Problem
Research questions and friction points this paper is trying to address.
serverless computing
data replication
function scheduling
edge-cloud continuum
consistency models
Data Replication Meets Function Scheduling in the Edge-Cloud Continuum
🤖 AI Summary
In the edge-cloud continuum, remote state access by serverless functions undermines the low-latency benefits of edge computing, as data placement is tightly coupled with function scheduling and consistency models, posing a challenging joint optimization problem. This work presents the first unified integer programming formulation that jointly models function scheduling and data replication. It introduces multi-granularity heuristics: a global greedy heuristic achieves near-optimal solution quality—within a few percentage points of optimality—even at scales of tens of thousands of nodes, while a local aggregate view method dynamically adapts to client demands, significantly reducing and stabilizing end-to-end latency in mobile scenarios. Experimental results demonstrate that data placement exerts a substantially greater impact on performance than function scheduling alone.
📝 Abstract
Serverless computing is an appealing model for the edge-cloud continuum, but its stateless assumption breaks down once functions need persistent data: fetching state from a distant cloud store erases the latency benefit of running at the edge. Keeping data close means replicating it, and replication forces a placement decision that is coupled with where functions execute and with the consistency each application demands. We study this joint problem of function scheduling and data placement under two consistency models, strong and eventual replication. We first formulate it as a Binary Linear Program that yields the optimal placement for a given system snapshot, and use it as a reference point. Because the solver does not scale past a few hundred nodes, we add two heuristics with progressively less information: a Global-View greedy method that works from the same complete snapshot, and an Aggregated-View heuristic in which each node decides from locally observed demand alone. Across a range of system sizes the Global-View heuristic stays within a few percent of the optimum while scaling to over $10^4$ nodes. The Aggregated-View heuristic sacrifices some solution quality, but adapts continuously to each invocation. Under client mobility, centralized policies suffer from stale snapshots and recurring latency spikes, while the Aggregated-View maintains low and stable client-observed latency. Across all experiments, data placement proves more influential than function scheduling in determining the outcome.
Innovation
Methods, ideas, or system contributions that make the work stand out.
function scheduling
data replication
edge-cloud continuum
consistency models
heuristic algorithms
🔎 Similar Papers
No similar papers found.
