FastScale: Fast Scaling Out of Network Functions

Scaling out network functions (NFs) offers great flexibility and elasticity to the management of service function chains (SFCs). However, existing solutions overlook the fact that scaling out an NF may influence downstream NFs. Their plans need several steps to scale out overloaded NFs, resulting in unpredictable SFC performance and slow convergence. In this paper, we propose FastScale, a novel framework that makes chain-wide scale-out plan within a single step. Once an overloaded NF is detected, it models the problem of scaling out NFs in consideration of both overloaded NFs and downstream NFs. It determines which NFs to scale and how many NF instances should be used by each NF within a single step so as to avoid unnecessary reconfiguration and reduce convergence time. After making scale-out plan, FastScale automatically updates the thresholds used to detect overloaded NFs to reduce user burdens. We have built a prototype of FastScale. Our evaluation demonstrates that FastScale reduces the convergence time of SFCs by up to 68% in comparison with existing solutions.