implementing nslookup

There is a very famous inteview question, what happens when you type google.com in your search bar The first and the most obvious jargon is DNS. DNS is a translation system which converts human readable domain names to computer understandable IP addresses.

I started with the concept of DNS and how it works. then came to know about UNIX utilities like nslookup and dig.

This blog is whole about my learning and building nslookup from scratch.

1. Intro and understanding (10 sept 2026)#

I started with reading about DNS headers and the DNS query process.

DNS is the Domain Name System, a hierarchical naming system for computers and services on the Internet.

DNS is based on UDP sockets but , if there is a big DNS query, it is split into multiple packets and transmitted with the help of TCP.

2. Understanding DNS query process (11 sept 2026)#

Next thing was to understand what is exactly DNS query process. For that we have to understand Types of DNS servers.

• Root DNS servers: The first level of DNS servers that are responsible for resolving the top-level domain (TLD) names.
• TLD DNS servers: The second level of DNS servers that are responsible for resolving the second-level domain (SLD) names.
• Authoritative DNS servers: The third level of DNS servers that are responsible for resolving the domain names.
• Recursive DNS servers: The fourth level of DNS servers that are responsible for resolving the domain names.

if you want more detail refer : Cloudflare docs

So we are just quering the Recursive DNS server and not recursively finding the authoritative DNS servers. some of them are google's DNS 8.8.8.8 and cloudflare's DNS 1.1.1.1 or provided by your ISP.

3. Making DNS request yourself (13 sept 2026)#

I used nslookup and wireshark to actually make DNS requests and understand the process.

4. DNS header structure#

struct dns_question_format{
    uint16_t qtype; // type of question
    uint16_t qclass; // class of question
    char qname[256]; // name of question
};
struct dns_resource_record_format{
    uint16_t type; // type of resource record
    uint16_t class; // class of resource record
    uint32_t ttl; // time to live
    uint16_t rdlength; // length of rdata
    char rdata[256]; // rdata of resource record
};
struct dns_request_format{
    uint16_t id; // for a application for multiple queries
    uint16_t flags; // some flags that indicate the type of query
    uint16_t qdcount; // number of questions
    uint16_t ancount; // number of answers
    uint16_t nscount; // number of name servers
    uint16_t arcount; // number of additional records
    struct dns_question_format questions[1]; // array of questions
    struct dns_resource_record_format answers[1]; // array of answers
    struct dns_resource_record_format name_servers[1]; // array of name servers
    struct dns_resource_record_format additional_records[1]; // array of additional records
};

DNS header is of 12 bytes and the content inside it are ~500 bytes.

5. Structuring the project#

I don't only want to implement DNS but to write code in more maintainable, production level manner so that it is easier to read and modify later.

For this project as the external dependency are very less so a simple MakeFile is enough

I have made separate modules for each work in project i.e. All dns related stuff at one place, networking stuff etc.

All the top level file imports are itself in the header file but will surely make a structure to include only 1 header file