What is round-robin DNS?

Round-robin DNS is the practice of publishing multiple records with the same name and type, commonly several A or AAAA records, so clients can use different endpoints.

Is round-robin DNS load balancing?

Only in a loose sense. It can spread requests, but it has no built-in health checks, capacity awareness, or per-request routing.

Does round-robin DNS fail over automatically?

No. If one IP is down and the record is still published, some clients can still receive it. Automatic failover requires health checks and record withdrawal or traffic steering.

Does TTL fix round-robin failover?

A low TTL reduces how long stale answers stay cached, but it does not tell DNS which endpoint is healthy. Low TTL is useful, not sufficient.

Can I use round-robin with IPv6?

Yes. You can publish multiple AAAA records just like multiple A records.

When is round-robin DNS a good fit?

It is fine for simple active-active services where every endpoint is healthy, equivalent, and able to handle traffic. It is weak for disaster recovery or precise load balancing.

Performance

Round-Robin DNS Explained

How round-robin DNS works, when multiple A or AAAA records are useful, and why round-robin is not the same as health-checked DNS failover.

Updated 3 June 2026

TL;DR

Round-robin DNS publishes multiple records for the same name, such as three A records for www.example.com. Resolvers and clients may rotate or choose among them, which can spread traffic roughly across endpoints. It is simple and cheap, but it is not real load balancing: DNS does not know whether an endpoint is overloaded, unhealthy, or close to the user unless your DNS platform adds separate health and steering logic.

What you'll learn

Explain how round-robin DNS works
Show valid A and AAAA record examples
Understand why caching makes traffic distribution uneven
Decide when round-robin is enough and when to use real failover

Round-robin DNS is the simplest way to put more than one endpoint behind one DNS name.

Instead of publishing one A record:

www 300 IN A 203.0.113.10

you publish several:

www 300 IN A 203.0.113.10
www 300 IN A 203.0.113.11
www 300 IN A 203.0.113.12

When a resolver asks for www.example.com, the authoritative server returns the set of addresses. Different resolvers and clients may use different orders or cache different answers, so traffic spreads out roughly.

What Round-Robin DNS Does Well

Round-robin is useful when all endpoints are equivalent:

same application version
same region or acceptable latency from all users
same capacity
same health status
same TLS and HTTP behavior

Good example:

three web nodes behind the same deployment, all healthy, all interchangeable

In that case, round-robin can reduce reliance on one IP address.

What It Does Not Do

Round-robin DNS does not perform active load balancing.

It does not know:

which server is overloaded
which server is down
which server is closest to a user
whether a deploy broke one endpoint
whether a firewall blocks one region

If this record set is published:

www 300 IN A 203.0.113.10
www 300 IN A 203.0.113.11
www 300 IN A 203.0.113.12

and 203.0.113.11 fails, DNS will still hand it out until the record is removed or a health-check system withdraws it.

That is the main trap.

Why Traffic Is Not Even

People often expect three A records to mean exactly one third of traffic per IP.

That is not how DNS works.

Unevenness comes from several places:

Recursive resolvers cache answers for many users.
Some resolvers preserve order; some rotate it.
Some clients try the first address only.
Some clients retry another address on failure.
Large public resolvers can represent many users behind one cache.
Low TTLs help but do not remove resolver behavior.

Round-robin is approximate distribution, not a traffic contract.

Round-Robin with A and AAAA Records

IPv4:

api 300 IN A 203.0.113.10
api 300 IN A 203.0.113.11

IPv6:

api 300 IN AAAA 2001:db8::10
api 300 IN AAAA 2001:db8::11

Dual-stack:

api 300 IN A    203.0.113.10
api 300 IN A    203.0.113.11
api 300 IN AAAA 2001:db8::10
api 300 IN AAAA 2001:db8::11

Dual-stack clients decide between IPv4 and IPv6 using their own address-selection logic. Round-robin does not control that choice.

Round-Robin vs Failover

Feature	Round-robin DNS	DNS failover
Multiple endpoints	Yes	Yes
Health checks	No	Yes
Removes dead endpoints	No	Yes
Predictable traffic split	No	Sometimes
Good for disaster recovery	Weak	Better
Operational complexity	Low	Medium

If you need failover, read DNS failover design patterns.

Round-Robin vs Anycast

Round-robin publishes several destination addresses.

Anycast publishes one address from many physical locations:

round-robin: one name -> many IPs
anycast:     one IP -> many network locations

For authoritative DNS itself, anycast is the normal modern answer. For your application endpoints, round-robin may still be useful, but it is usually less precise than a load balancer or traffic-steering system.

See What is Anycast DNS? for the routing side.

When Round-Robin Is Enough

Use it when:

all targets are healthy and equivalent
downtime on one target is tolerable
the app retries cleanly
you want simple distribution, not precise steering
you can remove bad records quickly

Avoid it as the only resilience mechanism when:

one bad endpoint creates user-visible errors
endpoints live in different regions with very different latency
health state changes often
you need compliance-grade disaster recovery
traffic weights matter

Safer Operating Pattern

If you use round-robin:

Keep TTL moderate, often 60-300 seconds for active endpoints.
Monitor every target independently.
Automate record removal for hard failures if possible.
Keep capacity headroom so remaining endpoints survive one failure.
Test client retry behavior, not just DNS answers.

DNS can point traffic at endpoints. It cannot make unhealthy endpoints healthy.

Frequently asked questions

What is round-robin DNS?: Round-robin DNS is the practice of publishing multiple records with the same name and type, commonly several A or AAAA records, so clients can use different endpoints.
Is round-robin DNS load balancing?: Only in a loose sense. It can spread requests, but it has no built-in health checks, capacity awareness, or per-request routing.
Does round-robin DNS fail over automatically?: No. If one IP is down and the record is still published, some clients can still receive it. Automatic failover requires health checks and record withdrawal or traffic steering.
Does TTL fix round-robin failover?: A low TTL reduces how long stale answers stay cached, but it does not tell DNS which endpoint is healthy. Low TTL is useful, not sufficient.
Can I use round-robin with IPv6?: Yes. You can publish multiple AAAA records just like multiple A records.
When is round-robin DNS a good fit?: It is fine for simple active-active services where every endpoint is healthy, equivalent, and able to handle traffic. It is weak for disaster recovery or precise load balancing.

Ready to manage your DNS with confidence?

DNScale provides anycast DNS hosting with a global network, real-time analytics, and an easy-to-use API.

Start free