Explainer

IPv6 in the cloud: what teams need to know

Learn how IPv6 in the cloud changes addressing, routing, firewalls, and DNS, plus the mistakes to avoid and a safe way to start.

By the Cloudverg engineering team · · 5 min read

What it is

IPv6 in the cloud is the same IPv6 protocol you meet on any network, run inside a provider's virtual network. Addresses are 128 bits, written as eight groups of hexadecimal digits. A cloud network gets an IPv6 prefix, subnets take slices of it, and resources get addresses from those slices. The big shift is that IPv6 has enough addresses to give every workload a globally unique address, so network address translation is no longer a design requirement.

That changes the mental model. In IPv4 cloud designs, private addresses hide behind NAT gateways or load balancers. In IPv6, an address can be reachable from the internet unless routing and firewall rules say otherwise. Privacy comes from policy, not from address scarcity. Teams still use private ranges, such as unique local addresses, for internal traffic, but public global unicast is normal.

Most teams run dual stack at first: IPv4 and IPv6 on the same network, load balancer, and host. Dual stack lets old clients keep working while new clients use IPv6. It also doubles some of the operational surface, because every rule, route, and monitor needs an IPv6 path.

How it works

In AWS, a VPC can have an IPv6 CIDR block, often a /56 from Amazon, and each subnet gets a /64. Instances receive addresses through the provider's IPv6 assignment method, which can be SLAAC or DHCPv6. Route tables need IPv6 routes, including a default route to an internet gateway or an egress-only internet gateway. Security groups and network ACLs must allow the traffic. Azure virtual networks and Google Cloud VPCs have their own dual-stack controls, but the shape is similar: a prefix at the network level, a subnet prefix, routes, and firewall policy.

DNS is where many migrations stall. You need AAAA records for names that should answer over IPv6, and clients need working IPv6 connectivity to use them. A common pattern is to publish AAAA only after the service, load balancer, and firewall are tested. Check internal and public DNS views. Resolvers can return different records, and an IPv6 record in one view but not the other causes odd failures.

Some clouds and services offer NAT64 and DNS64 for IPv6-only subnets that must reach IPv4 destinations. That is useful for constrained networks, but it adds a translation layer to debug. For most business systems, dual stack with clear routes is easier to run. Check each managed service, load balancer, database endpoint, and logging agent for IPv6 support before you promise it.

Why it matters for a business

The practical reason is reachability. Mobile carriers, broadband providers, and partner networks increasingly run IPv6. If your public endpoint is IPv4 only, some users take a slower path or fail when they are on IPv6-only networks. A dual-stack endpoint keeps both paths open.

IPv6 also helps with address space in mergers, acquisitions, and large internal networks. Two companies with overlapping IPv4 ranges create painful routing work. IPv6 prefixes give you room to number new environments without reusing the same private blocks. That is an architecture benefit, not a cost saving on its own.

Security and audit teams care because IPv6 changes what is exposed. A workload with a global address can be reachable if the firewall allows it. You need inventory, policy, and logs that cover IPv6, or you have a blind spot. Procurement and customer questionnaires may ask for IPv6 support. A tested dual-stack design answers those questions with evidence.

Common mistakes

The first mistake is copying IPv4 habits. Teams build NAT gateways for IPv6 or assume private addresses are hidden. IPv6 has no NAT at the cloud edge by default. If you need outbound-only access, use an egress-only gateway or firewall policy, not an ad hoc NAT box.

The second is blocking ICMPv6. IPv6 depends on neighbor discovery, router advertisements, and path MTU discovery. A blanket deny on ICMPv6 breaks address resolution and causes mysterious packet drops. Allow the required ICMPv6 types at the right points, and test with ping6 and traceroute6.

Other common failures: publishing AAAA records before the firewall is ready, forgetting that network ACLs are stateless, or missing IPv6 in VPN and peering configs. Monitoring tools that only collect IPv4 flow logs leave gaps. Address plans that hardcode prefixes into scripts make changes slow. Treat IPv6 as a first-class path in code, tests, and runbooks.

How to get started

Start with an inventory, not a big bang. List public endpoints, internal services, VPNs, load balancers, DNS zones, and monitoring tools. Mark which ones support IPv6 today and which need work. Pick one low-risk service, such as a marketing site or an internal API, and run it dual stack.

Write an addressing plan before you request prefixes. Decide who owns the cloud IPv6 CIDR, how subnets get /64s, and which ranges are for public services, private services, and transit. Put the plan in your IPAM tool or a versioned document. Then create the IPv6 CIDR, subnets, routes, security groups, and DNS records in code so the change is repeatable.

Test from outside and inside: IPv6-only clients, dual-stack clients, and IPv4-only clients. Check logs, metrics, and alerts for the new address family. Keep a rollback path, usually by leaving IPv4 in place until IPv6 is stable. If you want a second opinion on the design before you commit, the Expert architecture session at $450/month gives you 90 minutes with a senior architect and a written summary. Other options include a cloud consultant, your provider's support plan, or an internal network team.

Frequently asked questions

Is IPv6 required in the cloud?

No, most clouds still let you run IPv4-only. But public services, partner links, and mobile users may work better with IPv6. Treat it as a design requirement when a customer or vendor asks for it, or when you need address space.

Can I run IPv4 and IPv6 at the same time?

Yes. Dual stack is the normal first step. You keep IPv4 for old clients and add IPv6 routes, firewall rules, and AAAA records for the new path. Test both paths before you remove anything.

What breaks when I enable IPv6?

Common breakpoints are firewall rules, ICMPv6, DNS records, and monitoring. A global IPv6 address can be reachable if policy allows it, so review security groups and network ACLs first. Test with real clients and keep IPv4 in place during the rollout.

Monthly plan

Want help with this? Expert architecture session

$450/month

90 minutes a month · Within 5 business days

Ninety minutes a month with a senior architect on the cloud, AI or security question that is blocking you, plus a written summary.

Renews monthly. Cancel any time.
What is included

Related guides

More cloud networking guides   Cloudverg platforms

Smallest way to start

Get a senior architect on your problem this week

Ninety minutes a month on the cloud, AI or security question that is blocking you, a pre-read of your material and a written summary within two business days.

$450/month

Renews monthly. Cancel any time.