• I

Unit 1
Chapter 1 Introduction to the concept of security:
Computer S(curity
Security is the protection of assets. The tlu-ec main aspects are:
• Prewntion
• Detection
• Re-action
The need for computer security:
1. For prevention of data theft such as bank account numbers, credit card inforn1at;on, passwords, work
related documents or sheets, etc.
2. To make data remain safe and confidential.
3. To provide confidentiality which ensures that only those individuals should ever be able to view data they
an; not entitled to .
4. To provide integrity which ensures that only authorized individuals should ever be able change or modif;
information.
5. To provide availability \Vhich ensure that the data or system itself is available for use when authorized user
wants it.
6. To provide authemication which deals with the desire to ensure that an authorized individual.
7. To ·provide :-:c::-:-ep~di:it;c~. -,-, !-,;,::l:. dca!s c,,:it!·, ~:'" ability to ve11fy uial message has been sent and received
by an authorized user.

.
.· -. -·- -
.
.......
- - - ~~ - : . : _ ~ ~ - ~ _,c _=:(::-.. •'/~.. - . ;, - - - - - -- --~~. · ,..._..-.-
• 1. 3 ~fC UR ln' ;\?P ~C' .\CHE'S a

' ~' . . ' ,'!11~ .,,, :

."" ,' ." ~
' ...,

' , l. ,'l'". .,,,l}- 1._")

" t l' ', ' tt~ ll
}.. .": , , }- .;::, -> .I ..:., •.;,.. .:: , - ·. , ,,"" l "1.'·' ,,. _...,,.._, ' "'\' \' '. - .,, ~ '1\ ' ,,......,_1
~ ~ 1. . , '.. ' ! ,,ii.~ 1- ~ l.ltl't
(,\.11.,, ..
~ \..1,,.\. :" , ,,\..,\
l
\. .,.
' l'o "1' l th'l~
$h'\ll t1.•n1.pu tln~ ,,,s(.', 1..W 1, ,, . n
T .L -t .1 ... ~ .1 -\..~I.. ... , x,:_ __ , ............ ...... -. ..... ... ..

•' •
' \
•
1 .:i c-:10 • n 0 :.. t':,' • ..
::::::::-"' \1 : • tn-.' ::·1:
t o,,·:: .rct.s rc1...c •' h \ )
~ l.. • ..
.t1,,n '
1. !
l
,U'l::t \\'tn- ~. S 1..' l!l\\' .H'<..''. ,•,nu
c .... n:c_·-.:t. T Cn ,,-~1.s \1c?'a
> - l',"n1 l,1n,1
n~t; ,1$ :1
·~ ~(',:n rity p1.1fo.·y, Th(.' h)\\ '\'r
f_._-m,,·:1.:~ rc.sp1."ns i~lc :·t'T ~ni;_•r'.'il~ th,' sy~tc-1n
thi.s r~,i~t'S ~H'l inth.'l'<..'.nt
me- T('B. the" hi~h c"r the' ,1$.SULU).1..' c'. H1.' \\'c'\' t'f.
h) th1..' 1..1.l.'~,~~nin~ ,'I( (\p,'rn.tin~~
pr0~~c-m (qu:te ~1:1~ i~a:· h.") the- 1.h·ci.s:1..~ ,-~ r"'btt''-l
~l ~ f't)~~ibk , ttw ~\lrl' t.)lll h111 't~
systc>rn.s ). Ir \\'c rn:1.kc> the TCE ~'l.S $ ll'l:~ n bi~~!
h ~rLh,-~1.rc. sorrw~:.rc>. :1.nd •ilrm,,-:H'c :1.rt~ hk~h- - to be- 1:1. nih.' ' ~

The m2thc>matic-al i0mhiati1.)n fr,r tru~tt:.'d .

$\~t t'lns w-as l'lr1.. ' widt' i.1 lw . tw\,
In the yc.':n- l ~).-4 , Dm·itl Hdl t)
rebt iYely inde pen den t yt>t intc 'lTd at('d ,,·,:n-ks.
:1 kdr niq n~ fnlk tl tht' Ht.~11-
ai--.\1 Leo nard LaF adn b 0f ~IlT RE d,,yi~t'tl km is
t\\',)rtlw' l'\'H ll•n kr s,~
LaP adu b. 1no dd. In thi~ m1.."1..1t"'L :1 hi~~hh-.. tn1~ ~
\ '

k ,'t~.. ( )b_)t' d~ art..' p:,~~l\'f

de3i::;ned. ~ :S :1 ,Yl'. c:-1.·tion l't' 1.)b_kl.'t~ ,1t11.l .sul\
a~ fik~. _lll~k~. prin kr~ , d1.'.
rcpo :::it1.: ri1.". ':S 0 :· Je.stiu:iti,"n::-: 1>r 1.Lit:1, $llt'h
.'1.'~~ t'~- \,r ,hr1.' ad~ 1.)p1._'n\tllh~
Subj c-l·t::; J rc' ,11.-ci,·1.• 1.'lH: li,..>s . .:::llt'h :1~ ll~t' I'$. l'l'l..''1
1.,n h) tl \,,,. :rnh, n~~ ,)h.it'\'t~ .'
L) ll bc-lu ]r" of th1..).:::t' nst>r .: : . Snh_:1..' d .., 1.':\U.S~ in !~,1:m:,h
 · , Dorot h y Den nmg . . . .
Aro und. the san1e t ime at Purd ue Umvers1 ty was
-bas ed info rma tion flow s in
prep arin g for her doc tora te. It deal t with latHce
a partialJy orde red se t, in
co1:1puter syst ems . A mat hem atic al latti ce is
ices eith er dom inat es, is
whic~1 the rela tion ship ben veen any tvvo ve1i
raliz ed noti on of lcibels-
d_0 77:- 1nate d by or neit her. She devi sed a gene
sifie d mili tary doc um ents .
sim1 lar to the full secu rity mar king s on clas
Exa111ples of this are TOP SECRET.
's theo ry into thei r IvIITRE
\

Late r, Bell and LaP adu la inte grat ed Den ning

pute r Syst em: Uni fied
tech nica l repo rt, whi ch was title d Secu re Com
labels atta ched to objects
Exp osit ion and ]\fultics Inte rpre tatio n. Here ,
ed with in the object.
repr esen ted the sens itivi ty of data con tain
abo ut con fide ntia lity or
Inte rest ingl y, the Bell -LaP adu la mod el talk s only
the prob lem of inte grit y of
secr ecy of info rma tion . It does not talk abo ut
info rma tion .
1. 3. 2Sec urit y Mod els
t . l t its T
can t a l<:e seve ra1 appr oac11es o 1mp_emen_
Ct:) ...... ,,r~h
~~~': "~HJ
. 1011
An organ1za_
.J.
t'

mod el. Let us .su1nmarize thes e appr oach es.

JJ
1. No Sec udty
deci sion to imp leni e.nt no
In this siJ11plest case, the app roac h coul d be a
secu rity at all.
2. Sec urit y thro ugh Obs curi ty c\.\.,-,\: ,,\() w
r' )
nob ody kno ws abo ut its
In this mod el, a syst em is secu re simp ly beca ~se
k for too long , aE: ther e are
exis tenc e and con tent s. This app roac h cann ot wor
man y ways an atta cker can com e to kno w abo ut
it . .

3. Hos t Sec urit y

rced indi vidu ally . Thi s is a
In this sche me, the secu rity for each host is enfo
it can not scal e well . The
very safe app roac h, but the trou ble is that
izat ions mak es the task even
com plex ity and dive rsity of mod ern site s/or gan
hard er.
4. Net wor k Sec urit y
s grow and beco n1e n1ore
Ho st secu rity is toug h to achi eve as orga niza tion
<iiverse. In this technique, the' focus is to control netw ork access to
· a·1V1'dua1110st
ho~t? and their servi.ces, rathe r t h an 111 secun·ty. This isvati1__
a vei~,.
eff1c.1ent and sca1able mode l.
· ·
1. 3. 3 Security-Management Pract ices
Good secu rity- n1an age1 nent prac tices always talk of a secu
rity polic-y
being in place . Putti ng a secur ity polic y in place is actua lly quite
tough . A
good secur ity policy and its prope r imple ment ation go a long
way in
ensur ing adeq uate secur ity-n1 anage ment pract ices. _A good secur
ity policy
g~ne_r_ally_~akes care of four key aspec ts, as follows. - --·
• Affor dabil ity Hmv 1nuch mone y and effor t does this
secur ity
imple ment ation cost? •
• Func tiona lity Wha t is the mech anisr n of proVJding secur ity?
• Cultural Issues Does ,the policy comp lem~ nt the peop le's
expec tation s,
work ing style and beliefs?
• Lega lity · Doe:; Ll1e 1Julic.y 1neet tl1e 1egal reqtti reme nts?
Once a secur ity policy is in place, the fo110\ving point s shou ld be ensur
·~d.
(a) Rxpla natio n of the policy to all cc ·1cerned.
(b) Outli ne everybody's respo nsibi lities .
(c) Use si1nple langu age in a11 comm unica tions .
(d) Acco untab ility shoul d be estab lished .
(e) Provi de for excep tions and perio dic reviews.

■ 1.4 PRINCIPLES OF SECURITY ■

HaVJng discu ssed some of the attack s that have occur red in real
life, let us
now classify the princ iples relate d to security. This will help us unde
r$tan d
the attacks bette r, and also help us in think ing abou t the possi ble
solut ions
to tackle them . We shall take an exam ple to unde rstan d these concepts.
~)
Let us assum e that a pers on~_~ya1.1ts to send a check worth.$!P.9-to_.
a1;1,other
perso n B. Normally, what are the facto rs t}:iat A and B will think of,
in such a
case.iJ:\Vill write the check for $100 , put it insid e an envel
ope, and send it
to B.
 h .,·ccurity principles·
} · TI1e securit y principles are:
1. Confidentiality·. · The princ1'ple of con fid
I entia · on Iy sen der an d mten
· 1·1ty spec1·r,1es tnat · · ·
d e d rec1p1ents s h ou Id
( an unauthorized person is
be able to access the contents of a message. Confidentiality gets compromised if
able to access the contents of a message.
Example of compromising the Confidentiality of a message is shown in fig.

A , Secret

.... l
·"'.:l
~
() r,~ . -r· ,· ;'\;r ' · ·• ·
'Vt \,. ~ ~.J I
I :'!J

C a.\\ o. (.)c.

Here, the user of a computer A send a message to user of computer B. another user C gets access to this
message , which is not desired and therefore, defeats the pmpo se of Confidentiality. This type of attack is also
called as Interception.

2. Au then ti cation: Authentication helps to establish-proof of identities. The Authentication process ensures
that the origin of a message is correctly i Jenfffiecf -Fcr··examp)e, suppose that user C sends a message over the
int ernet to user B. howe,·er, the trouble is that user Chad posed us user A when he sent a message to user B.
l:a8'.'.' \ \''..".:~~ '..'.:e:- B !,_T').0'.' -' tlc~t tl,,, message has : ome from user C, who posing ao; u;;c, A 7 T1,i~ ,;um:epl i.;

Iu,~;u
_() shovm in fig . bdow. This type of attack is called as Fabrication.

CA L-----~
B

T
I C

3. Integrity:-when th·e contents of the message are changed after the sender sends it, but before it reaches the
intended recipient, we ~ay .that the integrity of the message is lost. For example, here user C tampers with a
message originally sent by user A, which is actually destined for user B. user C somehow manages to access
it, cha11ge its contents and send the changed message to user B. user B has no way of knowing that the
contents of the message were changed after user A had se:nt it. User A also does not .Know about this change.
This type of attack is called as Modification.
Ideal 1·oute orinessa.ge
A

Acru-"l route o±~ u:i.essage

4.A Y:iila bilit·y:

· En suring timel y and reliable access to and use of informatiori
· Lo ss of a\·ailability is the disruption qf access to or use of information
· Assures that systems work promptly and service is not denied to authorized users

·'·
_~·~:~;t_i::
· ._··,::..
~ - - - - -~--]•:__
 ,
1 11c p nn · c,p
1 l
e of n-< . , ,i/nW 1ty st ,itcs \-1I1a,, ". ..• ., 111 ( r•· 1 (i. ,; . infrJI 11 1; 1ti t1rl) ·'11<,, i\,I i
· , • •
:n11h m ·i1.cd ' pn rt ics :1 1 all t,m r•;. l·oli- , •x., 1,. du i- ' " 1h c 1111 , · 111 1r1:
. .
1. tl ;, , 1,, ".. 1•
,_11 1,1p . ', , 1111 1 !H'. ;, 1,1,: ' " ' , 11 11.
• \ C
un :1\ 1\ 1, on:.rc ·t: u s<: r .,, : ll •·1111h n 1 \7(• c 11 '-l' I 1
1 11 1. 11
, ,. ' :, , . , ·.- , . .. , ,
ns &hov.111 in Fig. l . [ i. T h b ,,·cn tld d e rC . . 1 . r i 111 i1,lc- I) (' ;1v;1i \; 1li 1l , 1 ) '· ~.. ,, Ii ,Oil .. 1: , ,, ·;
;i l I ,1 I' .
,1s i.l'l t c rt·n ptfo n .

t=i_g. 1..S: Atto, k on ovoiJobifity

5 .Accc ss Co ntr o l

' - I
I •

'T l1epr i.t1ci ple ofq.c cess contr ol c..lctc r,11i.i

1c s 1oho s h o uld b e J bic l<> ...t LLt.: !)::>
shou ld be able to .spec .if)' th;:i't user Ac.a tL• /,,,; _ !-;: :, i · ;j---.:;r..::::--:~.:::, . .. •-c
n vie"" t he 1·ccor ds in a d;..ct a b ;,. ~c.
them . 1-low ever, user B n 1igbt be all but ca n no t u pd:i i e
o wed t o
nism can be set np .to e .nsur e this. Acce ss n1ak c up cbi:c s a s well. .-\r, a cces s c o nt r ol n1e c li :·l -
c o ntrol is bP >.,dly n:bt e d to t~, o ,11·e:c_
..a.gc.1 1u.m1 a.nd nile mµna g emen l. R. c le n,1 an s: role ma 11 -
agen Jent conc en u ·.1 tc:s o n the 11 s -::r
qo what ), wher eas 1·ulc 1u.a nag·_e p1eni . focq scs o n the n ~sou, ·ces side ( w hich r 1
sic.I c: (h h1ch u s er c 2 n
sib.l e, ::ind un..ci er wha c c,i rcum st a.nce ~sou1 ·ce is acc•cs-
s). P.,ase cl on the deci s io n s ta k f.n here
·1 oa1:rL xis prert are'd , ,vl:l icl~_li.st.~ the· users , a n a c c ess co n , roi
.ag·;1;inst ..i list of it e ms the :' can ::icce
that user A .dm wrhc to 11,le :X., but c~ n ss (e . g . it can s ;1y
OQiy upda te file s Y a nd Z.) . An Acce ss
(A<;:L) is ar s'µ bsct, .o f ~n a~~ess . cc,ntr:ol• Cou trol List
m'.it:rix .
6 .Non Repu di a tion
• Nonr epudi ation preve nts either s.e nder
or recei ve r from de:1y ing a trans mitte d
• Thus , ·.vhen a mess a ge is sent, the receiv mes s age .
er can prov e th at the a llege d s ende r in
• S im il arl y, w h e n a ·mess a ge is re ce iv ed, fact se nt the mes s age .
the s ender can pro v e th a t the alle ge d r e
m essa ge ce ive r in fac t recei ved the
n:1.er,e arc s1tua uons wlH: re a user sc.n< ls
a n1es sage , and lat e r ·-on i-elu s es that
nesz age. For in.st...'\ n,c-e, user A co\ll :; he had sent th :.a
u send a fund s tr;-in sfe1· rcq\l est to
lnte, neL ~\fie r tJH: bank pe1·f onns bank B o v er the
the fund s trans fer .is per A's instr uctio ns, A: coul d cla.i n ,
hat she neve r s.ent the fun<l s trans fer
instr uct.i on to the b::in\:.! Thu s, A repu
<liat e s. or <len.ies.
*~ 'u.n B.e.rt..Q.A... oJ vi..iv....,o O,.)f~ c...lo .oo L~
~ ~~ ~ -'> oJ:::L =.dCR..A \._, (_)~ YY"LO. X vY"\. .i u
~ bcJ c:::vd~ CC4'YLp ~er
B.J ?v ..k ~ 0--t lCL cto -;) oJ::to...cJ:-_eA
°'f f efl--9. 0-n TV _9, '00-C Y-> f='-r-ei ·
~ ~..a ..J. ~ C.J' . - ) a.J::: i_o.. cleZ( - \-'\....L..a..A +a
q-b t :5_h .~ Q~ JU-½ 0-1 ce,n ''-P ~~.
- oJ,:;lQ?vC.k.Q.A- q... f:): -~_p ,O),..d d I B \J ~e,.p-") _\,G\._l~,....., -L::i
-hR ~..9 .-.\-e ~'~Vv OCl' n C..11.. - .\-e-•u2 crud .·z}.(?
\,;)C,4.....t' -H~-C:
1::J,C l.t ,-.-e h ao r...~ t-
6:t.A..cr:) ___.l..1 c l:m .L,.I,.~ 1- v-c::.,__~k.,-.
 .::.~er·f~_. 1ds transhf~r in str~ction. The principle of non-rt/iudiallon defeats ruch possibilitie! of.
oen)ing somet mg. havmg done it.
7

7.Authoriza tion
• AutJ1ori zation is a process of verifying th at a known person has the authorit y 10 perform ce rtain operation.
• AutJ10rization cannot occur without authenticati on.

-TYPES OF ATTACKS
\Ve can classify the types of att~cks on computers and network systems into two categories
for better understanding: (a) Theoretical concepts behind these attacks, and (b) Practical
approaches used by the ~ti.:atkers. Let us discuss these one-by-one .
I .S.1 Theoretical Conc(?j:>ts
A.s we have discussed earlier, the principles of security face threat from various attacks. These
.attacks arc generally classified· into four categories, as mentioned earlier. They are: - -
--.-fnte;c~pti.o n-Discussed in the context cf confidentiality, earlier.
• fa_b ricatiou-Disci1ss ed in the rnntext -of authentication, earlier.
• 'Modification-Di srnssed ih the contev:t of integrity, ~2-Y!.ie!".
• Interruptior~-Dis tussed in the context of avai/.nhility, earlier.
These attacks a re further grouped into two types: passive attic~ and acth·e attacks, as
shown in Fig. 1.6.
Let us discuss these two types of attacks now.
I • Passive attacks • C

Passive aiiacks are those, wherein the att.1cker indulges in evesdropping or monitoring of
~~~n. In otlier ,v(>ros-; the attacker auns to obtain intormauon that 1s m transit.
The term. passive indic~t~s that the afrac~er does not. ~UiITJ~E::rftmn.. any modllicattons
10 the~- In fact,_ thts 1_s ~lso why P;_~.1-~'. ~.-~!tacks ar·e har
er.ta detect. Thus, the g~eral
approach co d_eal w101 pa~s1ve attacks 1s to think about prevenuon, rather th._m detecuon or
corrective actions. ·

- - - • • - -•- -- - -n, - 1
 lntroduetJon to the Conce~u of
St~-
,u111y

Atlacks

Passive attacks
7
Active attacks

Fig, ._
1.6 Types of attacks
'rlate/ t ,
Figm-e l .7 sho ,,•s fur the
r cla ssif ica tion of passi,•e
cat ego rie s are rel eas e of att ack s int o two sub -ca
-- -- -- ~--- ~- -- -- --~.,--
me ssa ge co::1teots and tra
ffic an alµ is .
.
teg ori ~~ Th ese

Passive attacks (lnlercep

tk>n) j
[ J
Q" ele ase ol mes sag e con
tents --i
~ - - - - - - __ J
Tr.,ffic analysis

fig . l.7 Passive attQcks

Releast ofmessage cpn.len.t.s" is (lU
ite sim ple to un dc; sta nd.
me ssa ge to ou r fTie.ncl. we \'lh cn we sen d a con fid ent
cle sir e tha t on'.y she be abl ial em ail
of t'.h~ me ssa g~ rde ase cta e to ~cccss it. Oth ~rw ise , the
gii li" {.s tou i:-\~lslles'·Th-·fo-n con ten ts
~echarusms, we can pre ven lTo"ne else. Using cert.am
t release of message co1H eii:7s~Foi·e.xa-,npie; secunty
usi ng a cod e lan gu a'ge ,.so ·we can enc od e me ss~ ges
tha toh ly lhe des ire d par tie
bec am e onl y the y kno w s und ers t.,n d the con ten ts
the cod e lan gu age . Ho we ora me ssa ge,
thr ou gh , a passive att ack ver , if ma ny suc h 1111:ss
age s are pas sin g
wiifi som e sor t orp att ~I~ }h at try l<? fig ure ou t the similarities ~~-~ een the_12,1___to com e ue
er co\.lld
m:g,x.isJH.hs:r so~ ~- ~~ue's !.~gar3._~~
is taking_m_ace. Su ch ;i:u S, ~~ - com nic ati on tha t
patterns are the "ivo-;1;:inS'rtr · oL an~ l_yzing_(._~_ncoded) messages to com ewu
em 2ts
- eff ic mia/Jsis attack. ·~·· · ·
up with lik ely
l, Active otiotl<s
•-- ,;,- --:.•1.-1'..,., ---.- -- · --··
.. . ·. · - • • -· . • ...... .
·· · .... ·-- ··-- -
Unlike passiv e (1t1ftcks, ili£ .E.~
U~~~.!!11t~c_k:' .a~-~-~l_:!se~c\ on
mo
on of a raise messag~ ."''f\1e· dification of the:9r;gi_~~l _!; ~g e in
som e ma nne r, or on cre ati
- . -- .. se--attackscaili"1ol be 1ir c,·c nte d
eas ily.

_\~r~
\--~~
Ii
V
IO I Cr,-ptog-rophy ond Network Security

• b ccIe t ecte d 'Vil

However• the)' cin ·
· I, some clJort, and attempts can be made lo recover from
them. lnese attacks can _be _in _the fonu of intcnuplion, modification and fabrication.
----~~~~; ;:.~~ ___,. --- ----=·-""'·;.~-;.,-:_-:.,- -- - - - - - - - '

___;-}11'tel,-uption att.acks arc called as masquerade attacks.

~ f i c a t i o n attacks can be classified further imo replay attacks and alteration of
m.esyg-es. ·.
~-Fabricati on causes Denial Of Se.vice (DOS) attacks.
This classific.,1jon is shown 1n Fig. 1.8.

Active attacks

Interruption
(Masquerade)
Modification l raoncauon ,
(Denial Of Service-00S)J
I

Replay attacks Alleratlons

Fig. 1.8 Active attacks

(D haveM~quernde is c.,:::~~<:?_!\:J1~1,l.~ ~~!nit~~:·i_~~-~ -e!2~~t)'_pi:_:_~:.,~~~-be anothe!'~ti_LY:_As ~,·e

seen, user C rmglu pose as us~A ana send a_~ne~~-~-! t ~ ~ ~ - Os'e!.~--~~~ght be- le~
/;'')) to believe that tne message rndeect came Ii-om mer A. .
V In a :ep~a] rtl!::!5.k, ~-user captures~a seq~en~e of ~''. ~•~~s.\_or some data un!ts,_ and r~sends
. them. i;or mst:mce, suppose user M,·,~]JI Tou-ansfer some amount-to user Cs bank;,'IC:CQJ.mt.
- ~~ser,~~-~!~~~~J~~~<;~~.~~u:~!.~~~:L:~~,} ;i~~--~:: Ose~.:~ ngfil[senclanelec!f.?~~§essi~
co banO, requestmg for ~h~"J1-1_p_d_sJ,1.:a.u~(er. lJser C co"ufcl capture nus message, anasend
~"e'cood-~.P.f."'61iJfc's;i11-ie"i:o bank B. Bank B \\'OufuTi:ive·no rclea tlThl this 1s an u'ffiiuJio.n.zect
mess<ige, an.d""would~rcaCiii1.sas·a·second ;·aJ\ff-diJfertiiii ,· fuiid's tfans.fe'f reqv.esfl.foin·usei-"A,
r~,,- twice,:, once authorized. on<:e
Tnei'"efo'f'e:7.rst'!rCJ~l.•B~:J{el __tl1e ~-~!!~f!l --~f,1h,: ~~~~lld~~ tra~s_
r, through a repay auack. · .
(__i) ~!!i.E!!:....C!l!!1~!>/ 0f;1'.( i1wolvesy1111e char~ge to the '~E.lgi\~_:ll__~11_<;:~-~~$e, For instnnce, suppose
user ~s:= •~d s an ele((ronil message Trr1115E~-- ~!_()(~() tu D:S 11~:~_'/}_!>~!L!:9.b~nk B. User C might
c;~p~urc 1lii s , ;1_11d d1;111gc: it ~•.i r, :1@[1'1" S/.IJ!.)00 !.° c·.1 11t ((~ l!_ll(.___~o•e thnC!ioth the~beneliciarr
 /n:ro duc:i o n ro :he Conc epts c
( S«ur i:1 \
<1nd 1.hc :1m oun t hav e bee n
c han gcc l-in stcac .l, only one
altc rati on of 1.he mes sage . nC ihc.<c cn, ild h ;,n: ;\\ ~,, c;ru
s,-...,

0 D . J or S
~mn
--- ~.. ..-- '
· (DO S)
'J ·e1..•1u .
!'-om e se n ·1ccs , wbr cl\ (he')
k _k
alta c ·s 1;1.1 'C ~:
' are eG ~c for. For ms
pt l
t:.'lnc
O p1-cv
. .
;i[:.:.~2-..,. . ,... ,,. ,~-~--- ·'\·
c:1 t lcr,rJJJ11 ;-J 1C U5e r ~ rror n ;,cc
·1 - cl i, scr nii,,hi:'
e , ,rn 11:i; irrl ,_r~,-~_c:_ . ... __. _ 1?
.
,·ss11 :h
send
\ 0(:)
t"'.}~~y l~g:1_!1 !"~~ 1cst s-~o
- ~ - ~ - - • · ·· ·
.. · I , lt'·r· the o the r 111 q llr d:.... ____ ..
"-
so as to l1oo d .the netw ork ancl.:...
~ et;vc r ~ ~..£:..1..:<_l!].~.9!.11 ~~~~ 1. )~ ~ 9ne. ;i .~ _ ._ succ ess 1n n .
- . - · - ·· -- ----- - -------
den y 01.he1· l egtt lln;r lc .
ll SC JS ,in :itc t s~ 10 ili t: 11 1w m Y ·
1. 5.1 The Prac tica l Sid -·--- -- ··· ··- -··· ···· ··- .. - -· i:
e of Att. acks
The a u acks disc ussc ~\ tarli c1·
can c o me in a 1~t1m_bc1· of rn
cbs silie d in1.o two bro a9 cate n11 s in rc:il lire. The y c:,n
gori es: app lic;, tion -lc,·c l :1 lL~c be:
sho" • n in Fig. 1.9. ' ks :111 cl ;i-i:·i:,~:;,d,-lc\"C:I ,rt t: ,cks
. :rs

Secu rjty·;, ttack s tn pract ice

Appr. calio n·t,;,v.i lt-aU.a<;b

~Jetw orl< l o·,cl a11a&.s

Fig. i . 9' Pract ical side o( anod es

Let us. dis cu~s thes e new .
• i\..ppi. i.c:!.tlq:::. l::"":.:.! ::t+..::..=1...-:-:
'I -iha t .01e atta t:ke t· attemp,ts··.
T'u...::=;~ ~a.,,~ks · h ,,p~en at ;.u1
app lica 1ion levc i in th<.: ='e,
;, 0
fo. acce ss, mo d~ or pre veo ·~c ii.s io i',~·ro nn:a inn cf"
•..; ~
·par ncu lar .ippU Cltio i:i,,. or the
app lrca tion°"'.l5t'lf. ExampTes··o
so'i",c:on,e s c1;e·c hJ. catd. _i~-o rtrta r1his ;ire " :·m g to olTTaln
nit:, ·,ge to clrn ng e -·tl1e· -atr1:oun n 'o n on lire Cnte rnet . 01· chan gT1 1g t!Tc: " tc'lii le"T ffso f;1
!- ')1.r a· ti a.T".s:tr.rr,:,rr,-e1.c. ·- · ...
• N et._, r.·t ~!: Yel a'.t! f,¢k s~)l ie.se. ·· · · · · · ·· ··- · ·-- ·--·- •·-
att;i ~~s 1,g1.:.,n~e'fiil ly a11 nat r e
netw o rK Gv......a.J:Bim.ber ofp.._o:ss1 duc ing the cap abil ities !tl.
bls:•n1·~ -;. l ·11ese-rrccitk s · ge'n-e1·a '-'
eith er slow dmv n:, or coni p"le ll y m:ik c a .n atte •~E l to
te.),t ._~r ing tr, h ah , a com pute
j ·am orna tic.a lly can Iead .to app r tH!tw ork :-.'r-f nte that this
li-c.a tioo·leveT'a ttac b. "bec ause
gaiQ· .:icces.s ~o a :pen•io~k.. µsµ- onc e som eon e i~ ;,blc to
t iofo n:iµ tio4 .. 01-t i~-g lia\rbi;'.·
The se two t}',P.~ of~~ia:4u:-c an-,
all:y ,she is able to. acce ss/m odif
· .
y at leas t rnm e s en~ itive

:b
next ,. ,Ve ,l(ill, ri_o,~ '~ ~ 'ttre.~~-. e:-attempti;~fby 1.1si ng v=iD \.IS mec han :.sm s, as disc usse d
\ act6 $S app lica tion as ·we
t~c ~s ·intp the· aho ve two e:tte god
es, sinc e they can
I ll as nitWorl;, le•l <:ls. . sp an

.I
One can
.. "I cpl-JGU~ ~ - laun ch an -:ipp licaf i.o n-le vel att,rc ;k
v .~U J ;,. 01· a netw ork le vel .it.ta ck usin g
a viru s .
\ '1\R u.1 Co- .n ~ .. 4 ~ -.
I
! D Do.:vna..n..t ~ e •.- v ~ lo \.d. ,u . I-t
aJ ?~ e paeo-.t.l(!;)-n ·.- ,,;L9--U..l cop1JvO_
·t\-ee.l.h i! eet .d- , Ccif-'d
tv\o ...e
c~ ~ roea1:v-t:)
aJ " T ~ ~ ~- -A- ~ \Yl.D .....n -\.- \} \...ID_,u> \___.,._ ta - \-i:\._-;_A J
~ ~ 0-r> ~ ~~ Li: \.....OC..\.D ln. .-d : \ . o._'G_
r• ~J G, c~ .D r'- '? ~e ~- A -~ l,O '.P ~ Cl t \) \.,~I...A., t . .L-~ C..=...
~ ~ . °'- . d.a..otn.Jc.
D...,vc-
12:f. pv., ~\ 'yu t,U J o., \J '. o , • •
L) ·p~ v~ u : ~ ,· Q.:ti.a.c.J'1 ~

l
•r
} l L ~ r " - t g . Lev --er '"l \Jr.,s Q,l~ --io -<b<!e
"-.n .~(:_,o:"-d , b""--lQ ~o .e.
1..,J:(W ,
@~.hn.... ·-Q..~e, -~ -I-ct ~a .,d , l..t oe. \ ,l o.p.,.\..\10..d b
. \=f-==·=·=·· ==:.:::::.:.=.... ___ --~- -~ -- --- ..
---- ---·
II
1'

l( I
12 Gy p t o i;ropl,y and Nc1wor k 5ccuri ry

I
J>r or r~ 111 i

Add x lo y Dl: IC: IC ;-,It l1!c,.s

Por1orm Prinl-Job s~noj a copy o l
Perform Cloi;&-Jo b rr,:,•s~Jl IC" ;., II
wsing U,1~ u~cr ·!i
End
isddr~ss t,ook
Rcturri

,(a) Ori~inal :i::leari ce>do (b) Virus inlec:ed code (c) Virus code

Fig. 1.1 0 Virus

Vii·u.s .e s ca1J, ;iJso b ,e 3i·rgger c_d_1J) '. _specifi c ,cvent

s (e .g. a ,·i~·us could :iutom aticall y execu te
a t 12 PM ev<:1:y d~1y)., lJsual. ly vu:u.se s cause daro
a ge to co1'1pt lle1· ;!ncl nctwo1· k system s to the
"'. Xtent ~h~tt it c.;in l,e repa Lrt:d, ass uming Lhat
t11e org:ini_ ;,.ation deploy s good bar:ku
1·ecov ery proced ure,;;. p and

'1to u
2.. Worm
., Sin1ib r. in ,conc~ pt to a vi~1s, a """nn is acrnal
ly differe nt in imple menta tion . A vir.1s
1.nodiu es a proj;,rra111 ~Le. it .a ttache s itself to the
progra m under attack ). A w01·m, Fiow"0.·er,
-Boes n<)~ n1od{fy_ a pn~gy am . _I nstt!ad , it replica tes
Fig._·n -,:-:-r ~£!i .£.~ QW S ~fr)n4 cfi1Jia cul1:Ji'.na(cly tself ag;iin a :1cl again~- Th.is is shown in
the-e:0·1~1-e-ii~:r.~!:_l.~~E~~:~!. on
which tl\e"\,\~O,!Et-· esi$~-~:;J:>e com~~ i"s.'tJ' ,]!~~fi E~co·f
i!!?i1( to ,l h;~J~. Thus, Lhe basic purpo se
of a ,,;,qrn:,,,; ~~1.l,q; '-1;s:-i l3~et>~ fr_qn\ tfi•H q.f ;:t
V1n.1s. A worm attac:.-. a:uem pts to make the
c·o mpute r .~r the rietwb :f.k ·.ti'f1d'<:r aflli2k unusa bie
: . ·by eating all its rc:sour cc:s.
. . ,-·

3. Troj_Qn hP:tl,¢
- ·-A·'Il'~ ja1i lib'r.s~ t-s .:,; -h .t'dden,.piece: ;,p(cod e, like a virus.
Howev er, the purpo se ofa Trojan horse
is di!Ie.re:ne,. ''The'ti; i:liiii p'.u,rpose ·o'f a i.•iru:; is io make
some son o'T mo<lll tcat1on s to the target
comp uter ar net\yo rk', '\\'he·r ea,s a ·- r-roj.in horse ·
attemp ts lO rc\·eal confid ential inform ation
to au in_ra<;ke1: "I:lre: na m'e (Tr6ja n hgr.se) is di.le kl. the G'i=eek-s0Td1ers,·,.;.11ol'iid
·-·-......- . . ~--- ....... -..~ Tomi c:, a large
- .• -- ,.,._,. ~ ,. ,. ...

Goel -

_,,------::--~~ - -
0 - _,, ,1-~::s:,, ., ,.

i-( . . ....
,,
::.---
________ ,._.., .; - -~ - ---·
-- ~-- - ..... ----- .... ,- .
-- .--·-~--- ---- - ."- \r
 - I

P.erfom>
Replicate re60li'roo-eatlog
Itse lf tai;~.s. bu1 no
destruction

Rep lica te Perform

reso&rce~auhg ~ Peri onn
reGOOrce-oating

<$
Jtse lf :tasks.ti~ilC! taslcs but no
Pe,:for.m
res ou~ a~ g
·task s; 'but .f'iO.
0q .
· des.&uc
. iloti ·

Pe11crrn
,,, .d&sWction

di3stndlcin 1/i" res ou ~ti ng Perform

ia~ .bu ino
~-
re S-O U~ ating

worrn .cc:,,da
de~ _., task sJ but no
oostruc:tlon

-:Pe'ifolin-
~ij~Th:Eif.ta'.urig ·-
~~i15.!:l,l'no
~f!tµ_cqo_n ,.
..
'
}:i_g. I! 11 Wo rm
hcl low hor s.!, wh icb was- pul fo~
:t'b y ·Tr oy ·cil ize.n s, u11 .aw are
sol die rs cnt e.~ d the city 6f'T of its con tent s. On ce the Gre
:icry-; the y op: :ne d the gat es ek
1n ::;-:m . n:...
-, 1~ -.:r......; .•,.0.. - •. for th~ res t of Gre ek sol die
'
_, .. . i:;::r,y4_{.1 ·f18 ~-~ cg_mo •. ' s1·1 enu
··
• . . n1e' co d e fio r a 1 rs.
attaclti):1g itse lf't o it. \ .Y:bt~n t:1\'<; y sll- m -og m .
' :Q$'e ;r· ~tte t,i; the ~se r id .ind scr een b y
~P-t.4'.I::e f ~f ~ _<;i,l!t:ii,ls,, A~~ ~~:P pas swo rd, the Tro jan hor
:~9}1.ili}fmfo.tJl!a;tj.pJ:?. ~o the ~tta cke r se -
yje u 's er ,,·ho -ha ~·c rit~ .ed 'ilit !:id wit hom the kno wle dge of
irl an;r:I, pas swo rd to/ gaj,H a.c¢e.·w :to.i'n: g;pa:s:s.;;.,' orc kT )ie atta eke r can
the
0·ili'~1
·s'YsJ•mi :Tr us is sho wn in Fig . n1.1me 2.
.rri ly use the use ~

. .
_ ___
l -~A...._....._ Sp.et;Ht'c;
.
....,.,A
_,,_ ~~ -
~- -
011. tp.e lnte□I.et, c<?mpµt.ers ~

e}s.cha.nge·me ssa ~s wit h e:id

~. likcf a pos tal envdopeco1\fail1..{ the fo,· m of sm all gro ups
:of dat a, c.1l1ed as j5:1cti:c;=~- A paeke , oth er in
:an d the acld rcss -ing in.f od·n
at1 0n. Att ack ers targ et the se inc acniaTc.fo"t,1.-;:ioc~n. ·
sot ir~ it:o mp u te1.:..!Q..JJ.M:,..,~ pac ket s, a.s the y tr~1 vel fro
m, i~r un p.u .t.e .r..1iyc{ ru~_Jf m
mafo .forms~ (a.) }>a_<:~e~.s~n~~~~~i '~s o.cif:i'!~L --r1i ese a nacTs 'T:'ik~·-ctl11.:
,,.o
pr-o toc ol l_J sed ~~ u''i :m ,'i: o ca!l~~ a~_~n<>opi~~? and (b}..~2-
~ ~~ ~o o~ ~~: S i nce t~1c
the se two
1-. ~¥ J-CU1 'JS, ·c,Uie<l
at1;ac.ks are ; (a) IP·:s.nllfi:ng, and as
G1tei-net Pro toc ol (IP ).
Let us dils alss the se ~vo atta cks (b) 11> spo ofi ng.. Th e me ani ng oth er nam es lor
~ rem ain s the s;im t.•.
(a) Pa elt ct~ p.z !.. _ P.;:i~k:et shl .,
ffii !if i::s .a p:as si\'e atta ck on an
attacke_snee d JJ.Qt hfj acF a con ven ong oin g CO!)':':!~"!!!Q.I)_. An
~
.ati Cln , bu ~~ ad , ~--~ff~!i1n1
pad :.c1s ~ __!: h~ j~~:.£1.;. Cle
:arl·y. tn p-re ven t cm ~~u ack cr Ii-o m
·
r w• ·~§~:~r~~ (i.e . .
mi/ /)
1nf orm auo n tha t 1s l?4'SSJ.rJg sui ~ln g p ~1ck c~. the
ncc·d s lo be pro tec ted m ~om
two lev els: (D TI~ c dnt a tba i is. c- way s. 'll1 1s c~,n be drn w ;11
u:a vel ing can ·be cnc udc d
ll',m sn1 i:;;s-io11 link . itse lf can in sum e way s. t.w (ii) Th e
he enc ode d . lo rc:,L1...La. p:,c k1·1
l~~..i_l,...Ul..l.he...fii-:.st:p.b1cc. Th . the.: :,1t;1 ckl· r son wh ow 1w« ·ds
e sirnplc?t-~:;y lo 1.\11 this is"
wh ich the traf lic goe s thr oug lC.l l:lll\·l·1:.:;r ·,i·u m11>·1,1i.: ,. ,·i.1
h. Usu ally , this is a rn11t1: r.
l·! ,,,,e\f:•1-. rou lt·i ·s ar(• hig hly

- - ----- ---- ------·--- ---- ---- ---- ----·-- -----------------·•---· -

l
i
1I

I
r 21 I Cryt,toerophy and Network S,curity

pro tec ted reso urc

es. The refo re, ;in atta cke r mjg
atta ck a l~s -pr ote ctc d com put er ht not be ;iblc w atta ck it, and inst
on t.h'e ead ,
(b) Pac sam e path .
ket SEODhng: 1n thi_s le~ m\q ue,
sou rce add ress . \~e n thJs hap an atta cke r ;_e!!~s __ll?~~c~__!'!.Lt,h...
pen s, the rece 1ve i71. e . tnc par i!n inc oq:~ L
pac kets con tain ing a Ials c sotn 'tt!~;tdt lres s}'" '::,. ty who rece ives these
_!.~!$.~_cl ~?. .i\'. ~J:. \J;! ..1~ .L!! .;:l lies bac
a_ddi •~s (cal led as spo ofe d add .~ to tbis
thre e pos ~bi e case s: ress ), anti not to the atta cke i:..
T.h.is can lead to
·-~ -·-·-~---- ~ -- - ':":, · ,
(i) The atta cke r.ca n inte rce pt
the rep ly- If the atta cke r is betw
~ d .tl\e forg ed.s our ce, the ,ma een the des tina tion
cke r can see the r eply and use
, for hijaciting attacks. tha t inf. onn atio n
·
~ a t t a c k e r ue·e d not sec
tbc rep ly- If the atta cke r's inte
Of Ser vice (DO S) atta ck, the ntio n was -a Den ial
atta cke r nee ~ not bot her abo ut
(iii) Tl.le fl.ttaclsei- doe s not wan the rep ly.
t the rep ly- Th e att. ,cke r cou ld
,......--r he hos~, so it may put tha t hos sim ply be angry wit,h
t's ,ld<lress as the forg td sou rce add
the pac ket to the <lesr.ination . ress and sen d
The atta cke r doe s nol wan t a
des( im1 tion , as it wan ts the hos rep ly from the
t with the forg ed ,1c.Jdres s to rece
con fuse d. ive ic and get
p hQ; Y" ~ \.lij 0
~ • Ano the r 3 H-<lC~ - Whi.ch ~im ilar is
·
to thes e atta cb. is_the-J?1'!S spQ
usm g the Do ma m Nam e Sys tem ~:fin g~;i rt:ad .. A...s ~._,~ l, r.::_,:,,·,
j (DNS) , peo ple can 1dcnt1f-y \•Vebsn
nam es (suc h ::is· ,vw \\•.ya11.0.o.co m), es wllh hum an- rea dab le
and com put ers can con tinu e to trea
(suc h as l20 .10 ;8L 6'7). For this t the m as l P add ress es
, a spe cial sen•er com put er call
the map pin L , b.e ttVe er.. dom ah ed as a DN S serv er maint..aim,
n a mes and the corri:spondi.111:,
cou ld be loca l ' d ar,ywbc rc. Usu IP add ress es. The DN S serv er
ally, it is wi1J1 the Inte rne t Se!, 'icc Pro
Wit h this bac kgr oun d, the ~NS 1 vid er (lSP ) of the use rs.
sr,_9,~ ~1g_~~ S~2:.~:J'...S" ~ ~[o_Up.ws.
1. Sup pos e tJra nhc re isam erc
l1a nt {Bo who 3e site's dom ain .
-~1..1_!:_}-!'_..i!.llil.t.~~S.~is, X.QJL.10.~C CT uerb),cfo na1·n e is www.b ob.c om, and
~ s is.. mru n~afri,ed n:,..in:e' 1iN·s ~;;_·t~);Tor.·1~0G ·r11 airi.ne--
.)S foll.o,vs: r>""NS
...... ·- - - - .
ww w.b ob.c;:om
100 . 10.1 0.20
2. The atta cke r (Say Trn dy) · man
age s to hac k and repl ace the l P
9.~ !J..(sa y~ 0.2 0.2 0.2_2) in, E~~-P~ ~-S add rc_s~ 01· Bob wit11 her
The refo re, ilie l;:~1$"~en•c r rnaJ
-
~ l".~E .!ll~
Jn·t a1l-iecl . bythe· IS(>or·,i · l1Ser~ay"A1ice.
ntam ecfb )' the ISP of Ahc_c 11u1,·
ww--w.bo b.co m h:1s tht' foll owm g ent ry:
3 . ,vh en Alic e wa1its i.o com mu
l 00. 20. 20. 20
nic ate with Bob 's site , her We b
~ ~;c; J_!la.i nrn.ine.tl by her ISP for nob's lP :1e\d bro1\"ser qu<:rics the DN S
ww, v.bob.co m). Alice g~_ts'll!I.fi! r<ii:s, p1:7~0 c1~1 i_( ii ihc cloni.-lil1_ii~!
pT~s.
t !.l_! \1.e .
f . Now, Alic e sta-1 ·t.sc omm unic atin
:.
~ ·o·:e·
. ·-rh1~i.ly's) l I~--.t"dcln:ss . whi ch is I oo.~~.:~.9_19-
·._j Bob l
g with Tru dy, beli evin g tha t she
is com mu nic atin g with
Suc h :itta c-k s of ON S spo.9.fing.ar
e qui te con irno n, and cau se a lot
. :i.rtac.ke r (T1·uc.ly) doe s not have of ha voe . Eve n wor se, the· '
to listen lO 1!11: COll\'Crs;Hinn
t)J1 the wi r t''. She has
able co hac k the DN S serv er ln simply be
or
t.lic ISP and repbcc:: ;1 sinf,\k 1l'
;Hld n:~s with her own !
A pro toco l c:i.lled as DN SSc c (Se
cur e DN S) is bein g 11st'd to thll
unf ortu nat ely 1it is not wid ely 'arl s111: h aua cks . How eve r,
us\.'cl .

...,._,.._.,...~¥.... ~· ·- ...- ... ..... ....... _ __ __

~ ~
_ --- .
.. _...,....
- ··- --~ ,.,,- ·--- ·-
--~--~. --. -- .,,_-- - -~ .... --·· ··-·
f.r • ·- ...........
·•- - - ..
, ~-
- - - - - - --
- - ---·
-
~~ -
- _,,____
.
2. Phishing
P111s)1.1ng ha s becon1e a .
bi g pr o b1 en1 111 .
re ce n t ti1nes · ln 20 04 th e est:i..
1osses due to pl-n.sh1n .
° w er e to t11e h1ne O~f U~D 137 m·1. 1h. on' accoc.1~.
Tovver Grm.11). Attacl,eb f al w b ··t '"' 11\
1 ' .
sites. lt is quite silnple rs se t up ,e ~ e si es, Y !=~-= o 1< hk e re a1 ..,'vv
hi ch lo ,. ,
to-a o so, since cr ea tin g --,
-
W_eb pa ge s 1nvolves 1 . . 1
1...
SU11ple technologies
Sheets), etc. Le ar ni ng
su ch as HT11L, Java
Script,_ C~S (~asca_d
:e
at lv e {.
an d us in g th es e te ch no 1ng St y e
attacker's n1odus op er an lo gi es 1s qu1te si m pl
di works as fol1ows. e. T he
• T he attacker decides
to create hi s/ he r ov
icfei1tical t o·a-·re af v; re b·--~ite·;-Fo r ex m W eb si te , w hi ch lo ok s ve ry
a1nple, th e -·-atfadzei:··
CIBnanl2"sWeb-s1te~Tl1e ·c an cl on e
cloni~g is so clever th at
be ab le to di st in gu is h th e hu n1 an eye will no
si te . ~ -"· ...be hv ee n th e re al (C1tilian t
k:;s" fa nd fak~ · (affaclz
• T1{e at ta ck er ca n us
_- ·____________ ,,_,, ____,__ ... er's)
e
'\Te ill us tr at e"t E em os n1any te ch ni qu es to at ta ck th e ba nk 's
r·ec-m m m·toiie-helow. ·-·- cu st om er s.
·· - - -- -- --- ----· ~
Th e at tu ck er se nd s an
e1nail to th e le iit i1 na te
~1:·:-~ lj.tself a~pear·~
at ta ck er e).._rpl01ts th e~
to ·1~.2~ve' COD}LD;_O:g.~_!he
cu st om er s of th e ba nk
~ k .-·F ol =e ns ur in g
. T he
~a ,l sy s~ ~n _to su gg es th is , th e
Eon1e ba nk of fi da f (e t th at th e se nd er of th
.g. ac co un tm an ag er @ e em ai l'i s
·w ar ns th e us er th at ci tib an k. co m ). Th is fa
tl-1ere -h as be en s·01ne ke em ai l
co-inputer sy's te m s ,and
cu st om er s, or ve rif y tl1
th at th e ~~!1i-Y1ii1ti.\Q
so rt 9f at ta ck on C
_:bsu i.n ~~ -p as sw or ds
iti ba nk 's
eit ex is tin g PI N to all it s
is- askect.-toV1sifa lJ1tt-menuoned'"'1n::-s,tEete c.san1
Fo r th is pu rp os e, th e
cu s-t om er
shown in Fig. 1. 19 . e en1ail. Th is is co nc
· ep tu al ly
0
f Attacko,r
VlcUm 7

. . . '

f!i,e~~t : ~~~r :. ·. I
: ;f .. ~' . ::·:, ·:·;: •· . ·., ·, .. ~ .

0 Fig . 1.1 9 At tac ke r $e nd s

a for ge d em ail to the inn
oc en t vic tim (cu sto me r)
• Vvhen fn e cu sto me r (i.e.
th e victin1) in no ce nt ly click
in tn e e111aiT,neJshe i's tak s on th e UR L sp ec ifi ed
en to th e affacker's site,, an
or i~ al site. Th er e, ·111e--E cf rio fU ie ha nf s
ustom er°-·is - 1f ro mp ted to en
in fo rm ati on , su ch ~§ __h.i.~. te r·· co nf id en tia l
Llierpassword or PIN. S_in ce
sit e loo.ks exactly lik e the- th e att ac ke r's fake
in fo rm ati on . Th e ·.affacRef'
0-1 .Tginal barik sit e, th e cu sto me r p1:ovid-es th is
g1aa.lf.accepts tfo s i1i.fonn
Thcinl<J1ou- to the unsuspe-~ atiori an d disp1ays a
ting victim. In th e me an wh
nm"'.: uses th~.-~ 0t im1s passw ile 1 . th_? att_~~J;er
ca n p-er fo nn an y tra ns ac tio
or ·~-~T "X!N .JS UJ QG gss -Ihe.];:aJik;s. _real sit e an d
~ ----- - • --, -•--•• •,,,._..• ••-••-,,•-P•-
n as if he /sh e is th e victim!
-•,- v_r,.. ..... .--- -•-- ---
•... •~•·- -- - •,,., • 0 0
• • - • • - • • - , , _,..,.. . ~•o •- ••·

'
• I
 r-
USER-AUTHENTICATION
/
MECHANISMS
■ 7. 1 INTRODUCTION I
One of the key aspects of cry
ptography an d ne tw or k/I nte rn
authentication. Authenticatio et sec uri ty is
n helps establish tru st by
particular user/systen1. Authe ide nti f;i ng the
- . -· --
--.:-- :: - - .. -·- - .. ntication ens-ares that the -cl
who he /sh e clain1s J.o be. Th
.
'-·.
. aim an t is really
.

~u ·
H
c:1. Lihenb" hn n n,o <>h !ln i~n ,Q
f'':l
is
.l'-'""" -.&.'-'J r.....4 ·~--
cbap~er tlis::ticises
.._.. .-_.. ....... _
the va rio us asp ec ts cf
........... ...,.

-.
I
:

, ----. Th ere are n1any ways to -

authenticate a user. Traditiona

~--1 ) pa~§.W<?r~s ha
ve b1~en used. But there ·are · lly, us er ids an d'-
- n1echamsm. ·Passwords can ma ny sec uri ty conce.r:1s -in thi s
travel in clear text or can be sto
on the server, bo th of which are red ir clear tex t
dangerous propositions. M od ern
ba sed authentic2.tion techniques pa ssw ord -
use alternatives as encrypting pa
using something derived from the ssw ord s, or
passwords in ord er to pro tec t the
@ Authentication tokens add randomness to the m.
an d make it far mo re secure. Th password-based me ch an ism ,
the tokens. Authentication tok is -
~n echanism requires the u~e
ens m~e quife --p.o pu iir"-i'n-app ~:_! o~poss.ess
r~, dem·a1id li1gh security. ~ ~ ""'. ~
\ ~ · Certificate-based authentica J
--feJb .
tion ha s emei·ged as a mo de rn
Bcations tha t

mech anism, thanlcs to the em au the nti ca tio n

ergence of the PI<I technology
quite strong, if implanted co . This is also ID
rrectly. Smart cards can also
conjunction with this technolo be used in ~
gy. Sn1a:rt cards facilitate cry
operations inside the card, 1naki ptographic
ng -the whole proce~s a lot more
,
rehab.le. secure and
f

Biometrics is also .getting ~ lot

of attention these days, and .js
· human biolo_gical _____ characteris ba sed on
tics. However, it has still
con-ijJletcly. ---· . not -- matured

. 'I
 Thi~ chapter examines a11 these mechani sms of authentication jn great detail.
Jt d1scusses th e advantages and drawbacks of each one of th em. The chapter
I ~h en conclud ~s with the coverage of Kerb eros, a single sign on m echanis m
1mplemented m many real-life systems.

I 7. 2 AUTHENTICATION BASICS I
1V7w are you? This is a question that we ask evtryday and get asked. It has a
1ot of importan ce and significance in the world of cryptogra phy. As we
studied earlier, the whole concept of authenhc ation is based on determin ing
wl~ o an individua l user is, before allovling the user to go ahead and perform
actual business transacti ons using the system.
Authenti cation can be defined as determin ing an identity to the required
level of assurance-. Authentic ation is the first step in any cryptogra phic
solution. 'Ne say this because unless we know who is communi cating, there is
no point in encryptin g what is 1:-eing communi cated . As we know, the whole
purpose of encryJtio n is to steure commm1ication between two or more
p2.rtie~. lhiless ·we- c..i·c 2.b:;clutcly .s·uTe that the parties really are what they
claim to be, there is no point :n encryptin g the informati on flowing between
them. Otherwise, there is a chance th::i.t an unauthor ized user can access t}, e;.
info®ati on. I.n cryptogra phic tern1s, we can put this in other words: there is
·no use of encryption without authentic ation.
"'\11/e see authentic ation checks many times every day. We are required. to wear
and produce our identity cards at work, whenever demande d. To use our
ATM card, we must make use of the card as well as the PIN. n-1any such
examples can be given.
The w}-iole idea of authentic ation is based on secrets. Most likely, the entity
being authentic ated and the authentic ator both share the same secret (e.g.
the PIN in the ATM example). Another variation of this techniqu e is the case
where the entity being authentic ated knows a secret, and the authentic ator
knows a value that is derived from the secret. We shall study this during the
course of this chapter.

I 7. 3 PASSWORDS I
7.3.1 lntrod~ct ion

Passwor ds are the most common form of authentic ation. A password is a

sb-ing of alph abets, nun1bers and special characters, whic
h ~s supp ui
kno-wn only to the entity (usu ally a pers on) that is bein
g auth enti cate d.
are great n1yth s abou t passwords. l' eop1 e belie ve t_hat the
u se ?f p assvm:-1_
the simplest and th e least expe nsiv e c1.uthenticat10n mec
h an1 sm, b ecau sP. .
does not requ ire any sp ecial hard war e or soft ware supp
ort. H owe ver , as W t.
sh all see, this is quit e vn:on g a p erce ption !

7.3. 2 Clear-Text Password

1.Its Working
This is the simp lest pass word -bas ed auth enti catio
n mec hani sm . Usu ally,
every user in the syste rn is assig ned a user id and an
initi al pass wor d. The
user chan ges the pass word perio dica lly for secu rity reas
ons. The pass wor d is
stor ed in clea r te::\.'t in the user database agai nst the user
id on the serv er. The
auth entic ation mec hani sm vwrk s as desc ribe d belo w.
Step 1: Pro mpt for Use r Id and Pas s,vo rd Dur
ing auth enti cati on , th e
appl icati on send s a scre en to the user , prom ptin g
for the user id an d
pass wurd . This is shovm in Fig. 7.1.
Step 2: Use r Ent ers Use r Id and Pass -..vo rd The
user ente rs his / her id 1
and pass word , and pres ses the OK (or an equiv1.:ent)
butt on. This caus es th e
user id and pass wor d to trav el in clear text to the serv
er. This is shown -in Fi~.
7.2. . L

Login Screen
U~ed d: -
---
Password:
----

Fig. 7 .1 Prompt for user id and password

 Login request

Id= atul •• • ·~· - r • • • • •• \ .,.

Password = aprll .. .... ... -· . ·.,

Fig. 7. 2 User id and password travel in

clear tex t to the server

Ste p 3: Us er Id an d Pa ss, vor d Va

lid ati on Th e server consults the use r
dat aba se to see if this par ticu lar use
r id and pas sw ord com bin atio n exists
the re. Usually, thi s is the job of a use
r-a uth ent ica tor pro gra m, as sho wn ·
in Fig. 7.3. This is a pro gra m tha t tak
es a use r id and password, checks it
aga ins t the use r dat aba se, and retu
rns the res ult of the aut hen tica tion
(success or failure). Of course, the re are
ma ny ways to do this .

Id:;. atul
Password= april

User data bas e

Fig. 7. 3 User authenticator checks the
user id and password against tne user
database

Ste p 4: Au the nti cat ion Re sul t Depen

ding on the success or failure of the
validation of the use r id and the passwo
rd,. the use r-a uth ent ica tor pro gra m
ret urn s an app rop riat e res ult bac k to
the server. This is sho wn in Fig. 7-4.
He~e, yve assu.me tha t the use r was aut hen
tica ted successfully.
Ste p 5: Inf orm Us er Ac cor din
gly Depending on the outcome
(success/failure), the server sen ds bac
k an app rop riat e screen to the user. If
the user authentication was successful,
the server typically sen ds a me nu of
options for the user, which lists the act
ions the use r is allowed to perform. If

- -- - - - - -- -·- ------ ·---~...---.-- --~ -= - . , .

:.:. ,. . ~. ~,.-..-,-~-~~- r ~
. ___-___ ~::::::::::....;. ;__.:.:~;_,.·:-.:;::::::~±·~~ _:::_-..:,::::·.---::":'.".::-:-:-;: ::.::·.
:._: ::;::- --~ J
 •
User Autl1cnticnt/011 and Ktrlx:ros

< 3scr aulhcntlcalor program I

Success

User database

• F.1g..
L 7 4 User a11t/1e.nticr.lor program rel unrs asuccess orJai111re message lo lite scrocr

Login successful

i- Fig. 7.5 Server rdums asuccess or failure result back lo t1ieusa

'P~oblern LO~ ---t. h-e Scl-i t-m e ~
• Problem 1. - Database con!ams pnsswords in clcay text
r-irstly, the user..dat;iba,;c contains user ids .ind passwords in cle_nr_text. Therefore, if an
n_!!ackcr
succeeds in obtaining an access to tlie database, the whole list of user ids nnd r,assword
s ls nvnilable to
lhe alli!Cker. Consequently, it is advised-thnl. the passwords should" not ho
stored in clear te."<\ in the
database. Instead, they should firsl be encrypted and then stored fu' the· database. Wheneve
r a csc·r
anempts to log on, on·the server side, the user's pilssvtord llioutd·firs[l5c cncryptccf, compare
d with the
encrypted password in the d~1abasc and depending on whether they match or not, a decii.-ion
should be
taken. This ls shown in Fig. 7.6.
• l'roblcin 2 - Password tt::iYcls in clear text from lhe user's computer to the scn·cr
Even if we slorc encrypted passwords in the dntnbasc, the password would travel in clear
text from
Lhe user 10 the server. Tiicrcforc, if an attncker breaks into th1f
communica1ion link between the user's
computer and the. server, tl1c nt1t1ckcr CI.Ul cnsily ofaiiYii the clenr text pnssword. We sh11.U
~hid) hM, tftis
~~ tt1 tho n~t~o n.

7 ~melhing Derived Fr~ss word s

lntroducli~ i ~ alioil from the bas!~ d-bnscd aulhentlcatlon ls not 10
use 1he password
i1,e lf, but 10 use :;omething,l,!1_1 is clerlved from the pnssw . nt ls, lnsetnd of storing
tho pnssw(l~
:is i1 is. or in an encrypted formal, we can run some nlr,orithm on the pnss\vord 'nlld store
the output Clf

- - - - - -- - --------- . ---
 ·_. '

'

'
Step 1: The u~r'a Id end passwdrd travol
lo the servor In cJ~ar text. The ~erver
enetyPts the password using a password-enc.ry
ptlon program,

, , Login te(tuesl

Id= alul
Password: aprll

step 2: The server's useHiUthetlticalion

program MW checks
end ypl ed password against the databas Ule user id and the
e.

\d ;:: atul
Pa$Sword tt' [,,.9%-·

(_ . lt)gtn sutcesstul
\
User -database

L ~,tep 3: The $erv ~t s~nds bat~ an a_

ppro pnat·e menu or -~e~sa:ge to the w;er.

Logln StJccessfu-1

Fig. 7. 6 Encrypting passwords before

-they are stared and verified ·

Sev eral req uir em ent s nee d to be

m et to ens ure that thi s sch em e
cor rectly: \\'l) t1.s
 The Prob lem with pass word s 8
,.,a __ u_d ~ n/w
)0cu'U 1.AY' ~ (\ ~ a➔ °'-f-pJJ> , ~
•-

0
oC
, ~ --l'°.o -,
' J
t- unu:o..flPJ,
:P ch o,0( );\ c ~o ...n d.r ~ Ot..O .n \..).r>.e.t. i..d ;. f-8 · ' ·
U..XVU.•'V\.,bo_,.( ~ 0.X '\ ~ u:G f)-fg .
-g, 0 l,1: ~ 0
t.ip ,o ~
Q

f QNJL.e-r-J ~ oJJ 'rW •C ©"- -..,...e._,u::;~J>

th..Q..C ~d o.:l Der- Yu? r-1--o-~ .

0 e\.t.,cl~ - --~ -1.-,-~ !00l_r u~. I

~ ~;~ ,bo ard , or son1etimes e,:en stick to their monitor!
, .. -- . - ·· -- •r

· crea tes grea t concen1s for Eith er way, this

the security of J.he pa~sv~'ords an~l, ther efo re,
; access of the reso urce s. the
Pas s\•·ord n1~m't·eu anr P
., h: ,.:, vo,... ,, 1-,.;,.,. ,..,._n ~n::-t m a..l·-
-.,ia() .,....-••• ,-.orn f-,._~
"\ • L4. . -- ~ ~ - -
:m·n1'strat·o~ "
study shows tha t system administrators----· •.&
- - •-I'll.} 0
a."V~ ...,,.JV '--
:\.......i...,j i,.;>. L"l,.
spe nd abo ut 40% of the ir time ~
cr•2ahng rese ttin.g or cha ngi ng ll~er pas.5wo
rds! Thi s can trul y be a nig htm are
for them . .·
, Org;~!lizations sped:fy p~ on ;lp oU cie
s, whi ch man dat e the stru ctu re of
lJasswords. For inst anc e, an drga:n.i.zatio
n policy cou ld hav e som e cf tl:e
follow-ing poli cies gov emi ng th~ pas swo rds
ofit s usei·s:
_• The pr,sswo.rf\Jength must be at leas t 8 charact
ers.
· rt mus t not con tain any blanks.
• The re mus t b~ at leas t one lower•-case alph
abe t, one upp er-c ase alph abe t,
on c digit and one spe cial cha ract er in -~ e;p
as~ ord .
, The pas swo nln 1us t beg in \~h an alphabe
t. .
As we can see: this Oike a salt in PBE
, as discussed earlier) can be a
sign ific ant det erre nt to dictionary attacks
, ·whereby atta cke rs simply take
nor mal wor ds (frm n a dictionary} _an.d try
them as pas swo rds. How eve r, this
crealL·~ a pro blem of rem emb erin g
1 cryptic pas swo rds for the end use rs:
lh~rdo rl'. f'nd use rs reso rtlo ",,1-rithtg thei
r ·passwords so111ev1he1-e., whi ch can
deft'.;1t tlw ·.•,:hole purpo::;e of a password poli
cy!
In ;11111l s lil'll , lher e arc no easy solu tion s here
!

- - ----•
-----
-~- •-·-· --
•-•• ___ __
,. ,_ -.~· . ----•-•--~•...... .-~ -- ,,..,.,__
...,..... ..__..- ~
•--r~. ,,
-,, .;...,..;,._._,-";
- - - - ••'<•·• ,. -. . • '
--~-~---- -~ -~~ ·-,·- -- --
,~ f"!_-- •,,-- ..,_,.._.'0r
, ..

~; - ~•: -.;·,. -:.--'

.... _ _,,
"
• .,.._.. - --
 I BIOMETRIC AUTHENTICATION I
( 7.6.1 Introduction

B~o 1n .e tric -a~t hen ~ica ti~n ~1ec hanis ms are re ceivi
n g a lo t of publ ic
atten tion. A b10m etlJ c d ev'1ce 1s perh aps the ultim ate attem pt
in t ryi n g to
prov e ,vh? ? ou are.0 bi01 netri c devic e work s on
th e basis of som e hum a n
ch a r a<:_t ~n s t 1cs, S~C~1 _as finge rprii1 ~, V C?ice, or patte
r n of-lin es in th e j r is of
~~':r eye -) ~he u s er d 0"ab a,~~ co!}t ains a samp le
of th e u ser'::; b io m e tric
ch a racte nstic s . Duri ng authe ntica ticii-i, the user
is requ ired t o prov id e
anot her samp le of the user 's biom etric char acter istic
s. This is matc hed with
the one in the dat3. base, anc;l if the two samp les are
cons idere d to be a valid one. '\ the same th en the us er is
- ...__ - /
;The impo rtant idea in bio~ trics is that the samp
le prod uced durin g every
'a uthe ntica tion proc ess can vary sligh tly. This
is beca use the phy sical
char acter istic s of the u ser may chan ge for a num ber
of reaso ns. Fo r insta nce,
supp ose the finge rprin t of the user is capt ured and
used for authe nt:i r: ati on
ever y tin1e . The samp le take n for every auth entic a [ion
may not be the same ,
beca use the Enge r can b e dirty , can have cuts, othe
r mark s er the fin ger's
posit ion on the read er can b e diffe rent, ,n1d ~0 c-~. '!'h~r
efo:r c., an exc1ct n 1atch
of the sa}J}_p le need not be requ ired. An appr
f1 oxim ate matc h can be
acce ptabl e. ·
---- ~
/ This :.s alsc( the rL 3.Son why , durin g ilie user regis
tratio n proc ess, mult iple
°' .?am ples of the user biom etric
data are creat ed. They are comb ined and their
aver age store d in ilie user datab ase, so that the diffe
rent possi biliti es of the
user 's san1o les durin g the actua l auili entic ation can
roug hly map to iliis
avera ge samp le. Usin g this basic philo soph y, WY biom
etric auth entic ation
· sys~ m defin es two conf igura ble para mete rs:__ the
Fals e Acc ept Rati o
(FAR ) and the Fals e Reje ct Rati o (FRR ). The
FAR is a meas urem ent of
the chan ce that a user who shou ld be rejec ted is
·actu ally acce pted l;)y a
syste m as good enou gh. FRR is a meas urem ent of tl1e
·chan ce iliat a user who
shou ld be acce pted as valid is actua lly rejec ted by
a syste m as n ot good
e noHg li-:---T hus, FAR and FRR are exac tly oppo site of e.
,/ - ·- -- . -- . . .
. .
.
a ch oilie r.
. . .. . ....
Perh aps the best secu rity solut ion is to cou1 bine the
passv .rord /PIN , a smar t
card and biom etric s. It cove rs all the three key
aspe cts relat ed to
Q; auth entic ation : who you are, what you have , and wha
t you know . How ev er,
this can turn out to be an extre n1ely comp lex syste m
. to build and/ or use .
.
7. 6. 2 The Work ing of-B iome trics
A typ1c3] auth entic ation proc ess invol ving biom etric
s firstl y invol ves the
creat ion of th e u ser's samp le and its stora ge in the user
d a t abase . Duri ng the
 . t10
act ua] auth entica . n, th : req ·red to pro vid e a sam ple ot
· e use r 1s tu . .
nat ure (e.g. a ret ina sca n or a fing erp rint ). Th1s 1s usua11y sen t ~C'.\
enc iyp ted ses sio n (e.g. by using S~L) to
tl1e serve.r. On th e se~ :er, the /·
cur ren t sam ple is dec ryp ted , and con
1pared W1th the one ,)to r e<l_ · n "-
dat aba se. If the two sam ple s ma tch to the
exp~cted d~gree on th e bas1s_.°,f fh\
par ticu lar values of FAR or FRR, the
use r 1s cons1dered as aut hen t1c ate o
successfully. OtherV\rise, the use r is con sid
ere d as invalid.
7.6 . 3 Bio me tric Tec hni que s
Biometric tec hni que s are generally classifi
ed into two sub -ca teg ori es, nam ely
ph ysi olo gic al and beh avi ora l. Let us
discuss the se in brie f.
1. Ph ysi olo gic al Tec hni que s

.A5 the nam e sug ges ts, the se tec hni

que s rely on the physical cha rac teri stic
hum an bei ngs . Since the aim is to ide ntif y hum ans uni s of
cha rac teri stic s 1nust b •.: very pro min ent que ly, the se
and dis tiri gui sha ble fro m one per son
to ano the r. Several suc h tec hni que s are
used, as me nti one d below.
(a) Face Il..1. +-1 ... ~ ... +-,...~1
1...J.J.J .~
--. ..... ~,...,,~ +1--.,.,
"-'-''-'.l..J...1..1..1.'-i u ....... , ~,.:ic ,,, ~,.. +-o ,.,hef7'•1 r-u.llU
-.- - ., --- - .- ~--- -- ...,._~
\..lJ.'--' .a.\J..v
u. .... v .... '-' ...:i~~+ ~~"
O
111e a~ u1 c L.1.ic u ... v~u.£..-;...;....
\..,\..,f t
bet we en the var iou s facial fea ture s suc ..
h as eyes, nos e, and mo uth . Th is t
dis tan ce me asu rem ent is don e usi ng gem
net rica l tec hni que s.
(b) Voice Hu ma n voice can be uni
que ly ide ntif ied bas ed on· the
cha rac teri stic s of the sou nd waves
of a voice. Som e of the se
cha rac teri stic s-ar e the pitc h and ton e.
(c) Fin ger prin t Tvledical science tell s
us tha t every hum an bei ng has a·u niq ue
fing erp rint , at lea st afte r a specifi
c age. Th e fin ger pri nt- bas ed
aut hen ti~a tion use s two app roa che s: ·mi
nut iae -ba sed and ima ge- bas ed.
In the 1i1inutiae-based tec hni que , a
gra ph of Ll-ie ind ivid ual rid ge
pos itio ns is dra wn . In the . ima ge- bas
ed tec hni que , an ima ge of the
fin ger pri nts is tak en and sto red in
the dat aba se for sub seq uen t
com par iso ns. W11ile fin ger pri nts · can cha
nge due to agei!1g or dis eas es,
the y J1ave bee n use d extensively for aut hen
tica tion . ijj
(d) Iris A111azingly, eac h per son has som
e uni que pat ter n ins ide the iris. Thi s
tec hni que is bas ed on ide ntif yin g a
per son unique1y bas ed on thi s
pat tern . Thi s me cha nis m is con sid ere
d qui te sou nd and reliable. For
~be cki ng the iris pat tern , usu ally la.s er bea
I
• n1s are en1ployed. . ,
(e) Re tina Ret ina sca nni ng is not ver.y
con1mon. The n1ain rea son beh ind
this is its hig h cost. In this me cha nis m,
the vessels car ryin g b~ood sup ply
 6
at ~1e ~ack of a htunan eye are e~a~i~ed. They provi.de a unique patten ,
which 1s used to authen ticate an md1V1dual.
f 2. Behav ioral Techniques

The idea in be~havioral techniques is to observe a person to ensure that

he/ she is not trying to daim to he someo ne else. In other \Vords, here the
en1pha sis is on checki ng that a person 's behavi or is not unusua l or abnormal.
Two m.a in techni ques are used here, as d.iscuss~d below.
(a) Keystroke Severa l charac teristic s such as the speed of typing, strengt h of
keystrokes, tin1e beh\.·een t:\No keystrokes, error percentage and
frequency, etc,, can be n1ea.su.red for identifying users. However, it is not
as reliable ,as many other authentication mechanisms.
(b) Signat ure This is an old technique. C11eques ru1d 1nany other docum ents
are expected to be physically signed by the authorizer. This is no\v
extend ed l)y kee1)ing a scanned copy of a person's signatu re and
compa ring this co1.nput e1 ~based scanne d signatu re ,vith the paper
signat1.ne as and when the need arises.