Dynamic VPN Optimization by ALTO Guidance (DRAFT TITLE)

Hybrid cloud computing systems interconnect users and providers of resources by Virtual Private Networks (VPN). Due to elasticity, these VPNs are more dynamic than the traditional VPN services offered by network service providers. New usage patterns such as cloud bursting requires to dynamically scale out VPNs and to possibly to decide among several candidate destinations for a VPN. Managing and optimizing the topology of VPNs requires a insight into the underlying network topology. The Application-Layer Traffic Optimization (ALTO) protocol is a standardized solution that enables export of topology information to applications. This paper motivates using ALTO in a Virtual Private Network (VPN) environment. We explain how the topology data can be gathered and how it can be exposed to software applications that dynamically control VPN topology. A detailed analysis shows that application-driven management of VPN is a new, promising use case of ALTO. However, there are also challenges still to be solve. We also present a prototype implementation for this ALTO VPN service that uses a carrier-grade network management system.