What protocol does DNS use?

DNS mainly runs over UDP, but in some cases it uses TCP.

In simpler terms: ordinary DNS requests go over UDP, and TCP is used in special situations.

---

#The main protocol: UDP

#Why DNS usually uses UDP

By default DNS works over UDP on port 53, because:

• requests and responses are small
• no connection needs to be established
• the answer arrives quickly
• there is less overhead

DNS is a short question and a short answer, so UDP is a perfect fit.

---

#When DNS uses TCP

DNS switches to TCP (also port 53) in the following cases:

#1. A large response size

If the DNS response:

• does not fit in a UDP packet
• contains many records (for example, with DNSSEC)

TCP is used.

---

#2. DNS zone transfer

When synchronizing DNS zones between servers:

• primary -> secondary

TCP is always used, because:

• there is a lot of data
• reliability matters

---

#3. DNSSEC

When DNSSEC is used:

• responses become larger
• TCP is needed more often

---

#Briefly, in a table

---

#An important point for an interview

If DNS is an application-layer protocol, why UDP/TCP here?

The answer:

• DNS is an application-layer protocol
• UDP and TCP are the transport
• DNS uses transport protocols for delivery

---

#Real-life example

UDP is like:

• asking a short question and getting an answer

TCP is like:

• sending a package of documents with delivery confirmation

---

#Short interview answer

Remember this formulation:

DNS mainly runs over UDP on port 53, while TCP is used for large responses, DNSSEC, and for zone transfers between servers.