DNS

Computer Network & Network Security System — Application Layer, NEC licence examination syllabus (Nepal Engineering Council).

DNS: the internet's phone book, running silently in the background

You've never once typed a raw IP address into your browser, and DNS is the entire reason why.

Humans remember names ("google.com"). Computers route using numbers (IP addresses). DNS is purely the translation layer between the two.

🌳 Namespace and Hierarchy

RootThe top of the tree (represented by a trailing dot, usually invisible).

TLDTop-Level Domain — .com, .org, .np, etc.

DomainThe specific organization's name — google, wikipedia, etc.

SubdomainOptional further division — mail.google.com.

🔍 Recursive vs Iterative Resolution

RecursiveYour local DNS resolver does ALL the work on your behalf, chasing down the answer across multiple servers, and returns just the final result to you.

IterativeEach server queried gives back either the final answer OR a referral to the next server to ask — the ASKER has to follow the trail themselves.

Resource records are the actual entries stored in DNS — A record (IPv4 address), AAAA record (IPv6 address), MX record (mail server), CNAME (alias to another name), and more.

💡 One-liner: "What record type does DNS use to find the mail server for a domain?" → MX (Mail Exchange) record.

Why the root servers are not overwhelmed

A hierarchy with a single root looks like an obvious bottleneck. Every lookup starting there would be impossible — and the reason it works is the part the structure alone does not explain.

Caching at every level. Your resolver remembers answers, so the second lookup of a name never leaves your machine. It also caches the TLD servers for .com and .np, so subsequent lookups skip the root entirely. In practice the root is consulted rarely — mostly when a resolver starts cold.

⏱️ What TTL controls

Long TTLFewer lookups, less load, faster responses — but a changed record takes that long to reach everyone.

Short TTLChanges propagate quickly, at the cost of far more query traffic.

💡 This is why administrators lower the TTL days before a planned server move, then raise it again afterwards. The old value must expire everywhere before the change can propagate quickly — a practical detail that makes the mechanism concrete.

Recursive and iterative queries

RECURSIVE client → resolver: "give me the final answer" The resolver does all the work and returns one result. This is what your machine asks its ISP's resolver. ITERATIVE resolver → root: "who knows about .com?" root replies with a REFERRAL, not an answer. resolver → .com server → referral to example.com resolver → example.com server → the actual address

The distinction is who does the chasing. A client makes one recursive request; the resolver then performs several iterative ones on its behalf. Servers higher in the tree answer only iteratively — refusing to do the work for arbitrary clients is exactly what keeps them able to serve everyone.

💡 DNS normally runs over UDP, which fits the transport chapter's argument: one small question, one small answer, so a TCP connection would cost more round trips than the query itself. It falls back to TCP when a response is too large to fit in a datagram — most commonly for zone transfers between servers.

Syllabus points

Namespace and hierarchy
Resolution: recursive vs iterative
Resource records

Create a free account to tick topics off, take notes as you read, watch the video lessons and get a day-by-day study plan built around your exam date.