Dynamic Scaling of Virtualized, Distributed Service Chains: A Case Study of IMS

Abstract

The emerging paradigm of network function virtualization advocates deploying virtualized network functions (VNF) on standard virtualization platforms for significant cost reduction and management flexibility. There have been system designs for managing dynamic deployment and scaling of VNF service chains within one cloud data center. Many real-world network services involve geodistributed service chains, with prominent examples of mobile core networks and IMSs (IP Multimedia Subsystems). Virtualizing these service chains requires efficient coordination of VNF deployment across different geodistributed data centers over time, calling for new management system design. This paper designs a dynamic scaling system for geo-distributed VNF service chains, using the case of an IMS. IMSs are widely used subsystems for delivering multimedia services among mobile users in a 3G/4G network, whose virtualization has been broadly advocated in the industry for reducing cost, improving network usage efficiency and enabling dynamic network topology reconfiguration for performance optimization. Our scaling system design caters to key controlplane and data-plane service chains in an IMS, combining proactive and reactive approaches for timely, costeffective scaling of the service chains. We evaluate our system design using real-world experiments on both emulated platforms and geo-distributed clouds.

Reconfiguration Aware Orchestration for Network Function Virtualization With Time-Varied Workload in Virtualized Datacenters

“…The reconfiguration mechanism can be either reactive or proactive in nature. A reactive mechanism [11], [12], [15], [21] is used upon the initiation of an event. For instance, [15] reconfigures VNF placements to turn off idle PMs while [21] replaces the VNFs upon arrival of new requests by addressing the trade off between bandwidth consumption and PM's resource consumption.…”

“…For instance, [15] reconfigures VNF placements to turn off idle PMs while [21] replaces the VNFs upon arrival of new requests by addressing the trade off between bandwidth consumption and PM's resource consumption. On the other hand a proactive mechanism [12], [13], [20] uses forecasting techniques to predict the workload. [20] have predicted the workload and have used a k-medoids clustering technique for VNF reconfiguration while [13] have used a forecasting technique to predict the workload demands for VNF reconfiguration so that the number of VNF placement changes can be reduced.…”

“…[20] have predicted the workload and have used a k-medoids clustering technique for VNF reconfiguration while [13] have used a forecasting technique to predict the workload demands for VNF reconfiguration so that the number of VNF placement changes can be reduced. On the other hand [12] have used a combination of both reactive and proactive scaling technique. In proactive mechanism, they have predicted the workload for next time interval and a real time scaling is done based on a reactive event.…”

“…Since SFCs may have strict latency requirements, there can be significant impact on SLA too. Hence the orchestrator can choose reconfiguration solutions [11], [12], [12], [13] in order to address this situation, including (i) scale up, (ii) scale out, and (iii) migration.…”

Statistical Assessment of IP Multimedia Subsystem in a Softwarized Environment: A Queueing Networks Approach

“…Over the recent years, academia and industry alike have devoted an increasing interest to the characterization of 5G network architectures and their constitutive elements, with analyses ranging from optimal resource distribution of virtualized multimedia nodes [29] to availability characterization of virtualized IMS deployments [30]. Yet, research and practical developments in this area are incredibly fast-paced, and it would take a dedicated review paper to provide a comprehensive snapshot.…”