You might want to review some of those recommendations... They don't fall into best practices....and can be problematic.
There are many discussions and docs on this topic...A while ago I had a discussion on this topic on Nanog list.. There was a 'key' item pointed out to me and I had to grasp that to gain a full understanding on what size to assign. ... You have to think interns of subnet's and not qty of Addresses to assign properly. Here is an excellent presentation.. http://www.txv6tf.org/wp-content/uploads/2012/09/Doyle-IPv6-Address-Design.pdf http://meetings.ripe.net/ripe-49/presentations/ripe49-ipv6-guidelines.pdf -------- explanation:- There is no concept of NAT in IPv6, and IPv6 networks by design are 'stacked routers' vs a traditional flat LAN (as in ipv4 networks). so it is important to provide the ability to break down the subnet that you assign to your customers.. e.g. a /64 is / should be the smallest level of assignment (this of a LAN, or any Link (PTP or PTMP).) i.e. one lan segment a /56 would allow for 256 Lan Segments... or a 8 level deep network a /48 would allow for 65536 Lan segments or a 16 level deep network (I am going my memory on the network level depth, so I may be over estimating.. :) ) There is a lot of wisdom behind assign a /48 to each customer.. .and there is a lot of discussion on doing something a bit smaller for smaller customers e.g. a /56. Think of it this way... IPv6 is how IoT communicates.. and each device connecting to the 'home network' can act as a router, thus a layers.. e.g. Smart Cars, will have a router, to connect to devices in the car, and it will connect to the home network as a sub layer network... etc... Regards Faisal Imtiaz Snappy Internet & Telecom 7266 SW 48 Street Miami, FL 33155 Tel: 305 663 5518 x 232 Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net ----- Original Message ----- > From: "Paul Stewart" <p...@paulstewart.org> > To: "Animal Farm" <af@afmug.com> > Sent: Sunday, January 15, 2017 8:05:45 AM > Subject: Re: [AFMUG] Best Way to implement dual stack IPv4/6 > Ok .. lots of folks here can probably help but you’ll need to be a bit more > specific :) > > Advise on planning for it? Implementing it on specific gear? last mile to > customer? … over entire deployment? > > Here’s some info that may help (from network perspective): > > Full dual stack deployment across network with Juniper MX throughout > /126 for point to point > /128 for loopback > OSPFv3 for connected networks and loopbacks > iBGP for full internal and external routes (where needed) > > For DSL/Wireless (PPPOE): > /40 and /47 IP pools per device that is serving customers (varies with number > of > customers per device) > Customers receive /64 allocation via DHCP-PD over their existing PPPOE > connection > > For Cable (DHCP): > /48 and /56 IP pools per geographic region that is serving customers > Customers receive /64 allocation via DHCP-PD > > For fiber (static or BGP) > /126 point to point > /64 allocation by default, /56 upon request, larger if justified > > > I’ve pretty much followed this model in one form or other since 2008 when I > did > the first network end to end > > Paul > > > > >> On Jan 14, 2017, at 11:58 PM, Mitch Koep <af...@abwisp.com> wrote: >> >> Need some advise on implementing dual stack. >> >> Best practice or practical. >> >> Thanks >> >> Mitch Koep >> >> 219-851-8689 cell