Virtual Routing and ForwardingRead post
The term virtualization can have many different meanings depending on its context. In enterprise data centers, server virtualization is well established and has delivered improved utilization of resources. In networks, existing technologies such as virtual routing and forwarding (VRF) instances and VLANs can be viewed as forms of virtualization. In IP-based computer networks, virtual routing and forwarding (VRF) is a technology that allows multiple instances of a routing table to co-exist within the same router at the same time.
As a result an edge router can have multiple instances of routers inside. Because the routing instances are independent, they can use overlapping subnets, allowing to reuse same address space many times on one physical device. Those VRF provides a reliable mechanism for trusted VPNs to be built over a shared infrastructure.
Each virtual router has all of the functionality that is available from an actual physical router – route tables, IP address space, and routing protocol processes. In our implementation client is able to configure two separate routers and build two VPNs in his existing network. We used machine learning for optimisation of virtual paths, as it can detect seasonality, understand what a normal behaviour is, detect anomalies and get informations about their roots. It predicts when there will be peak traffic (for example on Friday morning) and dynamically anticipate congestion, take proactive measure to divert traffic onto an alternate path.
Typical use cases for network providers are:
Provider can sell separate VPN to his customers. From a configuration perspective, the provider no longer needs to worry about overlapping IP addresses among customers on the same physical router. Event management and troubleshooting features such as ping and traceroute are still available for each virtual router making troubleshooting easier.
Provider can group backbone logical networks into separate VPN on one physical router to handle different types of traffic separately – for instance separate payload from operation and maintenance data flow. Each logical network has a set of requirements with respect to connectivity, QoS, network availability, etc.
Monday, 06 March 2017