SD-Access single node vs three nodes

[Hashem1323]

Is there any difference between a single-node DNAC deployment and a three-node DNAC cluster, apart from high availability (HA), which I understand requires a three-node cluster?

What other differences or additional capabilities should I be aware of between the two deployment models?

Will it affect any feature for the SD-Access fabric? or SD-Access will work using either the single or three nodes deployment?

Thanks

[Torbjørn]

Hello @Hashem1323, 

They offer the same functionality regardless of clustering. The XL node size the clustering will provide you with a higher supported node count. For the other sizes it remains the same for both single node and clustered configurations.

I would however recommend going with a 3 node cluster if you decide to deploy SDA. Having the Cat-C available for network operations is essential when managing fabrics.

[Andrii Oliinyk]

Adding @Torbjørn , u'd like to have redundant connectivity with services CatC integrates with (ISE, Space, etc). Hence cluster is only viable solution for the prod.

[Sebastien CIELOCH]

Hi @Hashem1323 

No difference:) 

It will be cheaper with 1 node of course, specially when RAM and storage costs are insanes. Considere 1+1 as well to reduce appliances costs.

But you will definitely need CatC to run your fabric... It will run without it, but you won't be able to make any changes. Which is a real problem.

But to be honnest, you have a higher chance of a software defects from CatC/ise integration after an upgrade versus a hardware failure 😄