RSMLT vs VRRP

[jossaq]

What is the best redundacy routing protection RSML or VRRP?

[Dominik]

Hi jossaq

both has their pros and cons:

VRRP:
-pro interop with other vendors
-pro open standard
-pro active active loadsharing with backupmaster feature
-con limit of 256 VRRP instances
-con more complex to configure
-con constant network traffic for the VRRP hello packets

RSMLT
-pro easy to configure
-pro no interface limit
-pro no network management traffic
-con propritry

I prefer the RSMLT for Switchcluster network designs, worked very stable in my expirience.

Cheers

[jossaq]

WAW.

Tks a lot.

For clustering (Full meshed SMLT) what values for RSMLT Holup an Hold Down timers do yo use?

[Dominik]

I usual use a holddowntimer of 60sec, wich is the default timer.
For holduptimer I use the infinite timer of 9999. If one memeber of the switchcluster goes down,
usually you want to keep the still working unit to make the job with no time limit.

Donīt forget to safe the config when you first time enable the rsmlt feature, so that the peer-addr of the other switchcluster unit is written in your config.

Good Luck

[Jon Hurtt]

I usual use a holddowntimer of 60sec, wich is the default timer.
For holduptimer I use the infinite timer of 9999. If one memeber of the switchcluster goes down,
usually you want to keep the still working unit to make the job with no time limit.

Donīt forget to safe the config when you first time enable the rsmlt feature, so that the peer-addr of the other switchcluster unit is written in your config.

Good Luck

Please refer to the Super Large Campus Technical Configuration Guide for recommended RSMLT hold-up and hold-down timers. Make sure to only use infinite timer for RSMLT Edge (Client Gateway).  Not sure if this was a RSMLT Core or RSMLT Edge deployment.  Good Luck.

Link: Super Large Campus Technical Configuration Guide
http://support.avaya.com/css/P8/documents/100124746

[Dominik]

THX @ John

You are absolutly right I was regarding to the triangle or RSMLT edge design with only one switchcluster.

If you have two Switchclusters (4 switches) Avaya recommands this settings:

The hold-up timer defines how long the RSMLT-peer switch keeps routing for its peer after a peer switch failure. Leave the hold-up timer at default of 180 seconds when OSPF is used.
The hold-down timer defines how long the switch waits before activating Layer 3 forwarding for its peers‘ MAC address. Leave the hold-downup timer at default of 60 seconds when OSPF is used. If RIP is used, the value should be changed to 180 seconds

This settings in the Super Large Campus solution guide are related to the VSP9000.


You can find a similar information for the ERS8600 in the Large Campus Solution Guide (page 52)

http://support.avaya.com/css/P8/documents/100123838


As additional information the arguments for RSMLT from this guide:

-RSMLT is only limited by the number of IP interfaces
-VRRP is limited to 250 instances
-RSMLT requires significantly less control traffic
-RSMLT is much less intensive on CPU resources


Cheers