The massive ultra-reliable and low-latency communications (mURLLC) services are emerging as a new traffic type to support massive numbers of mobile users (MUs) demanding the stringent delay and error-rate bounded quality-of-services (QoS) requirements over 6G. Among multiple 6G mURLLC services, digital twins (DT) has been widely envisioned as a major intelligent application to support efficient interactions between physical and virtual objects. Moreover, multi-tier caching, which is one of the key distributed computing techniques, stores the frequently-demanded data items at different wireless network tiers to efficiently reduce mURLLC streaming delay and data move. However, how to efficiently cache mURLLC-based DT data items at different caching tiers of wireless networks and how to statistically upper-bound both delay and error-rate for DT communications remain challenging problems. To overcome these difficulties, in this paper we propose a multi-tier caching mechanism to support DT communications over 6G mobile networks. First, we propose the DT data adaptive collection scheme applying finite blocklength coding (FBC) to dynamically encode a physical object into its virtual representation according to the current network and wireless channel statuses. Second, we develop inter-tier and intra-tier collaborative caching mechanisms, where DT data items are selectively cached at different wireless network caching tiers according to their popularities including: router tier, massive-multiple-input-multiple-output (MIMO) basestation tier, and mobile device tier. Third, our proposed intertier collaborative caching mechanisms maximize the aggregate epsilon-effective capacity across all three caching tiers, and our proposed intra-tier collaborative caching mechanisms minimize the sum of data transmission delay for all DT data items cached in each caching tier. Finally, we numerically validate and evaluate our developed multi-tier hierarchical caching schemes over 6G DT-enabled mobile networks.
