Owing to the increasing popularity of online social media, Web 3.0, and Metaverse, mobile users can access intelligent and self-sustained services more conveniently than before. However, the enormous computation burden and data flow associated with these services pose tremendous challenges for future 6G networks that strive to enable diversified applications with varying user requirements. To address these grave challenges, wireless edge caching and computing are widely acknowledged as the two most successful enablers. Meanwhile, the trend of composing large services/applications as a suite of small and independent microservices (MSs) is leading to the cases where functional units can be distributed over edges, accelerating the transition of telecom architecture towards distributed MS-based paradigm. This article conceives an overarching perspective towards such distributed MS-aware cellular networks (DMCNs), which embrace the conventional service caching (and computing) system as a special case. The studies are expanded based on two aspects. One is the fundamental analysis of network architecture and the other is the leading design principles, including the user’s behavior modeling, MS placement, MS migration, and the joint consideration. These key issues are compelling problems in practices, yet have been largely ignored by existing works. Several potential solutions to the aforementioned issues are presented, and by addressing the challenges and exploring the potential enablers, we can pave the way for the successful implementation and adoption of DMCNs in the future. The article also includes a pair of case studies to validate the effectiveness of the developed framework. Finally, research opportunities and future trends are discussed.