TOWARDS A BLOCKCHAIN-BASED PARTIAL COMPUTATION OFFLOADING FOR THE METAVERSE IN IN-NETWORK COMPUTING

ABSTRACT

The Metaverse is anticipated to offer an immersive experience to many simultaneous users. Therefore, it is critical to optimally allocate computing resources to meet the massive users' demand. Meanwhile, the In-network computing (COIN) paradigm has emerged to reduce delay and meet quality of experience (QoE) by using unused network resources for performing some tasks. This paper considers the blockchain-based metaverse partial offloading approach, where the user's tasks can be partially offloaded to edge in-network computing (EIN) or a nearby fog in-network resource computing (FIN). First, we model partial offloading using an ordinal potential game from the users' perspective. Next, we modelled the problem as Markov Decision Process (MDP) from the network perspective and used a Double deep Q network ( DDQN) to obtain the optimal partial offloading policy considering changing users' demands and network conditions. The experimental results suggest that our proposed approach yield an optimal partial offloading policy for massive metaverse deployment in dynamic network condition over traditional baseline enabling massive metaverse task offloading.

Keywords: Blockchain, In-network computing, Metaverse, Tasks offloading