Replica from Multiple Sites to Veeam Cloud Connect

[jsalinas]

Hi,

I'm a Cloud Connect Service Provider. I have a Potential Customer / Tenant who has 4 physical datacenters located in different cities.

In each site, the customer has a few ESXi hosts servers and some production VMs. Each site has its own network segments, which are different from the other sites (No network overlapping exists between sites). All of the sites are connected through WAN links. All of the VMs have network connectivity regardless of the site they are running on.

They are interested on hiring my Veeam Cloud Connect DRaaS service.

The Replicas will need internet access while running on my Datacenter. Initially the customer will be replicating VMs from the 4 sites connected to 8 different network segments.

The Customer has required that in case of a Failover event (partial or full site), all of the VMs network connectivity remains fully operational.

Could you please let me know if Veeam Cloud Connect could support a DRaaS solution for this scenario?

Thank you!

[nielsengelen]

This is possible, you can create accounts per site specific to achieve the customer requirements.

[jsalinas]

Hi Vmniels,

I'm pretty sure that setup would do the trick for any Partial Site Failover scenario that I can think of .

Just wondering, would you have any recommendation on how to maintain full network connectivity in the case that all of the 4 sites do a Full Site Failover? In this scenario all of my customer's VMs will be running on my datacenter but each site will be independent from each other; so, the only option I can think of is telling the customer that He will need to have replicated some VPN servers which allow him to have full site VPNs between all of his sites so they can have full network connectivity.

By any chance, do you know if VeeamVPN HUB would help me to ease this situation? Can it be installed on any Cloud Connect infrastructure? I'm using my own VCC infrastructure, not azure.

Thank you!

Jorge

[nielsengelen]

It may be better to ask how they keep the connectivity now to each site. If they do a full failover this will be normally the same behaviour. You can ask them to utilize VPN software but this would mean you (in combination with the customer) would have to deploy VM's in the resource group (per site) which talk to eachother.

You can use VeeamPN for it however I would first look at the current networking state at the customer (site 1 to 4) before jumping on the VPN story.

[dellock6]

We did this exact design for one service provider that had to replicate 10 different remote locations of a customer. Let's start by ignoring partial failover, that's easy since the partially failed site will just connect back to its production site, and via existing networking solutions to the others.
The problem, as stated here, is full failover. In our specific case, we basically came to the conclusion that instead of managing 10 sites and their failovers, we would have been better to manage 20 sites, with 10 of them usually offline. There is obviously network overlap between any production site (and its networks and rules) and its failvoer sibling, I don't have the details of the specific appliance configuration that was used at this provider other than it was a Fortinet solution, but maybe this can give you an idea about how to attack the problem.

Luca

[jsalinas]

It may be better to ask how they keep the connectivity now to each site. If they do a full failover this will be normally the same behaviour. You can ask them to utilize VPN software but this would mean you (in combination with the customer) would have to deploy VM's in the resource group (per site) which talk to eachother.

You can use VeeamPN for it however I would first look at the current networking state at the customer (site 1 to 4) before jumping on the VPN story.

Sure thing, I'll follow your advice and I'll go and dig a little deeper into the network details. Thank you for taking the time to help me out!

[jsalinas]

We did this exact design for one service provider that had to replicate 10 different remote locations of a customer. Let's start by ignoring partial failover, that's easy since the partially failed site will just connect back to its production site, and via existing networking solutions to the others.
The problem, as stated here, is full failover. In our specific case, we basically came to the conclusion that instead of managing 10 sites and their failovers, we would have been better to manage 20 sites, with 10 of them usually offline. There is obviously network overlap between any production site (and its networks and rules) and its failvoer sibling, I don't have the details of the specific appliance configuration that was used at this provider other than it was a Fortinet solution, but maybe this can give you an idea about how to attack the problem.

Luca

Thank you Luca! I understand what you propose and I can see the overlap you are referring to. I'm going to talk to my Networking Team to see what they recommend.