15-441 Lecture 7 Outline What is DNS? How resolve name → IP addr?

Outline

15-441 Lecture 7

DNS

•DNS DNS Design •DNS DN Today d (Extra credit, remind me at end)

C Copyright i ht © S Seth th Goldstein, G ld t i 2008 Based on slides from previous 441 lectures Lecture 13

15-441 © 2008

1

Lecture 13

What is DNS?

15-441 © 2008

2

How resolve name Æ IP addr?

• DNS (Domain Name Service) is primarily used to translate human readable names into nto machine mach ne usable addresses, e.g., IP addresses. •DNS goal: – Efficiently locate resources. E.g., Map name Æ IP address y users over a large g area – Scale to many – Scale to many updates

Lecture 13

15-441 © 2008

3

Lecture 13

15-441 © 2008

4

Obvious Solutions (1)

Obvious Solutions (2) Why not use /etc/hosts? Or g nal Name to Address ddress Mapping Mapp ng •Original – Flat namespace – /etc/hosts – SRI kept main copy g y – Downloaded regularly •Mid 80’s this became untenable. Why? •Count of hosts was increasing: machine per d domain i Æ machine hi per user – Many more downloads /etc/hosts still exists. exists – Many M more updates d t

Why not centralize DNS? Single point of failure •Single •Traffic volume •Distant centralized database •Single point of update •Doesn’t scale!

Lecture 13

15-441 © 2008

5

Lecture 13

15-441 © 2008

15-441 © 2008

6

Programmer’s View of DNS

Domain Name System y Goals •Basically a wide-area distributed database ((The he biggest b ggest in n the world!) •Scalability Decentralized maintenance •Decentralized •Robustness Global scope •Global – Names mean the same thing everywhere •Don’tt need all of ACID •Don – Atomicity – Strong consistency •Do need: distributed update/query & Performance

Lecture 13

• Conceptually, programmers can view the DNS database as a collection of millions of host entry structures: /* DNS host entry structure struct hostent { char *h_name; ; / /* char **h_aliases; /* int h_addrtype; /* int h_length; /* char **h_addr_list; /* };

*/ official domain name of host */ / null-terminated array of domain names */ host address type (AF_INET) */ length of an address, in bytes */ null-termed array y of in_addr structs */

– in_addr is a struct consisting of 4-byte IP addr • Functions for retrieving host entries from DNS: – gethostbyname: query key is a DNS host name. – gethostbyaddr: query key k is an IP address. dd 7

Lecture 13

15-441 © 2008

8

DNS Message Format

12 bytes

Identification

Flags

No. of Questions

No. of Answer RRs

No of Authority RRs No.

No of Additional RRs No.

Name, type fields for a query

Questions (variable number of answers)

RRs in response to query

Answers (variable number of resource records)

Records for authoritative servers

DNS Header Fields •Identification – Used to match up request/response •Flags – 1-bit 1 bit to mark query or response – 1-bit to mark authoritative or not – 1-bit 1 bit to request recursive resolution – 1-bit to indicate support for recursive resolution

Authority (variable number of resource records)

Additional “helpful info that may be used

Additional Info (variable number of resource records)

Lecture 13

15-441 © 2008

9

Lecture 13

Different kinds of mappings are possible:

name, value, type, ttl)

•1 1 mapping between domain name and IP addr: •1-1

• DB contains tuples called resource records (RRs)

provolone.crcl.cs.cmu.edu maps to 128.2.218.81

• Classes = Internet (IN), Chaosnet (CH), etc. • Each class defines value associated with type

•Multiple domain names maps to the same IP addr: www.scs.cmu.edu and www.cs.cmu.edu both map to 128.2.203.164

For “IN” class:

• Type Type=CNAME CNAME

• Type Type=A A • Type=NS T NS – name is domain (e.g. foo.com) – value is name of authoritative name server for this domain Lecture 13

•Single domain name maps to multiple IP addresses: aol.com and www.aol.com map to multiple IP addrs.

• name is an alias name for some “canonical” name • value is canonical name

– name is hostname – value is IP address

•Some valid domain names don’t map to any IP addr: crcl.cs.cmu.edu doesn’t have a host

• Type=MX

15-441 © 2008

10

Properties of DNS Host Entries

DNS Records RR format: (class,

15-441 © 2008

• value is hostname of mailserver associated with name 11

Lecture 13

15-441 © 2008

12

DNS Design: Hierarchy Definitions root

org

net

gwu ucb

edu com uk cmu bu mit cs ece crcl

Lecture 13

DNS Design: Zone Definitions

• Each node in hierarchy stores a list of names that end with same suffix • Suffix = path up tree • E.g., given this tree, where would following be stored: • Fred.com Fred com • Fred.edu • Fred.cmu.edu • Fred.crcl.cs.cmu.edu • Fred.cs.mit.edu

15-441 © 2008

root

org

gwu ucb

cmu

bu mit

Subtree Single l node d Complete Tree

13

Lecture 13

15-441 © 2008

14

DNS: Root Name Servers

•Zones are created by convincing owner node to create/delegate g a subzone – Records within zone stored in multiple redundant name servers – Primary/master P i / name server updated d d manually ll – Secondary/redundant servers updated by zone transfer of name space

• Responsible p for “root” zone • 13 root name servers – Currently y {a-m}.root-servers.net

• Local name servers contact root servers when h they h cannot resolve a name • Why Wh 13?

• Zone transfer is a bulk transfer of the “configuration” of a DNS server – uses TCP to ensure reliability

•Example: – CS.CMU.EDU created by CMU.EDU admins – Who creates CMU.EDU or .EDU? 15-441 © 2008

edu com uk

cs ece crcl

DNS Design: Cont.

Lecture 13

net

• Zone = contiguous section of name space p • E.g., Complete tree, single node or subtree • A zone has an associated set of name servers • Must store list of names and d tree t links li k

15

Lecture 13

15-441 © 2008

16

Not really 13!

Check out anycast) Lecture 13

So Far •Database structure xyz – Hierarchy of labels x.y.z – Organized into zones – Zones have nameservers (notice plural!) •Database layout – Records which map namesÆnames, namesÆip, namesÆip etc. Programmer API API: gethostbyname, … •Programmer

10/08, from

www.root-servers.org

15-441 © 2008

17

Lecture 13

15-441 © 2008

Servers/Resolvers

18

Typical Resolution

•Each host has a resolver – Typically a library that applications can link to – Local name servers hand-configured (or DHCP) (e g /etc/resolv.conf) (e.g. /etc/resolv conf) •Name servers or – Either responsible for some zone or… – Local servers

root & edu DNS server

www.cs.cmu.edu

Client

• Do lookup of distant host names for local hosts • Typically answer queries about local zone

ns1.cmu.edu ns1 cmu edu DNS server

Local DNS server

ns1.cs.cmu.edu ns1 cs cmu edu DNS server

Hmm: Notice root server returned NS ns1.cmu.edu Lecture 13

15-441 © 2008

19

Lecture 13

15-441 © 2008

20

Typical Resolution

Lookup Methods R Recursive i query:

•Steps for resolving www.cmu.edu

root name server

• Server goes out and searches for more info • Only returns final answer or “not found”

– Application calls gethostbyname() (RESOLVER) – Resolver contacts local name server (S1) – S1 q queries root server (S2) for (www.cmu.edu) – S2 returns NS record for cmu.edu (S3) – What about A record for S3?

2 iterated query 3

4

Iterative query:

• Server responds with as much as it knows. • “I don’t d ’t k know thi this name, but ask this server”

• This h is what h the h additional dd l info f section is for f (PREFETCHING) ( E E )

– S1 queries S3 for www.cmu.edu – S3 returns A record for www.cmu.edu www cmu edu

7 local name server dns.eurecom.fr

1

8

intermediate name server dns.umass.edu

5

6 authoritative name server dns.cs.umass.edu

requesting host

•Can Can return multiple A records Æ What does this mean? Lecture 13

15-441 © 2008

gaia.cs.umass.edu

surf.eurecom.fr

Workload impact on choice?

• Root/distant server does iterative • Local server typically does recursive 21

Lecture 13

How to manage workload?

15-441 © 2008

22

Workload and Caching

•Does root nameserver do recursive lookups? What about other zones? •What •What about imbalance in popularity? – .com com versus .dj dj – google.com versus bleu.crcl.cs.cmu.edu? •How How do we scale query workload?

•DNS responses are cached – Quick response for repeated translations – Other queries may reuse some parts of lookup g NS records for domains • E.g.,

•DNS negative queries are cached – Don’t have to repeat past mistakes – E.g., misspellings, search strings in resolv.conf

•How do you handle updates?

Lecture 13

15-441 © 2008

23

Lecture 13

15-441 © 2008

24

Workload and Caching

Typical Resolution

•DNS responses are cached – Quick response for repeated translations – Other queries may reuse some parts of lookup • E.g., g NS records for domains

•DNS negative queries are cached

root & edu DNS server

www.cs.cmu.edu

– Don’t have to repeat past mistakes – E.g., misspellings, search strings in resolv.conf

Client

•Cached data periodically times out

ns1.cmu.edu ns1 cmu edu DNS server

Local DNS server

ns1.cs.cmu.edu ns1 cs cmu edu DNS server

– Lifetime (TTL) of data controlled by owner of data – TTL passed with every record

Lecture 13

15-441 © 2008

25

Lecture 13

Subsequent Lookup Example

15-441 © 2008

26

Reliability •DNS servers are replicated – Name service available if ≥ one replica is up – Queries can be load balanced between replicas

ftp.cs.cmu.edu

Client

Local DNS server

root & edu DNS server

•UDP used f for q queries

cmu edu cmu.edu DNS server

•Try alternate servers on timeout

– Need reliability Æ must implement this on top of UDP! – Why not just use TCP? – Exponential backoff when retrying same server

cs.cmu.edu cs cmu edu DNS server

Lecture 13

15-441 © 2008

•Same identifier for all queries – Don’t care which server responds

27

Lecture 13

15-441 © 2008

28

So far

Reverse DNS A Arpa: b k backronym Æ Address Add and dR Routing ti P Parameter t A Area

•Hierarchial name space

unnamed root

edu

arpa

in-addr

128

cmu

crcl

15-441 © 2008

29

Lecture 13

.arpa Name Server Hierarchy in-addr.arpa in addr.arpa

128

2

204

bleu 128 2 204 27 128.2.204.27

Lecture 13

•Method

– Maintain separate hierarchy based on IP names – Wr Write t 128.2.204.27 . . . 7a as 27.204.2.128.in-addr.arpa

•Managing

204

Lecture 13

– Given IP address, find its name

• Why is the address reversed?

cs

2

27

•Task

bleu 128.2.204.27

– Authority manages IP addresses assigned to it – E.g., CMU manages name space 2.128.in-addr.arpa

15-441 © 2008

30

Prefetching •Name servers can add additional data to response Why would they? •Why

a root-servers a.root servers.net net • • • m.root m root-servers servers.net net

chia.arin.net (dill henna, (dill, henna indigo, indigo epazote, epazote figwort, figwort ginseng) cucumber.srv.cs.cmu.edu, t-ns1.net.cmu.edu t-ns2.net.cmu.edu

mango.srv.cs.cmu.edu (peach,, banana,, blueberry) (p y)

• At each level of hierarchy, have group of servers that are authorized to handle that region of hierarchy 15-441 © 2008

31

Lecture 13

15-441 © 2008

32

Prefetching

Mail Addresses

•Name servers can add additional data to response Why would they? •Why •Typically used for prefetching – CNAME/MX/NS typically point to another host name – Responses include address of host referred to in “additional section”

Lecture 13

15-441 © 2008

•MX records point to mail exchanger for a name Eg – E.g. cmu.edu. cmu.edu.

2590 2590

IN IN

MX MX

10 CMU-MX4.ANDREW.cmu.edu. 10 CMU-MX5.ANDREW.cmu.edu.

•Addition of MX record type proved to be a challenge – How to get mail programs to lookup MX record for mail delivery? – Needed critical mass of such mailers – Could we add a new one now? 33

Lecture 13

15-441 © 2008

Outline

34

Root Zone •Generic Top Level Domains (gTLD) = .com, .net, .org, g etc… •Country Code Top Level Domain (ccTLD) = .us, .ca, .fi, .uk, etc… •Root R server ({ ({a-m}.root-servers.net) } ) also l used d to cover gTLD domains – Load on root servers was growing quickly! – Moving .com, .net, .org off root servers was clearly necessary to reduce load Æ done Aug 2000 – How significant an effect would this have?

•DNS Design

•DNS DNS Today

• On load? • On performance?

Lecture 13

15-441 © 2008

35

Lecture 13

15-441 © 2008

36

gTLDs

New Registrars

• Unsponsored – – – –

.com, .edu, .gov, .mil, .net, .org .biz Æ businesses .info i f Æ generall info i f .name Æ individuals

– – – – – – –

.aero aero Æ air air-transport transport industry .cat Æ catalan related .coop Æ business cooperatives there anything special about .com? .jobs Æ jobIs announcements goldstein as a gTLD? .museum Æ What museumsabout adding .goldstein .pro Æ accountants, lawyers, and physicians .travel Æ travel industry

•Network Solutions (NSI) used to handle all registrations, reg strat ons, root servers, etc… – Clearly not the democratic (Internet) way – Large number of registrars that can create new domains Æ However NSI still handles A root server

• Sponsored (controlled by a particular association)

• Starting up

– .mobi Æ mobile phone targeted domains – .post Æ postal – .tel Æ telephone related

• Proposed P d Lecture 13

– .asia, .cym, .geo, .kid, .mail, .sco, .web, .xxx – Whatever you want! 15-441 © 2008

37

Measurements of DNS

Lecture 13

15-441 © 2008

Measurements of DNS

•No centralized caching per site

•No centralized caching per site

• “Hit Hit rate for DNS:1 - (#DNS/#connections)Æ80%

• “Hit Hit rate for DNS:1 - (#DNS/#connections)Æ80%

• What does a typical page look like? Æ average of 4-5 imbedded objects Æ needs 4-5 transfers • This alone accounts f for 80% hit rate!

• What does a typical page look like? Æ average of 4-5 imbedded objects Æ needs 4-5 transfers • This alone accounts f for 80% hit rate!

– Each machine runs own caching local server y is this a problem? p – Why – How many hosts do we need to share cache? Æ recent studies suggest 10-20 hosts

– Each machine runs own caching local server y is this a problem? p – Why – How many hosts do we need to share cache? Æ recent studies suggest 10-20 hosts

– Is this good or bad? – Most Internet traffic was Web with HTTP 1.0

•Lower TTLs for A records does not affect performance •DNS performance really relies more on NS-record caching

Lecture 13

15-441 © 2008

38

– Is this good or bad? – Most Internet traffic was Web with HTTP 1.0

39

•Lower TTLs for A records does not affect performance •DNS performance really relies more on NS-record caching

Lecture 13

15-441 © 2008

40

Tracing Hierarchy (1)

Tracing Hierarchy (2)

•Dig Program – Allows querying of DNS system – Use flags fl to f find d name server (N (NS)) – Disable recursion so that operates one step at a time

•3 servers handle CMU names unix> dig +norecurse @e3.nstld.com NS kittyhawk.cmcl.cs.cmu.edu ;; AUTHORITY SECTION: cmu.edu. 172800 IN cmu.edu. 172800 IN cmu.edu. 172800 IN

unix> dig +norecurse @a.root-servers.net NS kittyhawk.cmcl.cs.cmu.edu

Zone

;; AUTHORITY SECTION: TTL edu edu. 172800 IN NS edu. 172800 IN NS edu. 172800 IN NS edu. 172800 IN Type NS edu. 172800 IN NS edu. 172800 IN NS edu. 172800 IN NS edu. 172800 IN NS Class edu. 172800 IN NS

Lecture 13

L3.NSTLD.COM. L3 NSTLD COM D3.NSTLD.COM. A3.NSTLD.COM. E3.NSTLD.COM. C3.NSTLD.COM. F3.NSTLD.COM. G3.NSTLD.COM. Value B3.NSTLD.COM. M3.NSTLD.COM.

– All .edu names handled by set of servers 15-441 © 2008

41

Lecture 13

Tracing Hierarchy (3 & 4)

NS NS NS NS

MANGO.SRV.cs.cmu.edu. PEACH.SRV.cs.cmu.edu. BANANA.SRV.cs.cmu.edu. BLUEBERRY.SRV.cs.cmu.edu.

•Quasar Quasar is master NS for this zone unix>dig +norecurse @blueberry.srv.cs.cmu.edu NS kittyhawk.cmcl.cs.cmu.edu ;; AUTHORITY SECTION: cs.cmu.edu. 300 IN Lecture 13

SOA

15-441 © 2008

42

•Motivations Æ large distributed database Scalab l ty – Scalability – Independent update – Robustness •Hierarchical database structure – Zones – How lookups are done •Caching/prefetching and TTLs •Reverse name lookup •What are the steps to creating your own domain?

unix> dig +norecurse @t-ns1.net.cmu.edu NS kittyhawk.cmcl.cs.cmu.edu

IN IN IN IN

CUCUMBER.SRV.cs.cmu.edu. T-NS1.NET.cmu.edu. T-NS2.NET.cmu.edu.

DNS (Summary)

•4 servers handle CMU CS names ;; AUTHORITY SECTION: cs.cmu.edu. 86400 cs.cmu.edu. 86400 cs.cmu.edu. 86400 cs.cmu.edu. 86400

NS NS NS

QUASAR.FAC.cs.cmu.edu.

15-441 © 2008

43

Lecture 13

15-441 © 2008

44