In DiscoLedger, we build on top of the results of the DiscoEdge project. There, we have studied how to share resources in today’s mobile networks that are populated by all sorts of devices that offer ubiquitous sensing capabilities and disparate categories of online services to mobile users, as well as a wealth of processing power, inexpensive storage and a wide range of computational and networking resources. In contrast with the traditional Cloud computing model in which the vast majority of online services and applications offload computationally expensive tasks to centralised and large-scale Cloud infrastructure services, we have shown the convenience of relying on network, storage and computational resources available in the proximity of users, i.e., in the Edge. These novel paradigms can effectively leverage an ecosystem of resources distributed all over the communication devices at the edge of the network (e.g., base stations and MEC hosts) and in the user devices (e.g., smartphones and IoT devices). In that context, we also explored economic and sociological challenges to guarantee user trust, fairness and security when accessing resources from third-party services.

How to concretely use Cloud/Edge resources efficiently within a communication network framework represents the next challenge. Indeed, little has been said about how to deploy, access and manage such resources in operational scenarios where the complexity of network, storage and computer systems imposes many operational and functional constraints. In addition, with the advent of machine learning, the need to integrate resource management interfaces and automatic intelligent decision processes has become of paramount importance and promises to offer novel classes of solutions to the resource efficiency problem. A major problem of applying machine learning to Cloud/Edge networked environments is that making and enforcing decisions automatically needs to be tracked and logs need to be secured. The rationale is twofold: (i) provisioning online services needs to be monitored towards the need of maintaining the service level high, according to service-level agreements, and (ii) the legal and economic responsibility of users, network and service operators needs to be determined in case of disputes, which might be cumbersome in case of applying machine learning decision making algorithms. Therefore, proposing the DiscoLedger project, we have identified the need of making the use of Cloud/Edge resources traceable and auditable in a framework in which the different players are not necessarily trustable. The deployment of distributed ledger technologies will be therefore explored. In particular, we consider that lightweight ledgers need to be provided in the Cloud/Edge as virtual network functions, which we call microledgers.

Specifically, in DiscoLedger, we will tackle both efficiency and traceability/auditability of services in the Cloud/Edge in a cellular network context, with network slicing features. To do so, we will evolve the results of DiscoEdge and build a proof of concept on (a) scaling/migrating online services in the Cloud/Edge by means of self-tuning and possibly interpretable machine learning algorithms and (b) embedding distributed ledgers technologies in the architecture, in the form of microledgers, with the associated support for their virtualization and management through intelligent algorithms.

Project PDC2021-121836-I00 funded by MCIN/AEI /10.13039/501100011033 and the European Union through the Next GenerationEU/ PRTR program.