Auto-Scaling VNFs Using Machine Learning to Improve QoS and Reduce Cost

Abstract

Virtualization of network functions (as virtual routers, virtual firewalls, etc.) enables network owners to efficiently respond to the increasing dynamicity of network services. Virtual Network Functions (VNFs) are easy to deploy, update, monitor, and manage. The number of VNF instances, similar to generic computing resources in cloud, can be easily scaled based on load. Hence, auto-scaling (of resources without human intervention) has been receiving attention. Prior studies on autoscaling use measured network traffic load to dynamically react to traffic changes. In this study, we propose a proactive Machine Learning (ML) based approach to perform auto-scaling of VNFs in response to dynamic traffic changes. Our proposed ML classifier learns from past VNF scaling decisions and seasonal/spatial behavior of network traffic load to generate scaling decisions ahead of time. Compared to existing approaches for ML-based auto-scaling, our study explores how the properties (e.g., startup time) of underlying virtualization technology impacts Quality of Service (QoS) and cost savings. We consider four different virtualization technologies: Xen and KVM, based on hypervisor virtualization, and Docker and LXC, based on container virtualization. Our results show promising accuracy of the ML classifier using real data collected from a private ISP. We report indepth analysis of the learning process (learning-curve analysis), feature ranking (feature selection, Principal Component Analysis (PCA), etc.), impact of different sets of features, training time, and testing time. Our results show how the proposed methods improve QoS and reduce operational cost for network owners. We also demonstrate a practical use-case example (Software-Defined Wide Area Network (SD-WAN) with VNFs and backbone network) to show that our ML methods save significant cost for network service leasers. 1

Virtualized Controller Placement for Multi-domain Optical Transport Networks

“…But, if the controller capacity constraint fails (line 9), the algorithm checks if the current hosting location (h) has enough compute and memory capacity to host the additional δ controllers (line 13). If yes, we turn on additional controllers and load balance the switches and traffic flows (line [14][15][16][17]. If host location h does not have enough resources, the algorithm finds the next optimal location to host all the instances (line 20) following constraints as in Eqns.…”

“…(1)(2)(3)(4)(5) and minimizing Eqn. (14). In this step, we utilize the benefits of consolidation in computing.…”

“…Each network domain requires domain-specific controller(s), that are connected to other domains via backbone optical links. For example, we consider six fixed-grid domains (4,5,7,9,12), two 'packet network' domains (8 and 11), and six flex-grid domains (1,2,3,10,13,14). Later, our study also considers larger domains in Fig.…”

Recent Advancement of Auto-Scaling in LTE M2M Communication.

“…More suitable for predictable traffics pattern. Hence the author Sabidur Rahman et al [6] proposed Machine Learning (ML) based classifier which would proactively take scaling decisions ahead of time, instead of traffic predictions.…”

“…Sabidur Rahman et al [6] Proposed Machine Learning (ML) based classifier which would take proactively scaling decisions…”