cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
992
Views
4
Helpful
15
Replies

traffic redirection when there's overload

harinirina
Level 1
Level 1

Hi,

We need to redirect internet traffic to ISP1 when ISP2 is down or overloaded.

When all work fine, LAN1 will go to ISP1 and LAN2 to ISP2

We need to allocate bandwidth for LAN1 when both LAN1 and LAN2 go one ISP.

How can we do that, how to detect that bandwidth is overloaded and to allocate bandwidth.

15 Replies 15

owaisberg
Level 1
Level 1

To cover a condition when link to one ISP goes

down and you want to redirect traffic

destines to that very ISP to ISP2 you can

just run BGP to both ISPs. As soon one link

is lost your exit point will shift to the second ISP.

Now, to be capable of making your redirection

decision based on the link load you can try

to use object tracking with IP SLA where you

can set desirable response time from the

target and change the exit point if it is not

within your requirements.

HTH,

OW

Hi,

and what about bandwidth limitation? should we configure Qos, and on which router , or there is other method for doing this ?

Hi

You can configure bandwidth allocation along these lines

LAN1 = 192.168.1.0/24

LAN2 = 192.168.2.0/24

LAN1 gets 30%

LAN2 gets 40%

class-map match-any lan1

match access-group 101

class-map match-any lan2

match access-group 102

policy-map restrict

class lan1

bandwidth percent 30

class lan2

bandwidth percent 40

interface s0

service-policy output restrict

access-list 101 permit ip 192.168.1.0 255.255.255.0 any

access-list 102 permit ip 192.168.2.0 255.255.255.0 any

Apply this to the router that connects to your ISP. Make sure that your interface accurately reflects the actual bandwidth available as this is what QOS uses.

HTH

Jon

Hi,

it will not be applyed all time, it will be just applyed backup is functionning.

is there a way to do this

hi my friend,

well u apply the QOS and let it monitor ur network. coz QOS policies are not applied till the time there is congestion on the interface. if enough bandwidth available it will let all data go. bt when there is a congestion it will follow the policy.

all my seniors here,correct me if i am wrong.

regards

Hi,

We won't apply qos when backup is not actif.

We'll apply it on traffic to isp1 when isp2 is down or saturated and lan2 should be redirected to isp1.

and when isp1 is down or saturated , we'll apply bandwidth limitation on traffic to isp2 and lan1 should be redirected to isp2.

From this and from your initial description of the problem it looks like simply configuring routing once and forever won't give you exactly what you expect - while routing (BGP) will detect link failures it doesn't care about congestions; on the other hand QoS care about congestions but doesn't care about routing. Have you considered (more complex) Optimized Edge Routing? It's an add-on in IOS that makes routing and IP SLA monitor to work together and dynamically adjust routing depending on network performance (rather than only availability).

Check following:

http://www.cisco.com/en/US/products/ps6628/products_ios_protocol_option_home.html

http://www.cisco.com/en/US/products/ps6599/products_white_paper09186a008036524d.shtml

Especially this looks like your scenario:

http://www.cisco.com/en/US/products/ps6599/products_white_paper09186a008036524d.shtml#wp2180497

Cheers,

iLya

Hi OW,

Could you give more info (config or interesting link) on setting desirable response time with IP SLA ?

Hi,

"type echo protocol ipIcmpEcho" will be used to test the reachability of ISP, won't it?

What's the command to be used for testing "link load"? how can we use it with track?

Now, if you look at the IP SLA configuration

you can see "timeout" value can be set which

indicates the amount of time to wait for a

response from its request packet. If that

response is higher than the value specified

then reachability ICMP test should fail and

trigger the client object (in our case it is

a static route). Idea I had is that during

heavy load you will have your RTT times higher

so IP SLA can catch this condition.

I would recommend for you to measure RTT

times over the backup link during normal

load and during overload conditions and then

use these values to configure IP SLA tracking.

HTH,

OW

Hi OW,

We'd like to know the response time during congestion before configuring timeout.

We've thougth of using path-jitter or icmp-jitter for this.We'd like to send the result of "ip sla statistics" to a syslog server.

We tried using "ip sla monitor reaction-configuration" with " ip sla monitor logging traps".

We got many log from router but none concerning "ip sla".

Is there any way to follow ip sla statistics?

To get response times during congestions

just load your link with different traffic valumes and measure RTT in each scenario,

then adjust your IP SLA accordingly.

As of logging make sure your logging on that

router is configured properly in the first

place.

HTH,

OW

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Innovations in Cisco Full Stack Observability - A new webinar from Cisco