In multi-stakeholder data ecosystems, digital resources (e.g. shared datasets) are often co-owned by independent organizations, making access governance a collaborative challenge. Each stakeholder brings its own security and business constraints, so agreeing on concrete access rules across organizational boundaries is hard. For example, in an industrial IoT supply chain, machine data can be valuable to the factory operator, the equipment supplier, and a maintenance provider, each must consent to who can access what. Manual agreement is slow and contentious, motivating an automated negotiation mechanism to help stakeholders efficiently converge on a shared set of acceptable rules. Consensus is only half the problem, once access rules are agreed upon, they must be enforced reliably in a distributed system with no single authority to push updates. Nodes go offline, networks partition, and rule updates arrive out of order. Without careful design, different parts of the system will enforce different decisions. Ensuring that all parties can consistently uphold the agreed rules under stated assumptions requires robust, fault-tolerant mechanisms. This licentiate thesis tackles both agreement and enforcement, providing methods to reach common ground on access rules and to apply them securely despite distributed-systems challenges. The thesis makes three contributions. First, it introduces a utility-based negotiation method that lets multiple stakeholders collaboratively arrive at a common access-rule set by quantifying preferences and using optimization to automate consensus, supporting more structured and reproducible agreement compared to informal negotiation. Second, it develops a formal verification toolchain for cloud-native infrastructures (shown on Kubernetes) to prevent misconfiguration and privilege escalation; RBAC and admission rules are translated into logical constraints, an SMT solver checks for unsafe conditions, and only verified rules are deployed; an integrated deny-overrides enforcement path then applies them consistently at runtime. Third, it outlines EQuack (to be detailed in forthcoming work), an offline-capable access control model for distributed ecosystems where continuous connectivity cannot be assumed. It ensures that even if nodes diverge while offline, they eventually converge to the same authorized state via deterministic deny-wins replay over update logs and tamper-evident audit trails, without relying on blockchains or other heavy consensus mechanisms. Taken together, the results suggest a path toward more secure collaboration by supporting structured agreement over access rules and providing enforcement mechanisms that can remain consistent in distributed settings, within the limits of the evaluated scenarios and stated assumptions.
Aditya Sissodiya (Thu,) studied this question.