Performance Routing with NAT

When Cisco IOS PfR and NAT functionality are configured on the same router and PfR controls the routing for a traffic class using static routing, some applications may fail to operate due to dropped packets. This dropping of packets behavior is seen when static routing is used to connect to multiple ISPs from the same router, PfR uses static routing to control the traffic class routing, and one or more of the ISPs use Unicast Reverse Path Forwarding (Unicast RPF) filtering for security reasons. Packets are dropped at the ingress router performing Unicast RPF because PfR changes the route for an outgoing packet for a traffic class from one exit interface to another after the NAT translation from a private IP address to a public IP address is performed. When the packet is transmitted, Unicast RPF filtering at the ingress router (for example, an ISP router) will show a different source IP address from the source IP address pool assigned by NAT, and the packet is dropped. For example, the figure below shows how PfR works with NAT.

Figure 1

PfR with NAT

The NAT translation occurs at the router that is connected to the internal network, and this router can be a border router or a combined master controller and border router. If PfR changes routes to optimize traffic class performance and to perform load balancing, traffic from the border router in the figure above that was routed through the interface to ISP1 may be rerouted through the interface to ISP2 after the traffic performance is measured and policy thresholds are applied. The RPF check occurs at the ISP routers and any packets that are now routed through ISP2 will fail the RPF check at the ingress router for ISP2 because the IP address of the source interface has changed.

The solution involves a minimal configuration change with a new keyword, oer, that has been added to theip nat inside source command. When theoer keyword is configured, new NAT translations are given the source IP address of the interface that PfR has selected for the packet and PfR forces existing flows to be routed through the interface for which the NAT translation was created. For example, PfR is configured to manage traffic on a border router with two interfaces, InterfaceA to ISP1 and InterfaceB to ISP2 in the figure above. PfR is first configured to control a traffic class representing Web traffic and the NAT translation for this traffic already exists with the source IP address in the packets set to InterfaceA. PfR measures the traffic performance and determines that InterfaceB is currently the best exit for traffic flows, but PfR does not change the existing flow. When PfR is then configured to learn and measure a traffic class representing e-mail traffic, and the e-mail traffic starts, the NAT translation is done for InterfaceB. The PfR static routing NAT solution is a single box solution and configurations with interfaces on multiple routers using NAT and managed by PfR are not supported. Network configurations using NAT and devices such as PIX firewalls that do not run Cisco IOS software are not supported.

NAT enables private IP internetworks that use nonregistered IP addresses to connect to the Internet. NAT operates on a router, usually connecting two networks together, and translates the private (not globally unique) address in the internal network into legal addresses before packets are forwarded onto another network. NAT can be configured to advertise only one address for the entire network to the outside world. This ability provides additional security, effectively hiding the entire internal network behind that one address.

NAT is also used at the Enterprise edge to allow internal users access to the Internet and to allow Internet access to internal devices such as mail servers.

For more details about NAT, see the "Configuring NAT for IP Address Conservation" chapter of the Cisco IOS IP Addressing Services Configuration Guide.

You can conserve addresses in the inside global address pool by allowing the router to use one global address for many local addresses. When this overloading is configured, the router maintains enough information from higher-level protocols (for example, TCP or UDP port numbers) to translate the global address back to the correct local address. When multiple local addresses map to one global address, the TCP or UDP port numbers of each inside host distinguish between the local addresses.

Perform this task to allow PfR to control traffic with static routing in a network using NAT. This task allows PfR to optimize traffic classes while permitting your internal users access to the internet.

When Cisco IOS PfR and NAT functionality are configured on the same router and PfR controls the routing for a traffic class using static routing, some applications may fail to operate due to dropped packets. This dropping of packets behavior is seen when static routing is used to connect to multiple ISPs from the same router, PfR uses static routing to control the traffic class routing, and one or more of the ISPs use Unicast Reverse Path Forwarding (Unicast RPF) filtering for security reasons.

In this task, the pfr keyword is used with theip nat inside source command. When thepfr keyword is configured, new NAT translations are given the source IP address of the interface that PfR has selected for the packet and PfR forces existing flows to be routed through the interface where the NAT translation was created. This task uses a single IP address but an IP address pool can also be configured. For a configuration example using an IP address pool, see "Configuring PfR to Control Traffic with Static Routing in Networks Using NAT" section.

Note	The PfR static routing NAT solution is a single box solution and configurations with interfaces on multiple routers using NAT and managed by PfR are not supported.

For more details about configuring NAT, see the "Configuring NAT for IP Address Conservation" chapter of the Cisco IOS IP Addressing Services Configuration Guide.

1.    enable

2.    configure terminal

3.    access-list access-list-number {permit | deny} ip-addressmask

4.    route-map map-tag [permit | deny] [sequence-number]

5.    match ip address {access-list access-list-name| prefix-list prefix-list-name}

6.    match interface interface-type interface-number [...interface-type interface-number]

7.    exit

8.    Repeat Step 4 through Step 7 for more route map configurations, as required.

9.    ip nat inside source {list {access-list-number| access-list-name} | route-map map-name} {interface type number| pool name} [mapping-id map-id | overload| reversible| vrf vrf-name][pfr]

10.    interface type number

11.    ip address ip-address mask

12.    ip nat inside

13.    exit

14.    interface type number

15.    ip address ip-address mask

16.    ip nat outside

17.    end