OmniSwitch os6900, Руководство по настройке сети

Страница: 380 / 942

Multiple VRF Overview Configuring Multiple VRF
page 16-6 OmniSwitch AOS Release 7 Network Configuration Guide June 2013
Multiple VRF Overview
The Multiple Virtual Routing and Forwarding (VRF) feature provides the ability to configure separate
routing instances on the same switch. Similar to using VLANs to segment Layer 2 traffic, VRF instances
are used to segment Layer 3 traffic.
Some of the benefits of using the Multiple VRF feature include the following:
•
Multiple routing instances within the same physical switch. Each VRF instance is associated with a set
of IP interfaces and creates and maintains independent routing tables. Traffic between IP interfaces is
only routed and forwarded to those interfaces that belong to the same VRF instance.
•
Multiple instances of IP routing protocols, such as static, RIP, IPv4, BGPv4, and OSPFv2 on the same
physical switch. An instance of each type of protocol operates within its own VRF instance.
•
The ability to use duplicate IP addresses across VRF instances. Each VRF instance maintains its own
IP address space to avoid any conflict with the service provider network or other customer networks.
•
Separate IP routing domains for customer networks. VRF instances configured on the Provider Edge
(PE) are used to isolate and carry customer traffic through the shared provider network.
This implementation of VRF functionality does not require a BGP/MPLS configuration in the provider
network. Instead, VRF instances can route and forward IP traffic between customer sites using
point-to-point Layer 3 protocols, such as IP-IP or GRE tunneling.
The illustration on page 16-7 shows an example of how the Multiple VRF feature is used to provide
independent routing domains that isolate and carry customer traffic through the provider network. In this
example:
•
Each PE switch maintains more than one routing and forwarding table, in addition to the default VRF
instance table.
•
One VRF instance is configured on the PE switch for each customer network to which the PE is
connected.
•
Each interface on the PE that is connected to a customer edge (CE) switch is associated with the VRF
instance configured for that customer.
•
When an IP packet for Customer A is received on a PE 1 or PE 2 interface associated with VRF A, the
VRF A instance determines how to route the packet through the provider backbone so that it reaches
the intended Customer A destination.
•
When an IP packet for Customer B is received on a PE 1, PE 2, or PE 3 interface associated with VRF
B, the VRF B instance determines how to route the packet through the provider backbone so that it
reaches the intended Customer B destination.
•
When an IP packet for Customer C is received on a PE 1 or PE 3 interface associated with VRF C, the
VRF C instance determines how to route the packet through the provider backbone so that it reaches
the intended Customer C destination.