ISP service Tracking via 8.8.8.8 packet drops

[kumarmh91282]

Hi All,

Have a query regarding IP SLA tracking to track Internet services.

We have Two ISPs and have configured IP Sla tracking [Tracking 8.8.8.8 and 1.1.1.1] for Iink failover.

Lately have noticed 8.8.8.8 and 1.1.1.1 are giving Packet drops even though ISP link is stable very frequently and causing issues.

tracking ISP DNS also not feasible in my case, we are using our own APNIC IP pool.

Requirement: What are the other best practices to track INTERNET  ?  please could you share your thoughts and best practices.

[balaji.bandi]

Depends on your network, how far you are for the Google DNS, yes you are right, even though if your ISP stable reaching google DNS some time much diffcult that is beyond ISP control once the packet leave ISP network.

Try 8.8.4.4 also 1.0.0.1

[kumarmh91282]

Thanks for the reply, Surely will try tracking 1.0.0.1 for few days and check the performance.

But just wanted to what is the industry best practices to track Internet ( Best IP to track in IP SLA)

[Georg Pauwen]

Hello,

according to Google itself (see link below), using their public DNS is not the best idea when it comes to performance measurement (and hence IP SLAs).

Then again, even if you use another IP address (such as 198.41.0.4, which is the IPv4 root name server), you still might run into the problem of (ISP induced) ICMP rate limiting.

What does the configuration of your IP SLA look like ? How often are you pinging ?

https://peering.google.com/#/learn-more/faq

[kumarmh91282]

Thanks for the reply sir, I'm using default IP SLA config on Cisco ISR 4451 router, i agree with your statement,

what is the ideal time value to track a IP before considering for IP Failover.

[Georg Pauwen]

Hello,

the default values for an IP SLA are:

Threshold: 5000 milliseconds.
Timeout: 5000 milliseconds.
Frequency: 60 seconds.

The command 'show running-config all | begin ip sla' should also give you the values that are configured and active.

60 seconds is actually rather long, because the SLA is considered down after 3 consecutive failures.
You might want to set this value to 5 (frequency 5).

[MHM Cisco World]

many reason lead to drop packet.