Application Layer2

8/6/2019 Application Layer2

http://slidepdf.com/reader/full/application-layer2 1/43

2: Application Layer 1

DNS: Domain Name System

People: many identifiers: SSN, name, Passport #

Internet hosts, routers:

IP address (32 bit) -used for addressingdatagrams

ńameµ, e.g.,hermite.cs.smith.edu -

used by humansQ: map between IPaddresses and name ?

Domain Name System: distributed database

implemented in hierarchy ofmany name servers

application-layer protocolhost, routers, name servers tocommunicate to resolve names(address/name translation)

note: core Internet

function implemented asapplication-layer protocol

complexity at network·sédgeµ




DNS

Why not centralize DNS?

single point of failure

traffic volume

distant centralizeddatabase

maintenance

doesn·t scale!

DNS services

Hostname to IPaddress translation

Host aliasing Canonical and alias

names

Mail server aliasing

Load distribution Replicated Web

servers: set of IPaddresses for onecanonical name




Structure of DNS Names

Each name consists of a sequence ofalphanumeric components separated byperiods

Examples:www.ssuet.edu.pk

ssuet.edu.pk

khi.comsats.net.pk

aurangzeb.ssuet.edu.pkNames are hierarchical, with most-

significant component on the right

Left-most component is, generally, acomputer name




Structure of DNS Names

T op Level Domains (right-most components; alsoknown as T LDs) aredefined by globalauthority Organizationsapply for names in a top-level domain: fsu.edu

macdonalds.com Organizations determine

own internal structure eng.fsu.edu

cs.purdue.edu




Root DNS Servers

com DNS servers org DNS servers edu DNS servers

poly.edu

DNS servers

umass.edu

DNS serversyahoo.comDNS servers

amazon.comDNS servers

pbs.orgDNS servers

Distributed, Hierarchical Database

Client wants IP for www.amazon.com; 1st approx: Client queries a root server to find com DNS

server Client queries com DNS server to get amazon.comDNS server

Client queries amazon.com DNS server to get IPaddress for www.amazon.com




DNS: Root name servers

contacted by local name server that can not resolve name

root name server:

contacts authoritative name server if name mapping not known

gets mapping

returns mapping to local name server

13 root nameservers worldwide

b USC-ISI Marina del Rey, CAl ICANN Los Angeles, CA

e NASA Mt View, CA

f Internet Software C. Palo Alto,CA (and 17 other locations)

i Autonomica, Stockholm (plus 3other locations)

k RIPE London (also Amsterdam,Frankfurt)

m WIDE Tokyo

a Verisign, Dulles, VAc Cogent, Herndon, VA (also Los Angeles)

d U Maryland College Park, MDg US DoD Vienna, VA

h ARL Aberdeen, MD j Verisign, ( 11 locations)




T LD and Authoritative

Servers T op-level domain (T LD) servers: responsible for com, org, net,edu, etc, and all top-level country domains uk, fr, ca, jp.

Network solutions maintains servers for com T LD

Educause for edu T LD

[2007 - TLD servers share responsibilities]

Authoritative DNS servers: organization·s DNS servers,providing authoritative hostname to IP mappings fororganization·s servers (e.g., Web and mail).

Can be maintained by organization or service provider

Local DNS servers: organization·s DNS servers located onvarious subnets to provide DNS lookups for hosts on thesubnet. May not be accessible from outside the subnet. T heirIP addresses are part of the host's network configuration(manual or DHCP).




Local Name Server

Does not strictly belong to hierarchy

Each ISP (residential ISP, company,university) has one.Also called ´default name serverµ

When a host makes a DNS query, query issent to its local DNS server

Acts as a proxy, forwards query into hierarchy.







Inverse domainCountry domains



DNS Name Resolution Example




DNS Name Resolution Example





DNS Operation




DNS: caching and updating records

once (any) name server learns a mapping, itcaches the mapping (Domain·s DNS = IP)

cache entries timeout (disappear) after sometime (usually 20 minutes)

T LD servers typically cached longer in localname servers

� Th us root name servers not often visited

update/notify mech anisms under design by IETF RFC 2136 h ttp://www.ietf.org/h tml.ch arters/dnsind-ch arter.h tml




DNS records

DNS: distributed db storing resource records (RR)

T ype=NS name is domain (e.g.

foo.com) value is IP address of

authoritative nameserver for this domain

RR format: (name, value, type,ttl)

T ype=A name is hostname

value is IP address

T ype=CNAME

name is an alias namefor some ćanonicalµ(the real) name

value is canonical name

T ype=MX value is hostname of

mailserver associated withname




DNS protocol, messages

DNS protocol : query and reply messages, both withsame message format

msg header

identification: 16 bit # forquery, reply to query usessame #

flags:

query or reply

recursion desired recursion available

reply is authoritative




DNS protocol, messages

Name, type fieldsfor a query

RRs in reponseto query

records forauthoritative servers

additional ́ helpfulµinfo that may be used



wireshark Display of DNS Response

ID is random nonce

used to authenticate

Response to Query

18



Pure P2P architecture

no always-on server

� arbitrary endsystems directly

communicate � peers are

intermittently

connected and change

IP addresses




File Distribution: Server-Client vs P2P




File distribution time: server-client




File distribution time: server-client





BitTorrent Overview

Website allowing peers to share music, video and other

media files Central server helps users find initial set of peers that have

pieces of the file

T racker server keeps track of peers possessing content of

individual files Users download the file by participating in exchange:

T hey exchange pieces that they have

for pieces that they don·t have

T herefore, for the system to work, users mus t have inc en tive to g ive

Users who just get, but do not give are called f ree r id ers

Protocol must discourage free riding



BitTorrent




BitTorrent




BitTorrent




BitTorrent




P2P: Searching for Information




P2P: centralized directory





P2P: problems with centralizeddirectory

Single point of failure

Performance bottleneck

Copyright infringement:

´targetµ of lawsuit is obvious

file transfer isdecentralized, butlocating content ishighly decentralized




P2P: decentralized directory

Each peer is either agroup leader orassigned to a groupleader.

Group leader tracksthe content in all itschildren.

Peer queries group

leader; group leadermay query other groupleaders.

ordinary ¡ ¡

r

group-leader peer

neighoring relationships

in overlay net¢

ork



Decentralized Indexing: Query Flooding




Query Flooding





P2P: more on query flooding

Pros

peers have similarresponsibilities: no

group leaders highly decentralized

no peer maintainsdirectory info

Cons

excessive querytraffic

query radius: may nothave content whenpresent

bootstrap node

maintenance of overlaynetwork




More about decentralized directory

overlay network

peers are nodes

edges between peers andtheir group leaders

edges between some pairsof group leaders

virtual neighbors

bootstrap node

connecting peer is either

assigned to a group leaderor designated as leader

advantages of approach

no centralized directoryserver location service

distributed over peers more difficult to shut

down

disadvantages of approach

bootstrap node needed

group leaders can getoverloaded



Gnutella: Peer Joining








Skype

IP telephony system

Allows users to make phone calls to Skype users

to regular phone users

Calls are routed via Skype nodes

Central login server

Super-nodes Nodes that know about location of other nodes







Skype P2P Network

Super-nodes are nodeswith powerful CPU andnetwork bandwidth

One has to qualify tobecome a super-node

Super-nodes knowabout locations ofother nodes

Ordinary nodes contact

super-nodes to placecalls




Skype Components

Skype client ² the client program used to make

phone calls Host cache ² list of other known Skype users,

maintained at each Skype client

Skype login server ² a centralized component.

Processes account information, authentication




Skype Bootstrap Proced u re

Skype client depends on host cache to find otherusers

How does a new Skype client populate its hostcache?

T here is a list of boo ts trap supern o d es hard-

coded into the Skype client Skype client contacts them to

Find the login server

Find names of other Skype peers

Skype client continues maintaining and refreshingits host cache throughout its lifetime It·s a good idea to let Skype running for a while before

making a phone call!




Skype User Searc h Proced u re

A Skype client making a phone call needs to findother users

It contacts super-nodes from its host cache,asking them to help find the user

Super-nodes return a list of nodes to contact

T he client contacts those nodes

If unsuccessful, the client asks for more nodes

Guarantees to find any user that has logged inwithin the last 72 hours

Not much specific information on Skype protocolis available«

Documents

Application Layer2