autonomous driving network (ADN) solution for data centers (DCs) and introduced its architecture, major features, key technologies, and use cases. As digitalization gains momentum, new applications are proliferating, and network devices are growing in quantity and category. This exponential growth puts a huge strain on DCNs, leading to more frequent network changes. One knock-on effect has been increasingly complicated network operations and maintenance (O&M). To keep up, more and more network engineers are needed. However, merely increasing manpower is neither sustainable nor practical. To address this, Huawei has consistently explored new ways to cope with the ever-increasing network complexity through technological innovations, finally proposing a vision for ADN — enabling networks to be highly autonomous, just like autonomous vehicles. Huawei defined six levels (L0–L5) of automation and intelligence for the ADN solution for DCs, as well as key capabilities and features at each level. The six levels span the entire network lifecycle from network planning, construction, O&M, all the way to optimization, paving the way toward self-fulfilling, self-healing, self-optimizing, and autonomous DCNs. ADN is currently going through a long-term evolution from L0 to L5. As well as researching ADN standards and technologies, Huawei is committed to joining forces with industry partners to build an ADN evaluation system as a standard reference to drive networks toward a higher plane of automation and intelligence. In September 2020, Huawei launched the L3 ADN solution for DCs, which implemented high-level automation on a single network. Post-launch, the industry environment has changed drastically in merely a year; the digital economy developed in full swing and became the key driving force of the global economy. To keep its ADN solution up to date with industry trends, Huawei joined hands with representative enterprises in multiple sectors to expand ADN exploration and practices. Since banding together, they have found that multi-cloud, multi-vendor heterogeneous DCNs already became the new norm, leading to siloed network management and diverse levels of network automation. To handle this, they reached a consensus that automating complex, heterogeneous networks in an end-to-end (E2E) manner should be a top priority to eliminate manual breakpoints throughout the lifecycle of DCNs. In response, Huawei released the L3.5 ADN solution for DCs in September 2021. The solution draws on a lineup of cutting-edge technologies, such as the industry's unique service designers, an open and programmable framework, and a unified database. With these strengths, the solution achieves high-level, E2E network automation capabilities, including unified and undifferentiated management and control of multi-cloud, multi-DC, and multi-vendor networks. Not only this, the solution is also capable of collaborative, flexible network orchestration and network-wide simulation. These features enable rapid provisioning of multi-cloud networks in seconds. Currently, DCNs are a cornerstone for rapid innovation and rollout of enterprise services. Implementing network as a service (NaaS) on DCNs can well help them achieve this. The implementation, however, is hindered by many challenges. For example, multi-cloud, multi-vendor networks lead to siloed network architectures and separated O&M views. Network engineers are buried in massive work orders and have no time for network optimization. This is where Huawei's L3.5 ADN solution for DCs comes in. It takes the lead in achieving unified, undifferentiated management & control, and collaborative, flexible orchestration of multi-cloud, multi-vendor heterogeneous networks. The solution can also seamlessly interwork with enterprises' IT management systems, implementing NaaS and accelerating agile innovation of enterprise services.