Hi all,
for a client (sort of ISP) of mine I was asked to develop an infrastructure where 60-70 tenants, the majority of which have already chosen their own VLAN and IP addressing with phisically segregated/cabled networks, will have their own segregated overlay. The idea is to onboard the tenants by providing them a delivery point in each room of the building they have rented and then have them use the ISP's (i.e. my client's) internet connection. Other services will be offered to each tenant like the possibility to have VMs in a small DC or the extension of their network throug the wireless.
We thought from day one of an underlay made by P2P L3 links on top pf which we could have many flavours of the L2VPN (EVPN BGP, VPLS, etc etc)
My client (the would be ISP) asked about QinQ and though I think it does what the client asks for but somehow I don't see it as future proof.
My opinion is that an underlay based on IP P2P links is more flexible and may accomodate what also comes in the near future in terms of underlay. On the contrary it's true that QinQ is way cheaper compared to SP-level L2VPN/L3VPN.
Would you reccomend any Cisco (or external) white paper or document where such approaches are compared? Where the differences between QinQ and L2/L3VPN are highlighted and will be more comprehensible?
TIA,
Gio
EDIT:
1) Obviously with QinQ you cannot delimit the L2 domain as you extend the tenant's L2 domain over the whole infrastructure. So QinQ provides just "L2VPN" and if the tenants manage poorly their L2 domain, depending on the bandwidth given to them, they may create problems.
