SSLB: Self-Similarity-Based Load Balancing for Large-Scale Fog Computing

As a novelty approach to achieve Internet of things and an important supplement of cloud, fog computing has been widely studied in recent years. The research in this domain is still in infancy, so an efficient resource management is seriously required. Existing solutions are mostly ported from cloud domain straightforward, which performed well in many cases, but cannot keep excellent when the fog scale increased. In this paper, we examine the runtime characteristics of fog infrastructure and propose SSLB, a self-similarity-based load balancing mechanism for large-scale fog computing. As far as we know, this is the first work try to address the load balancing challenges caused by fog’s ‘large-scale’ characteristic. Furthermore, we propose an adaptive threshold policy and corresponding scheduling algorithm, which successfully guarantees the efficiency of SSLB. Experimental results show that SSLB outperforms existing schemes in fog scenario. Specifically, the resources utilization of SSLB is 1.7$$\times $$× and 1.2$$\times $$× of traditional centralized and decentralized schemes under 1000 nodes.