You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cloudstack.apache.org by Chandan Purushothama <Ch...@citrix.com> on 2013/02/28 03:30:18 UTC

Question pertaining to Use of same public IP for static NAT, PF and LB in VPC

I referred to the Feature Information present at https://issues.apache.org/jira/browse/CLOUDSTACK-756

I faintly remember that we intentionally placed a requirement for a Public IP to be used for only one service in a VPC in the earlier release. One reason might be due to technical challenges that might arise due to different service provides present in a VPC (For Ex: Load Balancing can be done by an external device while other services are taken care of by the VPC VR.) .

Can anyone help me understand the possible technical limitations of this feature when VR is the all-services provider of the VPC? I am trying to figure out the cases that needs to be tested for this feature.

Thank you,
Chandan.

Re: Question pertaining to Use of same public IP for static NAT, PF and LB in VPC

Posted by Chiradeep Vittal <Ch...@citrix.com>.
1. You cannot upgrade to an offering which uses multiple devices to offer
the same services
2. There are port-level conflicts that need to be tested: e.g., LB on port
80 along with PF on the same port, static NAT + PF etc


On 2/27/13 6:30 PM, "Chandan Purushothama"
<Ch...@citrix.com> wrote:

>I referred to the Feature Information present at
>https://issues.apache.org/jira/browse/CLOUDSTACK-756
>
>I faintly remember that we intentionally placed a requirement for a
>Public IP to be used for only one service in a VPC in the earlier
>release. One reason might be due to technical challenges that might arise
>due to different service provides present in a VPC (For Ex: Load
>Balancing can be done by an external device while other services are
>taken care of by the VPC VR.) .
>
>Can anyone help me understand the possible technical limitations of this
>feature when VR is the all-services provider of the VPC? I am trying to
>figure out the cases that needs to be tested for this feature.
>
>Thank you,
>Chandan.