INFORMATION AND COMMUNICATION TECHNOLOGIES AND SERVICES VOLUME: 11 |NUMBER: 5 |2013 |SPECIAL ISSUE
Admission Con ol in IMS Ne wo ks
E ik CHROMY, Ma cel JADRON, Ma ej KAVACKY, S anisla KLUCIK
Ins i u e o Telecomunica ions, Facul y o Elec ical Enginee ing and In o ma ion Technology, Slo ak
Uni e si y o Technology B a isla a, Ilko ico a 3, B a isla a 812 19, Slo akia
c[email p o ec ed], [email p o ec ed], ka ac[email p o ec ed], [email p o ec ed]
Abs ac . In ou pape he e is an emphasis on sim-
ula ions o admission con ol me hods in MATLAB en-
i onmen . The main ask o admission con ol me hod
is o make a decision i he connec ion equi ing ne -
wo k access should be accep ed o he ne wo k o he
access should be ejec ed. I he connec ion is accep ed
o he ne wo k, he admission con ol has o ensu e ha
Quali y o Se ice o his connec ion will be sa is ied,
as well as Quali y o Se ice o all o he exis ing con-
nec ions. We ha e obse ed se e al Measu emen based
admission con ol algo i hms and he esul is he iden-
i ica ion o he sui able algo i hm which can es ima e
he equi ed bandwid h.
Keywo ds
Admission con ol, IMS, MBAC algo i hms,
Quali y o Se ice.
1. In oduc ion
T ends in elecommunica ion ne wo ks and se ices
end o IP Mul imedia Subsys em ne wo ks (IMS).
IMS ne wo k gua an ees Quali y o Se ice (QoS) and
Admission con ol me hods a e one o applied QoS
mechanisms [1].
Wi hin he scope o he p ojec ”Suppo o Cen e o
Excellence o SMART Technologies, Sys ems and Se -
ices II” unded by s uc u al unds o Eu opean union
we ha e buil he mos mode n IMS lab a he Ins i-
u e o Telecommunica ions. In his lab we can also
conduc esea ch aimed o admission con ol me hods,
which a e necessa y o Quali y o Se ice p o iding o
eal- ime se ices. The IMS a chi ec u e also con ains
he Resou ce Admission Con ol Subsys em (RACS) in
which he admission con ol me hods can be applied.
2. IP Mul imedia Subsys em
IP Mul imedia Subsys em a chi ec u e was de eloped
by 3GPP g oup. This a chi ec u e allows o p o ide s
o o e mul imedia se ices such as IPTV, VoIP and
many o he s. IMS is no dedica ed only o new se -
ices, bu i mus also suppo legacy se ices and
should be eady o de elopmen o new se ices.
Telecommunica ion p o ide s can deli e hei se ices
o cus ome s i espec i e o hei loca ion, access o
echnology o e minals. IMS de ines a chi ec u e
which allows con e gence o oice, ideo, da a h ough
IP based in as uc u e [2], [3], [4], [5], [6], [7], [8].
RACS is one o he impo an IMS componen s
o in e ac ion be ween con ol laye and ansmis-
sion unc ions o con ol o esou ces including e-
sou ce ese a ions, admission con ol and QoS sup-
po . The e o e RACS componen ensu es QoS in IMS
ne wo ks. Admission con ol block ecei es eques s
o QoS esou ces ia e e ence poin , e.g. bandwid h
equi emen . AC uses in o ma ion om QoS o ad-
mission con ol, i.e. AC checks i he equi ed QoS e-
sou ces a e a ailable and sends decision i he eques
is ul illed o no ia e e ence poin [9], [10].
3. Admission Con ol
Me hods
QoS in he ne wo k mus be gua an eed in o de o
suppo eal- ime eques s and eal- ime applica ions.
Th ee QoS classes a e de ined o In eg a ed se ices.
The i s is Bes -e o class. In he ne wo k wi h his
class all connec ions a e pe mi ed. Ne wo k sends
da a o hese connec ions wi h i s maximal ansmis-
sion a e. Each connec ion needs some ne wo k e-
sou ces he e o e he e is no QoS assu ance. Due o
his ac admission con ol o his class o se ice is
needed. The second class is Gua an eed se ices. This
class ensu es ha packe in he ne wo k will be no
los and gua an ees bounded end- o-end delay. This
gua an y needs pa icula bandwid h ese a ion. The
c
2013 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 373
INFORMATION AND COMMUNICATION TECHNOLOGIES AND SERVICES VOLUME: 11 |NUMBER: 5 |2013 |SPECIAL ISSUE
las class is ep esen ed by se ices wi h con olled a -
ic. Gua an eed se ices wi h con olled a ic equi e
some g ade o QoS, he e o e hey need admission con-
ol o es ima ion o a numbe o connec ion o which
hey can ensu e QoS. Admission con ol makes a de-
cision i he incoming connec ion will be accep ed o
ejec ed. The measu e is p o ision o QoS o incoming
connec ion and p ese a ion o QoS o exis ing connec-
ions.
The ield o admission con ol is di ided in o Pa ame-
e based admission con ol (PBAC) and Measu emen
based admission con ol (MBAC). PBAC me hods e-
ga d a ic cha ac e is ics o all connec ions, such as
peak ansmission a e. This me hod de e mines e-
qui ed ne wo k esou ces o all connec ions based on
such pa ame e . MBAC me hods a e aimed o mea-
su emen o ac ual a ic in he ne wo k. This me hod
accep s incoming connec ions upon ealized measu e-
men s.
Admission con ol is necessa y o admission o new
connec ion. I is possible o design he model o ad-
mission con ol which ensu es QoS by use o admission
con ol me hods. C ea ed models can be used sepa-
a ely o in combina ion o achie emen o be e QoS.
The main ask o admission con ol me hods is o es i-
ma e equi ed bandwid h o incoming da a low and o
decide i his bandwid h can be alloca ed. Admission
con ol me hods a e used mainly o se ices sensi i e
o he delay and ji e o o eal- ime se ices. The e
a e a ious admission con ol me hods and hey di e
mainly in di e en a ic ypes and me hod o ealiza-
ion. Some o he AC me hods a e based on ma hema -
ical calcula ions and s a is ical ma ke s while o he s
a e based on a ic measu emen s [1].
3.1. Condi ions o AC Me hods
Fo QoS p o ision admission con ol me hods mus
ull ill ollowing condi ions:
•p o ide QoS o incoming connec ion while exis -
ing connec ions a e no a ec ed,
• as decision (in o de o p e en delays),
•e icien capaci y u iliza ion and e ec i e band-
wid h alloca ion o pa icula lows,
•simple applicabili y in o he sys em,
•adap a ion o new se ice [1].
Admission con ol is impo an mainly in access ne -
wo k whe eas nodes in backbone ne wo ks ha e high
ansmission a es and in o ma ion abou bandwid h
calcula ions hey send o edge nodes. The ne wo k is
u ilized mos ly a he edge wha is also a eason o cen-
aliza ion o admission con ol o edge nodes. RACS
is one o he IMS componen s ha ensu es equi ed
QoS [9].
3.2. PBAC Me hods
PBAC can be p e e ed due o hei simple implemen-
a ion. They wo k wi h pa ame e s such as peak o
e ec i e bandwid h o incoming low ins ead o alues
measu ed in he ne wo k. Th ough PBAC me hods
we can limi cons ain s caused by measu emen s and
ne wo k moni o ing [11].
3.3. MBAC Me hods
MBAC me hods use measu emen o ac ual a ic in
he ne wo k o decision abou admission o new da a
low. MBAC me hods make he decision p ocess based
on measu emen s and QoS pa ame e s. In he case o
measu emen and ne wo k moni o ing he mo e e i-
cien ne wo k esou ces u iliza ion o agg ega e da a
lows o o lowe ansmission a e han he peak a e
is possible compa ed o PBAC me hods. In his sce-
na io emaining bandwid h can be used o o he da a
lows [12].
3.4. MBAC Algo i hms
Va ious measu emen based algo i hms a e known. In
he pape we deal wi h ollowing algo i hms:
•Simple Sum [13],
•Measu ed Sum [14],
•P edic ed Sum [15],
•Hoe ding Bound [16] and
•Accep ance Region [17].
1) Simple Sum Algo i hm
The algo i hm h ough he sum o exis ing lows simple
ensu es i sel agains no exceeding o a ailable band-
wid h. I accep s new da a low only i he ollowing
condi ion is ul illed:
+ α< C, (1)
whe e is a sum o ese ed ansmission a es
[kbi ·s−1], Cis link capaci y [kbi ·s−1], αis index o
incoming low and αis ansmission a e o new da a
low [kbi ·s−1].
c
2013 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 374
INFORMATION AND COMMUNICATION TECHNOLOGIES AND SERVICES VOLUME: 11 |NUMBER: 5 |2013 |SPECIAL ISSUE
Measu ed sum algo i hm is he simples admission
con ol algo i hm. The e o e i is he mos imple-
men ed algo i hm in ou e s and swi ches. I is o en
used in combina ion wi h WFQ me hod [18] in o de
o ensu e low enqueue delay. WFQ alloca es indi id-
ual se ice queue o each da a low, so da a bu s s a e
sepa a ed om each o he .
2) Measu ed Sum Algo i hm
This algo i hm ensu es ha sum o peak ansmission
a es o new lows and ac ual a ic is lowe han a -
ge ed link capaci y u iliza ion. I is exp essed h ough
he condi ion:
+p≤u·C, (2)
whe e is measu ed ac ual a ic o exis ing connec-
ions [kbi ·s−1], pis peak a e o incoming da a low
[kbi ·s−1], Cis link capaci y [kbi ·s−1] and uis link
u iliza ion pa ame e .
Ac ual measu ed a ic will no be usable o u-
u e connec ion admissions, because new connec ion
will occu in he ne wo k and a ic a y. Link u iliza-
ion pa ame e is se o a alue lowe han 1 in o de
o gua an ee he QoS o all connec ions. Measu ing
mechanism makes measu emen s equi ed o measu ed
sum and i is based on window size o his algo i hm.
This mechanism se s he ixed ime in e al o sam-
ple pe iod and a longe ime in e al o window pe iod
which is mul iple o sample pe iod. Link u iliza ion is
measu ed a he end o each sample pe iod and he
highes alue o a ic in one window pe iod is de ined
as he end o he window pe iod. End o he p e ious
window pe iod is ega ded as measu ed a ic. A he
end o each window he window closing pe iod is ese
o ini ial alue.
3) P edic ed Sum
P oposed algo i hm samples da a du ing de ined in e -
als. Such acqui ed samples ep esen inpu da a o
p edic ion. A e each sampling he p edic ion o he
nex da a a ic is done. Such p edic ion will be hen
used o u u e admission con ol o admission o ejec-
ion o connec ion. Algo i hm o p edic ed sum makes
a decision based on he condi ion:
ˆx(n+ 1) + pα≤uC, (3)
whe e αis index o incoming da a low, x(n+1) is p e-
dic ed agg ega ed a ic used in nex sampling n+1
[kbi ·s−1], pαis peak ansmission a e o new low
[kbi ·s−1], Cis link capaci y [kbi ·s−1] and uis sou ce
u iliza ion.
4) Hoe ding Bound
This me hod uses Hoe ding bound o es ima ion o
link a ic. Hoe ding bound se s he highe bound
o a ic o connec ions in he ne wo ks acco ding o
equa ion:
CH( , {pi}1≤i≤n, ε) = a g +sln(1
ε)Pn
i=1 p2
i
2,(4)
whe e a g is o al a ic o all connec ions [kbi ·s−1],
piis peak a e o i- h connec ion [kbi ·s−1] and εis
p edic ion ha a ic will exceed link capaci y (p ob-
abili y o packe losses).
Hoe ding bound algo i hm makes admission deci-
sion based on he equa ion:
CH+p≤C, (5)
whe e pis peak ansmission a e o new connec ion
[kbi ·s−1] and Cis link capaci y [kbi ·s−1].
I he sum o Hoe ding bound o all exis ing con-
nec ions and peak a e o new connec ion is lowe han
a ailable link capaci y he admission con ol accep s
new connec ion in o ne wo k. On he con a y, i his
sum is highe han a ailable link capaci y, connec ion
will be ejec ed. Compa ed o Measu ed sum algo-
i hm, he Hoe ding bound algo i hm will no ese e
he abo e capaci y o a sho - e m aised a ic, be-
cause Hoe ding bound is adap ed o his case. Mech-
anism o measu emen used in his algo i hm uses ex-
ponen ial a e aging. Fi s ly, a e age a e is measu ed,
hen he exponen ial a e age is calcula ed and inally
Hoe ding bound CHis es ima ed.
5) Accep ance Region
This algo i hm es ima es egion in which he link u i-
liza ion is maximized a he expense o packe losses.
Accep ance egion can be es ima ed on he basis o ol-
lowing pa ame e s: gi en bandwid h, memo y space o
swi ches, pa ame e s o bu e s ack il e , da a bu s s
o da a low and p obabili y ha ac ual a ic will ex-
ceed he accep ance bound. We suppose he Poisson
dis ibu ion o incoming independen eques s in ac-
cep ance egion calcula ion. In he case o pa ame e
alues a ia ion he algo i hm beha io is no p ecise.
Ve sion o he measu emen p ocess o his algo i hm
ensu es ha sum o measu ed a ic and ansmission
a e o new da a low will no exceed he accep ance
egion.
C(s) = 1
slog1 +
p(esp −1),(6)
whe e C(s) is es ima ed bandwid h o agg ega ed a -
ic [kbi ·s−1], is a e age ansmission a e o a ic
[kbi ·s−1], pis peak ansmission a e [kbi ·s−1] and s
c
2013 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 375
INFORMATION AND COMMUNICATION TECHNOLOGIES AND SERVICES VOLUME: 11 |NUMBER: 5 |2013 |SPECIAL ISSUE
is space pa ame e which alue is om in e al om 0
o 1.
4. Simula ions
This chap e deals wi h simula ions o AC me hods.
Ne wo k opology used in simula ions is shown in he
Fig. 1.
Fig. 1: Ne wo k opology.
Admission con ol me hod is applied in ou e . An
assump ion o his opology is ha he e is he numbe
o use s gene a ing s ochas ic eques s in ime. These
eques s ep esen s ochas ic ansmission a es om
a iable bi a e sou ces. Each sou ce equi es di e en
demands on ansmission a e in di e en ime sam-
ples. T ansmission a es a y om 0 o 128 kbi ·s−1.
I is app op ia e o use Gaussian dis ibu ion o sim-
ula ion wi h a high numbe o use s [19].
4.1. Simple Sum Simula ions
The Simple sum simula ions a e shown in he Fig. 2.
Simple sum algo i hm is he simples admission con-
ol algo i hm. The main goal o his algo i hm is o
admi o ejec incoming low on he basis o a ailable
bandwid h which can no be exceeded. The simula ion
esul s a e shown in he Fig. 2. The black cu e ep e-
sen s de ined link capaci y - 2 Mbi ·s−1. The ed cu e
ep esen s ac ual a ic o all use s in he ne wo k be-
o e decision p ocess. Based on he incoming admission
eques compa ison wi h a ailable bandwid h is done
and connec ion is hen accep ed (g een cu e). This
wo cou ses show si ua ion when he ne wo k is no
loaded o maximum.
Compa ison wi h opposed scena io ( a ic load nea
o he link capaci y 2 Mbi ·s−1) a e made h ough blue
and iole cu es. Blue cu e ep esen s a ic on he
bo de o bandwid h (i.e. 2 Mbi ·s−1, o 21 use s). Any
Fig. 2: Simple sum a ic simula ion.
new connec ion can no be admi ed due o lack o he
bandwid h. Accep ance o new connec ion in such a
si ua ion will lead o QoS deg ada ion ( iole cu e).
4.2. Measu ed Sum Simula ions
Measu ed sum algo i hm does no ega d he o al link
capaci y, bu only i s pa (Fig. 3) compa ed o sim-
ple sum algo i hm. The esul o he decision p ocess
Fig. 3: Measu ed sum a ic simula ion.
is shown in he Fig. 3. Red cu e and g een cu e
ep esen o al a ic be o e and a e connec ion ad-
mission. Blue and iole cu es ep esen bound cases.
Black cu e ep esen s 95 % o o al link capaci y (i.e.
1,9 Mbi ·s−1). Blue cu e ep esen s 20 accep ed con-
nec ions. Viole cu e ep esen s o e un o de ined
link capaci y, he e o e Measu ed sum algo i hm will
ejec his connec ion.
He e we can see he di e ence be ween Simple sum
and Measu ed sum algo i hm. Ano he connec ion
would be accep ed by Simple sum algo i hm because i
no uses he ansmission medium u iliza ion pa am-
c
2013 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 376
INFORMATION AND COMMUNICATION TECHNOLOGIES AND SERVICES VOLUME: 11 |NUMBER: 5 |2013 |SPECIAL ISSUE
e e . Measu ed sum algo i hm lea es ese e o he
case o unexpec ed bandwid h inc ease o connec ions.
4.3. Hoe ding Bound and
Accep ance Region Simula ions
In he case o unchanged ne wo k opology and highe
load he p ope selec ion o AC algo i hm is he key
elemen in o de o sa e some bandwid h. In he
Fig. 4 he bound scena io wi h Hoe ding bound ( iole
cu e) and Accep ance egion (blue cu e) algo i hms
a e shown. Red cu e ep esen s ac ual a ic in he
ne wo k.
Fig. 4: Accep ance o connec ions in ne wo k wi h highe load.
We can see ha Hoe ding bound algo i hm is mo e
sui able. A he momen 19 connec ions we e in he
ne wo k (ac ual a ic) and new connec ion is accep ed
only by Hoe ding bound algo i hm. Accep ance e-
gion algo i hm has ejec ed new connec ion despi e o
su icien bandwid h a ailable, he e o e om e ec i e
bandwid h u iliza ion, he Hoe ding bound algo i hm
is mo e p e e able.
5. Conclusion
Realized simula ions ha e shown he di e en alloca-
ion o bandwid h while each o simula ed algo i hms
wo ks on di e en measu emen p inciples. In he case
o w ong AC algo i hm selec ion he was e o he band-
wid h can occu , wha is ine icien and economic un-
p o i able. Based on he simula ions we ha e shown
ha in he case o highe a ic load he selec ion o
Admission Con ol algo i hm is e y impo an . F om
he e ec i e bandwid h u iliza ion poin o iew he
Hoe ding bound algo i hm is sui able.
Acknowledgmen
This wo k is a pa o esea ch ac i i ies conduc ed
a Slo ak Uni e si y o Technology B a isla a, Facul y
o Elec ical Enginee ing and In o ma ion Technology,
Ins i u e o Telecommunica ions, wi hin he scope o
he p ojec s ”G an p og amme o suppo young e-
sea che s o STU - Modelling o T a ic and T a ic Pa-
ame e s in IPTV Ne wo ks”, ”Suppo o Cen e o
Excellence o SMART Technologies, Sys ems and Se -
ices II, ITMS 26240120029, co unded by he ERDF”
and p ojec VEGA no. 1/0106/11 ”Analysis and P o-
posal o Ad anced Op ical Access Ne wo ks in he
NGN Con e ged In as uc u e U ilizing Fixed T ans-
mission Media o Suppo ing Mul imedia Se ices”.
Re e ences
[1] LAI, Y.-Ch. and S.-F. TSAI. Un ai ness o
measu emen -based admission con ols in a he -
e ogeneous en i onmen . In: P oceedings. Eigh h
In e na ional Con e ence on Pa allel and Dis-
ibu ed Sys ems. ICPADS 2001. Kyongju Ci y:
IEEE, 2001, pp. 667–674. ISBN 0-7695-1153-8.
DOI: 10.1109/ICPADS.2001.934882.
[2] VOZNAK, M. and J. ROZHON. App oach
o s ess es s in SIP en i onmen based on
ma ginal analysis. Telecommunica ion Sys ems.
2013, ol. 52, iss. 3, pp. 1583–1593. ISSN 1018-
4864. DOI: 10.1007/s11235-011-9525-1.
[3] VOZNAK, M. and F. REZAC. Web-based IP ele-
phony pene a ion sys em e alua ing le el o p o-
ec ion om a acks and h ea s. WSEAS T ans-
ac ions on Communica ions. 2011, ol. 10, iss. 2,
pp. 66–76. ISSN 1109-2742.
[4] VOZNAK, M. and F. REZAC. Th ea s o oice
o e IP communica ions sys ems. WSEAS T ans-
ac ions on Compu e s. 2010, ol. 9, iss. 11,
pp. 1348–1358. ISSN 1109-2750.
[5] ROKA, R. and F. CERTIK. Modeling o en i on-
men al in luences a he signal ansmission in he
op ical ansmission medium. In e na ional Jou -
nal o Elec ical Communica ion Ne wo ks and In-
o ma ion Secu i y. 2012, ol. 4, no. 3, pp. 144–
162. ISSN 2076-0930.
[6] KYRBASHOV, B., I. BARONAK, M. KOVACIK
and V. JANATA. E alua ion and In es iga ion o
he Delay in VoIP Ne wo ks. Radioenginee ing.
2011, ol. 20, iss. 2, pp. 540–547. ISSN 1210-2512.
[7] KOCKOVIC, L. and I. BARONAK. Al e na i es
o P o iding IPTV Using IP Mul imedia Sub-
sys em. In e na ional Jou nal o Compu e s and
c
2013 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 377

INFORMATION AND COMMUNICATION TECHNOLOGIES AND SERVICES VOLUME: 11 |NUMBER: 5 |2013 |SPECIAL ISSUE
Technology. 2012, ol. 3, no. 2, pp. 188–192.
ISSN 2277-3061.
[8] MISUTH, T. and I. BARONAK. Applica-
ion o M/G/1/K Model o Agg ega ed VoIP
T a ic Packe Loss Es ima ion. In: 2012
35 h In e na ional Con e ence on Telecommuni-
ca ions and Signal P ocessing (TSP). P ague:
IEEE, 2012, pp. 42–46. ISBN 978-1-4673-1118-2.
DOI: 10.1109/TSP.2012.6256249.
[9] ETSI ES 282 003. Telecommunica ions and In e -
ne con e ged Se ices and P o ocols o Ad anced
Ne wo king (TISPAN); Resou ce and Admission
Con ol Sub-sys em (RACS); Func ional A chi-
ec u e. V1.1.1. 650 Rou e des Lucioles: F-06921
Sophia An ipolis Cedex - FRANCE, 2006. A ail-
able a : h p://www.e si.o g/deli e /e si_
es/282000_282099/282003/01.01.01_60/es_
282003 010101p.pd
[10] WUTHNOW, M., M. STAFFORD and J. SHIH.
IMS: A New Model o Blending Applica ions.
Boca Ra on: CRC P ess, 2010. ISBN 978-1-4200-
9285-1.
[11] YERIMA, S. Implemen a ion and E alua ion o
Measu emen -Based Admission Con ol Schemes
Wi hin a Con e ged Ne wo ks QoS Managemen
F amewo k. In e na ional Jou nal o Compu e
Ne wo ks &Communica ions (IJCNC). 2011,
ol. 3, no. 4, pp. 137–152. ISSN 0975-2293.
[12] ALIPOUR, E. and K. MOHAMMADI. Adap i e
Admission Con ol o QoS Gua an ee in Di e -
en ia ed Se ices Ne wo ks. In e na ional Jou nal
o Compu e Science and Ne wo k Secu i y (IJC-
SNS). 2008, ol. 8, no. 6, pp. 93–98. ISSN 1738-
7906.
[13] JAMIN, S. and S. SHENKER. Measu emen -based
Admission Con ol Algo i hms o Con olled-load
Se ice: A S uc u al Examina ion. Michigan:
Uni e si y o Michigan, 1997, Technical Repo .
CSE-TR-333-97.
[14] JAMIN, S., P. DANZIG, S. SHENKER and L.
ZHANG. A Measu emen -Based Admission Con-
ol Algo i hm o In eg a ed Se ices Packe Ne -
wo ks. IEEE/ACM T ansac ions on Ne wo king.
1997, ol. 5, iss. 1, pp. 56–70. ISSN 1063-6692.
DOI: 10.1109/90.554722.
[15] EGYHAZY, M. and Y. LIANG. P edic ed
Sum: A Robus Measu emen -Based Ad-
mission Con ol wi h Online T a ic P edic-
ion. IEEE Communica ions Le e s. 2007,
ol. 11, iss. 2, pp. 204–206. ISSN 1089-7798.
DOI: 10.1109/LCOMM.2007.061127.
[16] FLOYD, S. Commen s on Measu emen -based
Admission Con ol o Con olled-Load Se ices.
Compu e Communica ions Re iew. 1996, ol. 26,
iss. 3, pp. 1–16. ISSN 0146-4833.
[17] GIBBENS, R., F. KELLY and P. KEY. A
Decision–Theo e ic App oach o Call Admis-
sion Con ol in ATM Ne wo ks. IEEE Jou -
nal on Selec ed A eas in Communica ions. 1995,
ol. 13, iss. 6, pp. 1101–1114. ISSN 0733-8716.
DOI: 10.1109/49.400665.
[18] ASHOUR, M. and T. LE-NGOC. Pe o -
mance o Weigh ed Fai Queuing Sys ems
wi h Long Range Dependen T a ic Inpu s.
In: Canadian Con e ence on Elec ical and
Compu e Enginee ing, 2005. Saska oon: IEEE,
2005, pp. 1002–1005. ISBN 0-7803-8885-2.
DOI: 10.1109/CCECE.2005.1557145.
[19] BRICHET, F. and A. SIMONIAN. Conse a i e
Gaussian models applied o Measu emen -based
Admission Con ol. In: Six h In e na ional Wo k-
shop on Quali y o Se ice (IWQoS’98). San F an-
cisco: IEEE, 1998, pp. 68–71. ISBN 0-7803-4482-
0. DOI: 10.1109/IWQOS.1998.675222.
Abou Au ho s
E ik CHROMY was bo n in Velky K is, Slo akia,
in 1981. He ecei ed he M.Sc. deg ee in elecommuni-
ca ions in 2005 om Facul y o Elec ical Enginee ing
and In o ma ion Technology o Slo ak Uni e si y o
Technology (FEEIT STU) in B a isla a. In 2007 he
submi ed Ph.D. wo k om he ield o Obse a ion
o s a is ical p ope ies o inpu low o a ic sou ces
on i ual pa hs dimensioning. Nowadays he wo ks as
assis an p o esso a he Ins i u e o Telecommunica-
ions o FEEIT STU B a isla a.
Ma cel JADRON was bo n in Le ice, Slo akia, in
1990. He is a s uden a he Ins i u e o Telecommu-
nica ions, FEEIT STU in B a isla a.
Ma ej KAVACKY was bo n in Ni a, Slo akia, in
1979. He ecei ed he M.Sc. deg ee in elecommuni-
ca ions in 2004 om FEEIT STU in B a isla a. In
2006 he submi ed Ph.D. wo k “Quali y o Se ice in
B oadband Ne wo ks”. Nowadays he wo ks as assis-
an p o esso a he Ins i u e o Telecommunica ions
o FEEIT STU B a isla a.
S anisla KLUCIK was bo n in B a isla a,
Slo akia, in 1984. He ecei ed he M.Sc. deg ee in
elecommunica ions in 2009 om FEEIT STU in
B a isla a. In 2012 he submi ed Ph.D. wo k om he
ield o Sou ce and QoS a ic pa ame e s modeling
c
2013 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 378
INFORMATION AND COMMUNICATION TECHNOLOGIES AND SERVICES VOLUME: 11 |NUMBER: 5 |2013 |SPECIAL ISSUE
in NGN ne wo ks. Nowadays he wo ks as esea ech a
he Ins i u e o Telecommunica ions o FEEIT STU
B a isla a.
c
2013 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING 379