I want to verify my concepts regarding EoMPLS. Please Answer Yes if I am right, otherwise please correct me.
Point #1: The ATOM is P2P tunneling mechanism in which we would have MPLS L2 VPN. For example EoMPLS, HDLC over MPLS, FR over MPLS.
Point #2: It is used exclusively used for Ethernet. The main feature of VPLS is that is going to allow us to form MULTIPOINT Tunnel. L2TPv3 and AToM feature set is Point to Point. VPLS is Point to Multi-Point.
Point #3: We want to replicate databases from main site to DR-site, therefore, it is best option to use MPLS L2 VPN between three data centers. Two of them are Site 1 and Site 2 and one of them is the DR-Site of Site1 and Site2.
The MPLS L2 VPNs that I am using is EoMPLS
P2P link between Site 1 and DR-Site
P2P link between Site 2 and DR-Site
Please see three cases given in attachment VPN-1.jpg , VPN-2.jpg , VPN-3.jpg below and let me know if I am okey.
Point #4: This point is more important to me. As far as my knowledge is concerned, we can configure MPLS L2 VPNs in two ways.
1) We configure L2VPN to have Point to Point link as given in attached picture VPN-1.jpg. In this case, we assign IP address of same subnet to CE devices.
2) We configure L2VPN to stretch or expand L2 across data centers. Please see the attached picture Stretch.jpg In this case, we normally connect PE device to Customer Switches and configure trunk to pass VLANs information. Therefore, stretching L2 across multiple data centers.
I highly appreciate your response to my questions/concepts.
Have a read over this. I think it explains all the options quite well.
In particular, check out A-VPLS. I think it is a good fit for you based on what has been described. It is talked about in the above document as well.
Let me Thank you very much for the help you provide. Actually I need to propose a final solution to a client and I am a bit confused.
This is the requirement of the client.
1) Client wants same IP address schema in Primary and DR-Site. They cannot change the subnets as they will have to change all scripts for Applications.
2) Client wants to do partial fail-over. If a business process is dependent on 10 applications and 1 application goes down in Primary site, then, only this application should be run from DR-Site. This would be done manually.
It ultimately means that they want L2 extension across data centers.
3) Client wants to replicate the Applications and Databases to DR-Site.
I have attached diagram of the environment where they have Cisco 6513 in their Primary site and Huawei Cloud Engine in their DR-Site.
I know for the fact that we will need a link from MPLS cloud to Cisco and Huawei Switch to extend layer 2.
But Keeping above requirements in mind, what technology should I go for:
Question1: As far as my knowledge is concerned, all the aforementioned technologies are for L2 extension. But which technology will be used when we have cisco 6513 and Huawei switch.
Question2: Do we need to just configure trunk on 6413 and huawei switch to extend Layer 2 and all the configurations will be done by MPLS provider OR there is something that we must configure on our switches in data center.
Thank you for your help.
It appears there is no device or circuit redundancy between the two sites. That being the case, a simple L2TPv3 pseudowire doing QinQ will work fine.
To do LAN extension you are going to need to plug a port from R1 and R2 directly into the local LANs (assuming these devices have the capability and performance). These are going to be layer 2 ports - with no IP config on them.
Then whatever is presented at one routers port will "pop" out the other side. It acts just like a long piece of wire. They can configure whatever VLANs they want to be presented to R1 and R2's port.
I will surely connect switches with the MPLS cloud and keep it on L2. But what technology do we need to go for.
I am not wrong should we consider EoMPL (P2P) or VPLS (Multipoint) as the service provider is MPLS.
Any suggestion ? Thanks.