Naming

IPs are not human-readable.

Not always the appropriate granularity

The address names an interface
This however does not give information about the kind of service / hardware

A file maps names to addresses

Unix & Linux
- /etc/hosts
Windows
- C:\Windows\System32\drivers\etc\hosts

These are simple but neither automatic nor scalable, which led to DNS.

Was initially RFC882
Now is RFC1035, 1987

DNS is a consistent namespace

No reference to addresses, routes etc
Is hierarchical, distributed and cached
- All of which help with scalability
Federated - sources control trade-off
- This just means DNS is worldwide
Flexible - many records
Simple client-server name resolution protocol

Components

Domain name space and resource records
- Tree-structured name space
- Data associated with names
Name server
- Contains records for a subtree
- May cache information about any part of the tree
Resolver
- Extracts information from the tree upon client requests
- gethostbyname()

Root

Ultimate authority with the US Dept. of commerce (NITA)
Managed by IANA, operated by ICANN, maintained by Verisign
Started with only thirteen root server clusters
- Now many more
Top-level domains, TLDs
- Operated by registrars, delegated by ICANN
Delegate zones to other registrars
- and so on down the hierarchy
Eventually, a customer rents a name - their zone
- Registrar installs appropriate resource records
- Associated with names within the zone

Query

Query generated by resolver
- e.g. call to gethostbyname(), gethostbyaddr()
Carried in single UDP/53 packet
- Or more rarely TCP/53 in case of truncation
- UDP is not smart and therefore does not follow traffic routing (it is selfish)
  - It is beneficial for the internet as a whole to use UDP sometimes
Header followed by question
- ID, Q/R, opcode, AA/TC/RD/RA, response code, counts
- Query type, query class, query name

Response consists of three RRsets following the header and question

Answers: RRs that the server had for the QNAME
Authoritatives: RRs pointing to an authority for the name
Additionals: RRs related to the question but don’t answer it

Common Resource Records

A / CNAME / PTR

www.cs.nott.ac.uk.			61272	IN	CNAME	pat.cs.nott.ac.uk.

pat.cs.nott.ac.uk.			68622	IN	A		128.243.20.9

pat.cs.nott.ac.uk.			68622	IN	A		128.243.21.19

9.20.243.128.in-addr.arpa.	39617	IN	PTR		pat.cs.nott.ac.uk.

cname refers to the mapping of the domain name to its IP (or another domain) & ports

Can have 2 authoritative records

NS

cs.nott.ac.uk.	10585	IN	NS	ns1.nottingham.ac.uk.
cs.nott.ac.uk. 	10585	IN	NS	ns2.nottingham.ac.uk.
cs.nott.ac.uk. 	10585	IN	NS	marian.cs.nott.ac.uk.
cs.nott.ac.uk.	10585	IN	NS	extdns1.warwick.ac.uk.
cs.nott.ac.uk.	10585	IN	NS	extdns2.warwick.ac.uk.

It is good practice to have an external DNS, UoN uses Warwick as an external DNS.

MX

nott.ac.uk.	3600	IN	MX	1 mx191.emailfiltering.com.
nott.ac.uk.	3600	IN	MX	2 mx192.emailfiltering.com.
nott.ac.uk	3600	IN	MX	3 mx193.emailfiltering.com.

What happens when the resolver queries a server that doesn't know the answer? There are two solutions:

Iterative (required)
- Server responds indicating who to ask next
- This method is slower and more difficult to retrieve an answer
Recursive (optional)
- Server generates a new query to the next server

Load Balancing

DNS may have multiple servers. When a query arrives, various algorithms can be used to choose the best one, for example, based on geographical location.

Operational & Security Issues

Usually need primary and secondary servers
- Separate IP netblocks, physical networks - more robust
- DNS is a very common single point of failure
Cache poisoning
- Caching and soft-state means bad data propagates and can persist for some time
- Even if through simple mistakes (or of course malicious attacks)
Man-in-the-middle attacks
- Can happen with both iterative & recursive queries

4.1 KiB Raw Blame History Unescape Escape