A Flexible Contention Resolution Scheme for QoS Provisioning in Optical Burst Switching Networks

Size: px
Start display at page:

Download "A Flexible Contention Resolution Scheme for QoS Provisioning in Optical Burst Switching Networks"

Transcription

1 A Flexible Conenion Resoluion Scheme for QoS Provisioning in Opical Burs Swiching Neworks Ashok K. Turuk a, Rajeev Kumar b,,1 a Deparmen of Compuer Science and Engineering, Naional Insiue of Technology, Rourkela, Orissa , India b Deparmen of Compuer Science and Engineering, Indian Insiue of Technology, Kharagpur WB , India Absrac Conenion resoluion is a major issue in bufferless opical burs swiching (OBS) neworks. The exising conenion resoluion schemes consider prioriy and arrival ime o resolve conenion. For mos mulimedia applicaions prioriy and delay are key parameers for QoS provisioning. In his paper, we propose a new signaling proocol for reducing conenion in OBS neworks and consider hree parameers, namely, prioriy of he burs, number of hops raversed, and burs size ino accoun o resolve conenion. The source node in he proposed proocol can be informed of he conenion up o halfway along he pah of he burs, and hus, can reschedule he burs accordingly. The scheme is adapable o boh prioriized and delay consrained raffic. We call he scheme OBS-F lex. For selecing a daa channel, we propose hree channel selecion algorihms, namely, Leas Recenly Used (LRU), Firs Fi (FF), and Prioriy Se (PS). We simulae OBS-F lex and compare wih preempive prioriy jus-enough-ime (PPJET) conenion resoluion scheme. We show ha OBS-F lex ouperforms PPJET in erms of burs loss raes. For simulaion, we have considered Poisson and bursy raffic models. Key words: Opical burs Swiching, wavelengh division muliplexing, QoS, conenion resoluion, blocking probabiliy, channel selecion algorihm, Poisson and bursy raffic. Corresponding auhor. addresses: akuruk@nirkl.ac.in (Ashok K. Turuk), rkumar@cse.iikgp.erne.in, raj@iik.ac.in (Rajeev Kumar). 1 Presen address: Deparmen of Compuer Science and Engineering, Indian Insiue of Technology, Kanpur UP , India Preprin submied o Compuer Communicaions

2 1 Inroducion There has been a phenomenal increase in he number of Inerne users and he variey of Inerne applicaions in recen years. This has resuled in exponenial growh of Inerne raffic and demanding a huge bandwidh a he backbone nework. To mee his growing demand for bandwidh, wavelengh division muliplexing (WDM) nework has become he de-faco choice for he backbone nework. IP over WDM neworks have drawn much aenion among researchers, and many inegraion schemes beween IP and WDM layers have been proposed e.g., [1, 6, 17]. To carry IP raffic over WDM neworks hree swiching echnologies have been sudied: opical circui swiching, packe swiching and burs swiching. Opical circui swiching and packe swiching have heir own limiaions when applied o WDM neworks. Circui swiching is no bandwidh efficien unless he duraion of ransmission is greaer han he circui esablishmen period [12]. I is shown ha esablishmen of circuis (lighpahs) in opical neworks is an NP-hard problem [3]. Many heurisics and approximaion algorihms exis for esablishing lighpahs in opical neworks e.g., see [5] and he references herein. Packe swiching is hop-by-hop sore and forward scheme and needs buffering and processing a each inermediae node. I is flexible and bandwidh efficien. However, echnology for buffering and processing in opical domain is ye o ge maure for his scheme o commercialize. Fiber delay lines proposed in he lieraure provide limied buffer and are suiable only when delays are fixed. In his conex, opical burs swiching (OBS) is emerging as a new swiching paradigm for nex generaion opical neworks. I combines feaures of boh circui and packe swiching. As such here exiss no formal definiion of OBS; he feaures defined by Yoo and Qiao [20] for OBS have become he de-faco sandards. OBS schemes are based on eiher one-way (for mos cases) and wo-way reservaion proocols. The burs-size granulariy (which lies beween circui and packe swiching), separaion of conrol and daa burss, one-way or wo-way reservaion scheme, and no opical buffering are he imporan characerisics of OBS paradigm. In a bufferless opical nework one of he conending burss is dropped. Therefore, burs loss ha should be minimized in OBS neworks is he key design parameer. Several echniques have been proposed o reduce burs loss and provide QoS in OBS neworks. In one-way reservaion proocol daa burs follows he conrol burs afer a predeermined offse-ime. Examples of such proocols are Tell-n-Go [13, 16], Jus-Enough-Time (JET) [18, 19] and burs segmenaion [14, 23]. Two-way reservaion proocols require an explici release of reserved resources [4, 15]. An example of a wo-way reservaion proocol is a Jus-In-Time (JIT) [15] scheme. The offse ime in one-way reservaion proocol is aken o be he sum of processing delay of conrol packe a each inermediae node. This ime is oo shor o resched- 2

3 ule he ransmission in case of conenion and he conending burs is dropped. In wo-way reservaion proocol, as in JIT, daa burs is sen before receiving an acknowledgmen, resources are reserved from he ime reques is received and remain reserved unil a release message is received. Loss of a release message leads o he wasage of bandwidh. Since he burs is sen before receiving an acknowledgmen, in case of conenion i is dropped. In Prioriy-Jus-Enough-Time (PJET), Yoo e al. [21] assigned an addiional offse ime, in addiion o he base offse ime for each class of raffic o reduce burs loss. The higher prioriy raffic is assigned an addiional offse in ime. Kaheel and Alnuewiri [7] proposed a preempive prioriized JET (PPJET) scheme. In PPJET, a higher prioriy reques can preemp he reservaion of lower prioriy reques if heir ransmission has no sared. This can, however, increase he loss of lower prioriy raffic significanly. Vokkarane e al. [14] proposed a prioriized burs segmenaion approach. In heir approach when a conenion occurs he overlapped porion of he burs is dropped. This mehod needs a complex implemenaion echnique o segmen he burs for a drop. Zhang e al. [22] proposed a wavelengh grouping and early drop scheme where a low prioriy raffic is dropped firs in case of congesion in he nework. In heir scheme, burss belonging o lower prioriy class are dropped inenionally wih a pre-deermined probabiliy before possibly conending wih he burs of a higher prioriy class. To provide guaraneed services, core rouers have o mainain raffic saisics for each suppored class of raffic. There are few oher sudies oo. For example, Lu e al. [10] proposed an inermediae node iniiaed reservaion where an inermediae node along he pah can iniiae an reservaion o reduce burs loss due o conenion. Boudriga [2] assigned differen delay unis o each class of raffic in order o isolae higher prioriy class from he lower prioriy class. Lee and Griffih [9] presened raffic engineering echnique o suppor QoS in opical Inerne. The mechanism proposed by hem ries o uilize he available wavelengh efficienly in order o provide lower delays. Kim e al. [8] proposed deflecion rouing mechanism o reduce burs losses. They defined hreshold funcion o reroue he conending burss. In heir scheme, defleced burss may ake longer pah o reach and require large buffers a is desinaion. Mos of he researchers have aemped o reduce blocking probabiliy of differen class of raffic in order o provide prioriy based services. To reduce burs loss differen delays are pre-assigned o each class of raffic. In all of he above menioned proocols, in case of conenion, one of he burss is dropped. There is no way o reschedule he burs in case of conenion eiher because he offse ime is oo shor o reschedule as in one-way proocol, or he burs is sen before receiving an acknowledgmen as in wo-way proocol. In his paper, we presen a new signaling proocol o reduce burs loss due o conenion in OBS neworks. The source node in he proposed scheme can be informed of he conenion up o halfway along he pah of he burs, and can hus, reschedule he burs accordingly. We call he proposed scheme OBS-F lex. We consider packe 3

4 loss and number of hops raversed, in addiion o prioriy, for resolving conenion. The scheme is generic and can easily be adaped o saisfy delay consrains. The main aim of his work is o reduce blocking probabiliy of he burss arising due o resource conenion a inermediae nodes as well as o mee he delay consrains of he delay sensiive raffic. The proposed OBS-F lex guaranees ha he burs succeeds when conenion occurs up o halfway along he pah; he conenion should be resolved in accordance wih saisfacion of QoS parameers. To selec daa channel in OBS-F lex, we propose hree channel selecion algorihms (i) Leas Recenly Used (LRU), (ii) Firs Fi (FF), and (iii) Prioriy Se (PS). Channel selecion algorihms are run a he ingress rouers o selec a daa-channel for reservaion and for subsequen ransmission. We evaluae he proposed OBS-F lex wih he above hree channel selecion algorihms, and presen resuls for Poisson and bursy raffic models. Res of he paper is organized as follows. Secion 2 explains he proposed conenion resoluion echnique including assumpions and noaions used in his work. Channel selecion algorihms are explained in Secion 3. Simulaion resuls are presened in Secion 4 and compared wih PPJET scheme. Finally, conclusions are drawn in Secion 5. 2 Conenion Resoluion Scheme OBS-F lex 2.1 Assumpions and Noaions We model an opical nework by means of an undireced graph G(V, E) where V is he se of verices (nodes) and E represens he se of links/edges in he nework. Two ypes of nodes (hereafer, we use he erms node and rouer inerchangeably), namely, edge and core rouers are idenified. Every edge rouer has (n e 1) P elecronic buffers where n e is he number of edge rouers, and P is he number of prioriy classes suppored in he sysem. Each buffer belongs o a specific pair of prioriy class and an egress rouer. The core rouer has no buffer; his is a desirable feaure of OBS neworks. Besides processing and forwarding he conrol packe, core rouer has he capabiliy of generaing is own conrol packe. A core rouer acs as a ransi rouer for daa raffic. Thus, daa raffic remains in opical domain from ingress o egress rouer. We consider propagaion delay,, o be he same beween every pair of adjacen verices in he graph G; his assumpion simplifies he simulaion. Though his is a highly resriced assumpion, his is reasonable because main aim of he simulaion is o demonsrae low burs loss in case of conenion. Processing delay of he conrol packe a each rouer is assumed o be. Few more noaions used in res of he paper are defined below: original burs: A burs for which resources are already reserved a he core rouer, 4

5 S A D S A D Time T X T X (a) (b) Fig. 1. Principle of OBS-F lex conending burs: A burs whose reservaion reques has resuled in a resource conenion a he core rouer, H sd (z): Toal number of hops for he reques z beween he source-desinaion pair (s, d), and Hi sd (z): Remaining number of hops for he reques z beween he source-desinaion pair (s, d) a node i. We idenify he following hree siuaions ha can occur when an inermediae rouer receives a reservaion reques: no conenion (NC):- when no conenion occurs for resources a he inermediae core rouer. conenion resolved (CR):- when a conenion occurs a an inermediae core rouer i and for he conending burs s reques Hi sd (z) > H sd (z)/2. conenion-no-resolved (CNR):- when conenion occurs a an inermediae core rouer i and for he conending burs s reques Hi sd (z) > H sd (z)/ Proposed OBS-F lex Scheme In case of conenion, he conending burs is dropped in OBS neworks. The basis of his work is ha if ransmission of burs is delayed a he source for he duraion of he conenion period hen he ransmission of he burs would be successful. To delay he burs a conrol signal is o be sen o he source from he node where conenion has aken place. The conrol signal should reach he source before he expiry of he offse ime as shown in Fig. 1. In Fig. 1(a) conenion has occurred a node A and he burs is dropped a node A. However, in Fig. 1(b) a conrol packe is sen from node A where conenion has occurred and is received by he source S before he offse ime T. Transmission of he burs is furher delayed by he source S for he conenion period. Thus, ransmission of he burs is successful. We use he above mechanism in he proposed OBS-F lex o resolve conenion. Offse ime is aken o be he propagaion delay beween source-desinaion pair. We use wo conrol packes: Forward (F ) and Reverse (R) conrol packes. F - conrol packe is sen o reserve resources along he pah, and R-conrol packe is 5

6 s A B C d s A B C d s A B C d Time T T T (a) (b) (c) Fig. 2. Timing diagram of a burs swich nework: (a) no conenion occurs a inermediae nodes, (b) conenion occurs a node A, and (c) conenion occurs a node C s A B C d s A B C d T T T d Time (a) (b) Fig. 3. Timing diagram in OBS-F lex: (a) conenion a node A is resolved, and (b) conenion a node C hough he burs is dropped a node B. sen from he inermediae node where conenion has aken place eiher o delay he ransmission a source or o release he reserved resource. Processing of F and R conrol packes is explained in he following paragraphs. Working of OBS-F lex is explained wih he help of he iming diagrams illusraed in Figs. 2 and 3. The oal delay encounered by a conrol packe for he sourcedesinaion pair (s, d) is no greaer han = H sd (z). The offse-ime, T, in OBS is aken o be a leas. In Fig. 2, he number of hops beween sourcedesinaion pair (s, d) is 4. Therefore, he offse-ime T in OBS is 4. In OBS, if a conenion occurs say a node A or a node C hen he burs is dropped a A or a B as shown in Figs. 2(b) and 2(c), respecively. Wih his offse ime a conending burs canno be furher delayed. In OBS-F lex, he offse, T, beween he source-desinaion pair (s, d) is aken o be (+)H sd (z). In he above example, he offse ime beween he source-desinaion pair (s, d) is 4( + ). Le us consider Fig. 2(b) where conenion has occurred a node A, and d is he duraion of he conenion period. F -conrol packe has aken one hop o reach he node A from source s, and R- conrol packe is sen from node A o source s o delay he ransmission of burs for he conenion period d. The packe will reach s a T = 2( + ) unis afer he source s has sen he F - conrol packe (Fig. 3(a)). The offse-ime T > T i.e., source s will receive he R-conrol packe o delay he ransmission before expiry of he offse-ime. Hence, 6

7 he ransmission of he burs is delayed and he burs is no dropped a node A; his is illusraed in Fig. 3(a). Nex, we consider Fig. 2(c) where conenion has occurred a node C, and d is he duraion of he conenion period. R-conrol packe is sen from node C o source s o delay ransmission of he burs for he conenion period d; i will reach s a T = 6( + ). The offse-ime T < T, i.e., source s will receive R-conrol packe afer i has ransmied he burs and he burs is dropped a node C. Therefore, in he presen case, R-conrol packe is sen o release resources reserved a he inermediae node raher han delaying a he source. Hence, R-conrol packe will release he resources reserved a node B before he burs arrives a node B and is dropped a node B raher han a node C. This gives rise o beer uilizaion of he resources on link BC which was earlier occupied by he reques. The OBS-F lex differs from oher OBS schemes in wo aspecs - one, he offse ime, and second, he mehods adoped for conenion resoluion. In oher OBS schemes, he offse ime is H where is he processing delay of conrol packe a each node, and H is he number of hops beween source-desinaion pair. In OBS- F lex, we ake he offse ime o be P + H where P is he addiional propagaion delay beween source-desinaion pair. The minimum laency of burs in oher OBS schemes, is P + H which is he same if a burs is sen along wih conrol packe in opical packe swiching. The minimum laency in opical circui swiching is 3P + H. In OBS-F lex, he minimum laency of a burs is 2P + H. In oher OBS schemes, where wo way reservaion proocol is used he minimum laency is 2P + H. Thus, we can say ha he minimum laency of OBS-F lex is idenical o he OBS scheme wih wo way reservaion proocol. However, OBS-F lex is a one way reservaion proocol where each burs experiences an addiional delay of P unis. OBS-F lex is also unable o delay sensiive raffic. For delay sensiive raffic, he offse ime in OBS-F lex is aken o be H which is he same as ha in OBS. However, his offse can be made adapive o he needs of he applicaions. In OBS-F lex, if a conenion occurs and he siuaion is a CR one (as menioned in Secion 2.1) hen a burs is furher delayed for he conenion period. However, his delaying echnique of OBS-F lex is no applicable in case of delay sensiive raffic. For delay sensiive raffic if he required resource is no available wihin ha amoun of ime, he burs is dropped. Secondly, OBS-F lex differs from oher OBS schemes in he mehod adoped for conenion resoluion. In oher OBS schemes, he resource conflic is resolved on he basis of he reques prioriy and he ime insance for which he reques is made. In addiion o he above wo parameers, we ake burs size and he number of hops raversed o resolve conenion. A high prioriy reques is given a prioriy. However, for he same prioriy requess, he one ha has raversed he maximum number of hops, is acceped. For same prioriy and he number of hops raversed he one ha has larger burs size is acceped. For all he hree parameers having idenical values, he insance of reservaion is aken for conflic resoluion. Therefore, ies in 7

8 f pah r pah h T w s d rid m H Fig. 4. Fields of F-conrol packe f pah h T w s d rid r H Fig. 5. Fields of R-conrol packe conenion resoluion are resolved in he following order: prioriy, number of hops raversed, burs size and he delay. 2.3 Signaling Proocol and Implemenaion We use wo ypes of forward (F ) and reverse (R) conrol packes in OBS-F lex. In he following subsecions, we describe F and R conrol packes and OBS-F lex signaling proocol Conrol Packes F -conrol packe: An ingress rouer sends ou F -conrol packe when a burs arrives requesing for reservaion of resources a he inermediae core rouer. Resources are reserved using he delayed reservaion echnique, analogous o he one discussed in [20]. We skech he srucure of F -conrol packe in Fig. 4, and explain he fields of F -conrol packe below: f-pah is an explici forward pah ha F -conrol packe akes from he ingress o he egress rouer. Burs follows his pah once ransmied. r-pah is a reverse pah of he forward f-pah. For example, if f-pah is a b c d, hen r-pah is d c b a. h is he number of hops F -conrol packe has raversed/compleed. When a rouer receives F -conrol packe, i updaes value of h o h + 1; iniial value of h is se o zero. w is he wavelengh requesed for reservaion by he ingress rouer. s is he source/ingress rouer. d is he desinaion/egress rouer. Value of T indicaes he duraion of he conenion period. Iniially he value of T is se o zero by he ingress rouer. When a conenion occurs he value of T is se o he duraion of he conenion period. Value of m equals o one indicaes ha F -conrol packe is modified (iniially, value of m is se o zero by he ingress rouer). An inermediae node modifies F -conrol packe by seing he value of m o one. When value of m in F -conrol packe is se o one he resource reservaion is deferred for he conenion period 8

9 F Conrol Packe f pah r pah h T w s d rid m H f pah h T w s d rid r H R Conrol Packe Fig. 6. Formaion of R-conrol packe from F -conrol packe. r pah of F Conrol Packe 7 > 5 > 6 > 3 > 1 3 > 1 f pah of R Conrol Packe Fig. 7. Copying of f-pah o r-pah. menioned in he T field. rid is he reques ideniy. Value of r indicaes wheher he resources are o be rescheduled or released. H is he oal number of hops he reques rid has o raverse. When an inermediae core rouer receives F -conrol packe, one of he following hree possible siuaions arises : (i) NC, (ii) CR, or (iii) CNR as described in Secion 2.1. The acion aken by he core rouer depends on he value of m in F -conrol packe and one of he above hree siuaions. The inermediae core rouer updaes he value of h in F -conrol packe o h+1. The acions aken by he core rouer for boh he values of m and for all he hree possible siuaions are discussed below. In he following paragraphs we lis all he possibiliies depending on he values of differen fields in he conrol packes. Case I: When he value of m in F -conrol packe is equal o zero and one of he following siuaions occurs: 1. NC : Required resources are reserved a he core rouer and F -conrol packe is forwarded o he nex node in he pah. 2. CR : The following acions are aken a he core rouer for he conending reques: (i) he conenion period is deermined and he reservaion is deferred for ha period, (ii) he value of T in F -conrol packe is se o he above found conenion period, (iii) he value of m in F -conrol packe is se o one, (iv) R- conrol packe is formed (formaion of R-conrol packe is explained in he following paragraphs) and is sen o he ingress rouer s, and (v) F -conrol packe is sen o he nex node in he pah. 3. CNR: The following acions are aken a he core rouer for he conending reques: (i) R-conrol packe is formed and is sen oward he source s o release 9

10 he resources reserved for he conending reques, and (ii) F -conrol packe is dropped. Case II: When he value of m in F -conrol packe is equal o one, and one of he following siuaions occur: 1. NC : Defer he reservaion reques for a period as menioned in field T. For example, suppose resources are o be reserved a ime x and he value se in field T is x. Then he resources will be reserved a ime x + x. 2. CR : The following acions are aken a he core rouer for he conending reques: (i) R-conrol packe is formed and sen oward source s o release he resources reserved for he conending reques, and (ii) F -conrol packe is dropped. In our conenion resoluion scheme, we reschedule a reques only once. If he required resources are no available o he already rescheduled reques in any of he subsequen hop hen ha reques is dropped. 3. CNR : The following acions are aken a he core rouer for he conending reques: (i) R-conrol packe is formed and is sen oward he source s o release he resources reserved for he conending reques, and (ii) F -conrol packe is dropped. R-conrol packe : R-conrol packe is formed a he inermediae core rouer where he resource conflic has occurred. The srucure of R-conrol packe is shown in Fig. 5. Fields of R-conrol packe are explained below: f-pah is an explici pah ha R-conrol packe akes from he core rouer o he ingress rouer s. Semanics of he h, T, w, s, d and rid fields of R-conrol packe are idenical o ha of F -conrol packe. Value of r equal o zero indicaes resources reserved are o be rescheduled o a laer ime as specified in field T. For example, if he resources is reserved a a node from ime x and value of T field is x hen reservaion a he node is rescheduled o a ime x+x. A value equal o one indicaes ha he resources are o be released. R-conrol packe is formed from F -conrol packe and he formaion is explained in he following paragraphs. r-pah of F -conrol packe is copied ino f-pah of R-conrol packe and all he oher fields of F -conrol packe are copied o he corresponding fields of R-conrol packe (Fig. 6). Value of r is se o zero if resources are o be rescheduled, oherwise se o one if resources are o be released. Copying r-pah of F -conrol packe ino f-pah of R-conrol packe is illusraed in Fig. 7. In his illusraion, we have assumed ha resource conflic has occurred a core rouer 6. Remaining elemens of r-pah of F -conrol packe excluding node 6 are copied ino f-pah of R-conrol packe. R-conrol packe follows his f-pah o reach he ingress rouer 1 for whose reservaion reques, he resource conenion has occurred. Processing of R-conrol packe : On receiving R-conrol packe, a node updaes value of h in he conrol packe o h + 1. If value of H h 0 and value 10

11 of r is zero hen he resource reservaion for reques rid from ingress rouer s o egress rouer d is rescheduled for he ime as specified in T field else resources are released. If he node is ingress rouer s, R-conrol packe is dropped afer processing. If value of H h 0 hen R-conrol packe is forwarded o he nex node in f-pah else R-conrol packe is dropped a ha node. If a node on receiving R-conrol packe finds ha he requesed resources are subsequenly reserved by anoher reques han i does he following. If prioriy of he reques for which R-conrol packe is generaed is higher han he reques ha has subsequenly reserved he resources hen i de-reserves he reques and re-schedules he reques corresponding o R-conrol packe else R-conrol packe is dropped OBS-F lex Signaling Proocol The signaling proocol specifies he acions aken by boh ingress and core rouers. The following acions are aken a he ingress rouer: 1) F -conrol packe is sen ou when a burs arrives, 2) Burs is ransmied a he ime for which resources are reserved, and 3) On receiving R-conrol packe depending on he value of r-field of R-conrol packe resources are eiher released or reservaion is rescheduled o a ime as specified in he conrol packe. The acions aken a he core rouer are : 1) On receiving F -conrol packe i is processed as explained in previous subsecion, and 2) On receiving R-conrol packe i is processed as explained in previous subsecion. Summarizing, acions ha are needed o ransmi a burs are: (i) send F -conrol packe, (ii) process F -conrol packe, (iii) process R-conrol packe, if any, and (iv) ransmi a burs during he reserved ime. 2.4 Correcness of OBS-F lex In his subsecion, we show wih an illusraion ha OBS-F lex operaes as desired afer rescheduling of reservaion requess. We consider Fig. 8 for illusraion. In Fig. 8(a), burs b 1 has reserved resources for duraion 1 o 2 a node i and for duraion 5 o 6 a node j. Burs b 2 has reserved resources for duraion 3 o 4 a node i and for duraion 7 o 8 a node j. In he above scenario here is no conenion among burss for resources. So boh he burss are ransmied successfully. 11

12 Le us assume ha resource conenion has occurred a node j for burs b 1 and siuaion is CR as menioned in Secion 2.1. Le σ be he duraion of he conenion period. In OBS-F lex, R-conrol packe is sen from he conenion node, in his case node j, o he source o delay ransmission of burs b 1 for conenion period σ. On receiving R-conrol packe, node i will reschedule he ransmission of burs b 1. Reschedule of burs b 1 a node i may overlap wih burs b 2 which is already scheduled for ransmission a node i, depending on value of α as shown in Fig. 8(a) and duraion of conenion period σ. There are now wo possible cases: (i) σ < α, and (ii) σ > α. For σ < α reschedule of burs b 1 by node i is shown in Fig. 8(b). I can be seen from Fig. 8(b) ha burs b 1 afer reschedule does no overlap wih burs b 2. Only release of resources by burs b 1 and acquire of resources by burs b 2 is decreased from α unis in Fig. 8(a) o α σ unis in Fig. 8(b). Transmission of burs b 1 will be delayed for he conenion period. Thus, boh he burss are ransmied successfully. For σ > α, reschedule of burs b 1 will overlap wih burs b 2 as shown in Fig. 8(c). However, when burs b 2 arrives a node j i will conend wih burs b 1 for he period σ α as shown in Fig. 8(c). Therefore, node j will no schedule ransmission of burs b 2 a 7 bu o a laer ime a 7 + σ α. Node j also sends R-conrol packe o he source o delay ransmission of burs b 2 for conenion period σ α. Thus, node i reschedules he burss as shown in Fig. 8(d). As seen from Fig. 8(d), burss b 1 and b 2 do no overlap afer heir reschedule. Thus, hey are ransmied successfully. Therefore, i is illusraed wih he help of a diagram ha OBS-F lex operaes correcly afer rescheduling of he reservaion reques. 3 Channel Selecion Algorihms In his secion, we describe hree channel selecion algorihms, namely, (i) Leas Recenly Used (LRU), (ii) Firs Fi (FF), and (iii) Prioriy Se (PS) algorihms used for he proposed conenion resoluion scheme (Secion 2.2) for channel selecion. The channel selecion algorihms are run only a he edge rouers o find he daa channel for which reservaion reques is o be made and ransmi subsequenly daa burs. In LRU, a daa channel which is idle for he maximum duraion is seleced. In FF, daa channels are searched from he lowes index and he one which is available firs, is seleced. Consider Fig. 9, LRU channel selecion algorihm selecs daa channel 2 as i is idle for he maximum duraion where as FF channel selecion algorihm selecs daa channel 0. In PS approach, we decompose he se of daa channels, S, ino P subses, S i, of daa channels where P is he number of prioriy classes suppored. S = S 0 S 1 S P 1. A prioriy class i selecs daa channel from se S i. If no daa channel is available in se S i hen i selecs from se S i 1 and if no available hen from se 12

13 Time i j k b 1 b 2 α 1 + σ 2 +σ σ 6 + σ 7 8 i j k b 1 b 2 α σ (a) α σ (b) α 1 + σ 3 2 +σ σ σ 8 i j k b 1 b σ 2 +σ σ α (c) 3 + σ α 4 + σ α σ α i j k b 1 b 2 σ α (d) + σ 5 + σ σ α σ α 8 + σ α Fig. 8. Illusraion of he working of OBS-F lex: (a) no conenion, (b) no overlap due o rescheduling of burs b 1 [σ < α], (c) overlap due o rescheduling of burs b 1 [σ > α], and (d) no overlap due o rescheduling of burs b 1 and b 2 [σ > α]. S i 2. This process is ieraed ill he lowes prioriy se S 0 is searched. If no daa channel is available in se S 0 hen he burs is dropped a he ingress rouer. For prioriy class 0, if no daa channel is available in se S 0 hen he burs is dropped a he ingress rouer. To illusrae working of PS approach, we consider wo prioriy classes 0 and 1; class 1 has higher prioriy han class 0. We divide available daa channels as shown in Fig. 10 in wo ses S 0 = {0, 1} and S 1 = {2, 3}. Le class 1 burs arrives a a and i is o be ransmied a s afer base offse ime offse. Since all daa channels in se S 1 are busy a s, channel 0 from se S 0 is seleced. Inpus o he above channel selecion algorihms are burs arrival ime, a, and offse ime offse. PS algorihm has an addiional inpu of burs prioriy. Oupu of each of he algorihms is he seleced daa-channel dc. A negaive value of he oupu indicaes ha no daa channel is available. The funcion Channel Available Time in each of he algorihms reurns he available ime of each daa channel. Pseudocodes of LRU, FF and PS algorihms are included in Algo-1, Algo-2 and Algo-3, respecively. Inpu: a, offse 13

14 Time a s offse Fig. 9. Illusraion for selecion of daa channel in LRU and FF algorihms. 0 Time a offse s Fig. 10. Illusraion for selecion of daa channel in PS algorihm. Oupu: dc Algorihm: for i 0 o Num DaaChannel do av[i] Channel Available T ime() end of for loop. dc F ind LRUChannel(av, a, offse ) if dc is negaive hen drop he burs a he ingress rouer and repor no daa channel is available else repor daa channel dc. Algo-1: Leas Recenly Used Channel Selecion Algorihm Inpu: a, offse Oupu: dc Algorihm: for i 0 o Num DaaChannel do av[i] Channel Available T ime() end of for loop. dc F ind F F Channel(av, a, offse ) if dc is negaive hen drop he burs a he ingress rouer and repor no daa channel is available else repor daa channel dc. Algo-2: Firs Fi Channel Selecion Algorihm 14

15 Edge Rouer Core Rouer Fig. 11. Simulaed Burs Swiched Nework Inpu: a, offse, prioriy Oupu: dc Algorihm: for i 0 o Num DaaChannel do av[i] Channel Available T ime() end of for loop. dc F ind P SChannel(av, a, offse, prioriy) if dc is negaive hen drop he burs a he ingress rouer and repor no daa channel is available else repor daa channel dc. Algo-3: Prioriy Se Channel Selecion Algorihm 4 Simulaion Resuls We simulaed burs swiching nework as shown in Fig. 11; dark circles indicae edge rouers (ingress and egress rouer) and squares indicae core rouers. We made he following assumpions in he simulaion. The propagaion delay,, beween any wo adjacen nodes in he burs swiching nework is assumed o be 1ms. This assumpion is carried o simplify he simulaion ask; main aim of simulaion is o demonsrae effeciveness of OBS-F lex sraegy in reducing burs losses, herefore, his is a reasonable assumpion. Processing ime of he conrol packe a he rouer is assumed o be 2µs. We assume ha here is no wavelengh conversion and here exiss no opical buffer in he swich. For simpliciy and wihou loss of generaliy, we consider wo classes of raffic: class 0 (low prioriy) and class 1 (high prioriy). We generae high prioriy raffic wih a probabiliy of 0.4. Traffic is generaed a he edge rouer only, and he load is measured in Erlang. Performance of a nework is srongly influenced by he saisics/paerns of he arriving raffic. To sudy he effec of raffic on he nework performance, we consider he following wo cases: 15

16 LRU FF PS PPJET burs blocking probabiliy load (Erlang) Fig. 12. Overall burs loss in OBS-F lex wih differen channel selecion algorihms and PPJET. Pareo disribued burs size and Pareo disribued iner arrival of burs is considered. Case-I: Bursy Traffic: Pareo (α = 1.1) burs lengh disribuion and Pareo (α = 1.1) iner arrival ime disribuion, and Case-II: Poisson raffic: We consider wo sub-cases (a) Poisson disribued burs size and exponenial iner arrival ime disribuion, and (b) Fixed burs size and exponenial iner arrival ime disribuion. We compare he simulaion resuls obained wih he proposed scheme o ha of PPJET [7]. We consider burs blocking probabiliy as he performance meric for comparison. We have aken seven number of wavelenghs available on each link. 4.1 Bursy Traffic Traffic in he Inerne is repored o be bursy in naure [11]. We consider Pareo (α = 1.1) disribued burs lengh and Pareo (α = 1.1) disribued iner arrival ime. We include plos for burs loss under hree siuaions, namely, overall, high prioriy and low prioriy for OBS-F lex wih he hree proposed channel selecion algorihms, in Figs. 12, 13 and 14, respecively. We also include burs loss obained from PPJET in each of he graph for comparison. The overall burs loss increases wih increase in load as shown in Fig. 12. I is observed from he figure ha he overall burs loss in OBS-F lex is lower han ha in PPJET. Of he hree proposed channel selecion algorihms, LRU gives compara- 16

17 OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 13. High prioriy burs loss in OBS-F lex wih differen channel selecion algorihms and PPJET. Pareo disribued burs size and Pareo disribued iner arrival of burs is considered. ively lower and PS gives higher overall burs loss. The higher overall burs loss in PS is due o he higher low prioriy burs loss shown in Fig. 14. The plo for higher prioriy burs loss wih load is included in Fig. 13. I is observed from he figure ha OBS-F lex has lower higher-prioriy burs loss han in PPJET. Of he hree proposed channel selecion algorihms, PS algorihm has lower and FF algorihm has higher high-prioriy burs loss. Low high-prioriy burs loss in PS is due o he channel selecion sraegy ha is adoped in PS. In PS algorihm, a high prioriy raffic can selec a channel ha is marked for low prioriy raffic. Burs loss for low prioriy raffic is ploed in Fig. 14. I is observed from he figure ha OBS-F lex has lower burs loss for LRU and FF algorihms han in PPJET for all loads. However, OBS-F lex wih PS algorihm experiences higher low-prioriy burs loss han in PPJET a higher load. The increase in he burs loss a higher load is aribued o he daa channel consumed up by he higher prioriy raffic from hose marked for low prioriy raffic. From Figs. 12, 13 and 14, we can conclude ha OBS-F lex gives lower burs loss han ha in PPJET for bursy raffic. For bursy raffic, if a low overall burs loss is desired hen OBS-F lex wih LRU algorihm can be used. If low burs loss of high prioriy raffic is desired hen OBS-F lex wih PS algorihm is he obvious choice. 17

18 OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 14. Low prioriy burs loss in OBS-F lex wih differen channel selecion algorihms and PPJET. Pareo disribued burs size and Pareo disribued iner arrival of burs is considered. 4.2 Poisson Traffic Nex, we consider Poisson disribued burs size and exponenial iner arrival of burss. The mean exponenial iner arrival of burs is assumed o be 1ms. The plos for burs loss under hree siuaions overall, high prioriy and low for Poisson disribued burs size and exponenial iner arrival ime are included in Figs. 15, 16 and 17, respecively. Overall burs loss is ploed in Fig. 15. I is observed from he figure ha OBS-F lex wih LRU algorihm gives lower burs loss, for all loads, han in PPJET. OBS-F lex wih FF and PS algorihms gives higher burs loss in PPJET for all load. Of he hree channel selecion algorihms, LRU gives he lowes burs loss for all load. The higher prioriy burs loss is ploed in Fig. 16. From he figure, i is observed ha OBS-F lex wih PS algorihm has lower burs loss han in PPJET for all load. OBS-F lex wih LRU algorihm has almos he same burs loss as ha of PPJET for all load. Of he hree channel selecion algorihm, PS algorihm gives lower burs loss han wih LRU and FF algorihms. The lower burs loss in PS can be aribued o he selecion of daa channels which were marked for low prioriy raffic. Burs blocking probabiliy for lower prioriy raffic is ploed in Fig. 17. From he figure, i is observed ha OBS-F lex wih LRU algorihm gives he lower blocking han in PPJET for all load. OBS-F lex wih FF algorihm and PPJET scheme have almos he same blocking for all load. Of he hree channel selecion algorihms, PS has he higher blocking probabiliy; his is due o he selecion of daa channel 18

19 OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 15. Overall burs loss in OBS-F lex wih differen channel selecion algorihms and PPJET. Poisson disribued burs size and exponenial iner arrival of burs is considered OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 16. High prioriy burs loss in OBS-F lex wih differen channel selecion algorihms and PPJET. Poisson disribued burs size and exponenial iner arrival of burs is considered. by higher prioriy raffic from he daa channel marked for lower prioriy raffic. From Figs. 15, 16 and 17, we can conclude ha for raffic wih Poisson disribued burs size and exponenial iner arrival of burss, OBS-F lex gives lower burs loss han ha in PPJET. If low overall burs loss is desired hen OBS-F lex wih LRU algorihm may be he choice. If low high prioriy burs loss is desired hen OBS- F lex wih PS algorihm will be he choice. 19

20 OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 17. Low prioriy burs loss in OBS-F lex wih differen channel selecion algorihms and PPJET. Poisson disribued burs size and exponenial iner arrival of burs is considered. Nex, we consider fixed size burss and exponenial burs iner arrival ime of burss. We keep size of he burss o be 100µs. Plos for overall blocking probabiliy, blocking of high and low prioriy raffic are included in Figs. 18, 19 and 20, respecively. Fig. 18 shows he overall blocking probabiliy of OBS-F lex and PPJET wih load. I is observed from he figure ha OBS-F lex wih LRU algorihm gives lower blocking han in PPJET for all load. OBS-F lex wih PS algorihm has lower blocking han in PPJET a lower load hough a higher load he differences become marginal. Of he hree channel selecion algorihms, LRU algorihm gives lower blocking and FF gives higher blocking. The blocking probabiliy of high prioriy burss is ploed in Fig. 19. From he figure i is observed ha blocking probabiliy in OBS-F lex wih PS and LRU algorihms is lower han in PPJET for all load. Of he hree channel selecion algorihms, PS algorihm has lower blocking probabiliy; his is aribued o he selecion of he channel marked for low prioriy raffic. The blocking probabiliy of low prioriy burss is included in Fig. 20. From he figure, i is observed ha he OBS-F lex wih LRU algorihm gives lower blocking probabiliy han in PPJET for all load. I is also observed from he figure ha LRU algorihm gives lower blocking probabiliy han oher proposed channel selecion algorihms. From Figs. 18, 19 and 20, we can conclude ha for raffic wih fixed burs size and exponenial iner arrival of burs OBS-F lex gives lower burs loss han ha in PPJET. For low overall burs loss OBS-F lex wih LRU algorihm is he choice. If 20

21 OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 18. Overall burs loss in OBS-F lex wih differen channel selecion algorihm and PPJET. Fixed burs size and exponenial iner arrival of burs is considered OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 19. High prioriy burs loss in OBS-F lex wih differen channel selecion algorihms and PPJET. Fixed burs size and exponenial iner arrival of burs is considered. lower high prioriy burs loss is desired, OBS-F lex wih PS algorihm may be he choice. Comparing Figs. 15 and 18, i is observed for raffic wih fixed burs size ha he overall burs loss is lower han raffic wih Poisson disribued burs size. Similarly, from Figs. 16 and 19, i is observed ha he burs loss for high prioriy raffic for fixed size burs is lower han ha in Poisson disribued burs size. 21

22 OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 20. Low prioriy burs loss in OBS-F lex wih differen channel selecion algorihms and PPJET. Fixed burs size and exponenial iner arrival of burs is considered OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 21. Overall blocking probabiliy considering delay consrained raffic in OBS-F lex wih differen channel selecion algorihms and PPJET. Pareo disribued burs size and Pareo disribued iner arrival of burs is considered. 4.3 Delay Consrained Traffic In his secion, we simulae o sudy he effec of conenion resoluion scheme for delay consrained raffic. In our simulaion, we consider 20% of he oal raffic as delay consrained raffic. The offse ime for delay consrained raffic is aken o be H and for ohers P + H, as menioned in Secion 2.2. We include plos for 22

23 OBS-Flex (LRU) OBS-Flex (FF) OBS-Flex (PS) PPJET burs blocking probabiliy load (Erlang) Fig. 22. Blocking probabiliy of delay consrained raffic in OBS-F lex wih differen channel selecion algorihms and PPJET. Pareo disribued burs size and Pareo disribued iner arrival of burs is considered. overall blocking probabiliy, and blocking probabiliy for delay consrained raffic in Figs. 21 and 22, respecively. From he plos for overall blocking probabiliy (Fig. 21), i is observed ha OBS- F lex wih LRU algorihm gives lower overall blocking for all load. OBS-F lex wih FF algorihm and PPJET scheme have almos idenical blocking probabiliy a lower load. However, a higher load, blocking in OBS-F lex wih FF algorihm is marginally higher. OBS-F lex wih PS algorihm has lower overall blocking han in PPJET a lower load, however, a higher load blocking shows a marginally increasing rend. The blocking probabiliy for delay consrained raffic is shown in Fig. 22. I is observed ha OBS-F lex wih PS algorihm has lower blocking han in PPJET for all load. Blocking in OBS-F lex wih LRU algorihm and PPJET are almos idenical a all load. OBS-F lex wih FF algorihm has higher blocking han in PPJET a all load. From he experimens, we can conclude ha he proposed scheme OBS-F lex gives lower burs loss han PPJET if delay consrained raffic is also aken ino consideraion. If we need a lower overall burs loss hen we can use OBS-F lex wih LRU channel selecion algorihm and if lower loss of delay sensiive burs is desired hen OBS-F lex wih PS channel selecion algorihm is a superior choice. 23

24 5 Conclusions In his paper, we proposed a conenion resoluion scheme called OBS-F lex for OBS neworks. The scheme akes hree parameers, namely, prioriy, number of hops raversed, and burs size ino accoun o resolve conenion. The proposed scheme is adapable o boh prioriized and delay consrained raffic. We also proposed hree channel selecion algorihms called, Leas Recenly Used (LRU), Firs Fi (FF), and Prioriy Se (PS) algorihms o selec daa channel a he ingress rouer. We simulae OBS-F lex wih each of he hree channel selecion algorihms. We considered boh bursy and Poisson raffic in our simulaions. Simulaions were carried ou for boh prioriized raffic and delay consrained raffic. We observed ha LRU channel selecion algorihm gives lower overall burs loss for boh prioriized and delay consrained raffic. In addiion, PS channel selecion algorihm gives he lowes burs loss for prioriized and delay consrained raffic. We compared OBS-F lex wih anoher conenion resoluion scheme called PPJET. We found lower overall blocking probabiliy in OBS-F lex using LRU channel selecion algorihm han in PPJET scheme for all load in boh ypes of raffic ha we have considered. OBS-F lex using PS channel selecion algorihm gives lower blocking for high prioriy raffic and delay consrained raffic han PPJET. Thus, we can conclude ha irrespecive of he ype of raffic, if a lower overall burs loss is required han OBS-F lex wih LRU channel selecion algorihm can be used. If a low blocking of high prioriy raffic or delay consrained raffic is desired hen OBS-F lex wih PS channel selecion algorihm may be he choice. The lower blocking in OBS-F lex comes wih an addiional delay for prioriized raffic. In PPJET, an incoming burs is delayed by an amoun of ime which is equal o he oal processing ime of he conrol oken a each node. However, in OBS- F lex an addiional delay which is equal o he propagaion ime beween source o desinaion, is involved for he prioriized raffic. Fuure work includes assessing he effec of increased delays, in order o minimize burs losses due o conenion, on mulimedia applicaions involving delay consrained raffic. References [1] EURESCOM Projec P918-GI, Deliverable 2. hp:// [2] Noureddine Boudriga. Opical Burs Swiching Proocol for Supporing QoS and Adapive Rouing. Compuer Communicaions, 26: , [3] I. Chlamac, A. Ganz, and G. Karmi. Lighpah Communicaions: An Approach o High Bandwidh Opical WANs. IEEE Transacions on Communicaions, 40(7): , July [4] Michael Düser and Polina Bayvel. Analysis of a Dynamically Wavelengh- 24

25 Roued Opical Burs Swiched Nework. Journal of LighWave Technology, 20(4): , April [5] R. Dua and G.N Rouskas. A Survey of Virual Topology Design Algorihm for Wavelengh Roued Opical Neworks. Opical Nework Magazine, 1(1):73 89, January [6] Amaury Jourdan, Domonique Chiaroni, Emmanuel Doaro, Ger J. Eilenberger, Francesco Masei, and Monique Renaud. The Perspecive of Opical Packe Swiching in IP-Dominan Backbone and Meropolian Neworks. IEEE Communicaions Magazine, 39(3): , March [7] Ayman Kaheel and Hussein Alnuweiri. A Sric Prioriy Scheme for Qualiyof-Service Provisioning in Opical Burs Swiching Neworks. In Proceedings of he Eigh IEEE Inernaional Symposium on Compuers and Communicaion (ISCC 03), pages 16 21, [8] Hyun-Sook Kim, Su-Kyoung Lee, and Joo-Seok Song. Opical Burs Swiching wih Limied Deflecion Rouing Rules. IEICE Trans. Commun., E86- B(5), May [9] Su-Kyoung Lee, David Griffih, and Joo-Seok Song. Lambda GLSP seup wih QoS requiremen in opical Inerne. Compuer Communicaions, 26(6): , [10] Kejie Liu, Jason P. Jue, Gaoxi Xiao, Imrich Chlamac, and Timucin Ozugur. Inermediae Node Iniiaed Reservaion (IIR): A New Signaling Proocol for Wavelengh-Roued Neworks. IEEE Journal on Seleced Areas in Communicaions, 21(8): , Ocober [11] Vern Paxson and Sally Floyd. Wide Area Traffic: The Failure of Poisson Modeling. IEEE/ACM Transacion on Neworking, 3(3): , June [12] Rajiv Ramaswami and Kumar N. Sivarajan. Opical Neworks: A Pracical Perspecive. Morgan Kaufmann, [13] E. Varvarigos and V. Sharma. A Ready-To-Go Virual Circui Proocol: A Loss-Free Proocol for Muligigabi Nework Using FIFO Buffers. IEEE/ACM Transacions on Neworking, 5(5): , Ocober [14] Vinod M. Vokkarane, Qiong Zhang, Jason P. Jue, and Biao Chen. Generalized Burs Assembly and Scheduling Techniques for QoS Suppor in Opical Burs-Swiched Neworks. In Global Telecommunicaions Conference, 2002, GlOBECOM 02., volume 3, pages , November [15] J. Y. Wei, J. L. Pasor, R. S. Ramamurhy, and Y. Tsai. Jus-In-Time Opical Burs Swiching for Muli-Wavelengh Neworks. In Proceedings of 5h Inernaional Conference on Broadband Communicaions (BC 99), pages , [16] I. Widjaja. Performance Analysis of Burs Admission-Conrol Proocol. IEE Proceedings Communicaions, 142(1):7 14, February [17] Shun Yao, S.J. Ben Yoo, Biswanah Mukherjee, and Sudhir Dixi. All-Opical Packe Swiching for Meropolian Area Neworks: Opporuniies and Challenges. IEEE Communicaions Magazine, 39(3): , March [18] M. Yoo, M. Jeong, and C. Qiao. A High Speed Proocol for Bursy Traffic in Opical Neworks. In SPIE Proceedings All Opical Communicaion Sys- 25

26 ems: Archiecure, Conrol and Nework Issue, volume 3531, pages 79 80, November [19] M. Yoo and C. Qiao. Jus-enough-ime(JET): A High Speed Proocol for for Bursy Traffic in Opical Neworks. IEEE/LEOS Technologies for a Global Informaion Infrasrucure, pages 26 27, Augus [20] M. Yoo and C. Qiao. Opical Burs Swiching (OBS) - A New Paradigm for an Opical Inerne. Journal of High Speed Nework, 8(1):69 84, [21] M. Yoo, C. Qiao, and Sudhir Dixi. QoS Performance in IP over WDM Neworks. IEEE Journal on Seleced Areas in Communicaions, Special Issues on Proocols for Nex Generaion Opical Inerne, 18(10): , Ocober [22] Qiong Zhang, Vinod M. Vokkarane, Biao Chen, and Jason P. Jue. Early Drop and Wavelengh Grouping Schemes for Providing Absolue QoS Differeniaion in Opical Burs-Swiched Neworks. In Global Telecommunicaions Conference, 2003, GLOBECOM 02, IEEE, Volume: 5, 1-5 December 2003, pages , December [23] Qiong Zhang, Vinod M. Vokkarane, Biao Chen, and Jason P. Jue. Early Drop Scheme for Providing Absolue QoS Differeniaion in Opical Burs- Swiched Neworks. In Workshop on High Performance Swiching and Rouing, 2003, HPSR, pages , June

Efficient burst assembly algorithm with traffic prediction

Efficient burst assembly algorithm with traffic prediction Efficien burs assembly algorihm wih raffic predicion Mmoloki Mangwala, Boyce B. Sigweni and Bakhe M. Nleya Deparmen of compuer science Norh Wes Universiy, Privae Bag X2046, Mmabaho, 2735 Tel: +27 8 3892,

More information

Architectures for Resource Reservation Modules for Optical Burst Switching Core Nodes *

Architectures for Resource Reservation Modules for Optical Burst Switching Core Nodes * 4. ITG-Fachagung Phoonic Neworks, May 5. - 6., 2003, Leipzig, Germany Archiecures for Resource Reservaion Modules for Opical Burs Swiching Core Nodes * Sascha Junghans, Chrisoph M. Gauger Universiy of

More information

Performance Evaluation of a MAC Protocol for Radio over Fiber Wireless LAN operating in the 60-GHz band

Performance Evaluation of a MAC Protocol for Radio over Fiber Wireless LAN operating in the 60-GHz band Performance Evaluaion of a Proocol for Radio over Fiber Wireless LAN operaing in he -GHz band Hong Bong Kim, Adam Wolisz Telecommunicaion Neworks Group Deparmen of Elecrical and Compuer Engineering Technical

More information

ECMA st Edition / June Near Field Communication Wired Interface (NFC-WI)

ECMA st Edition / June Near Field Communication Wired Interface (NFC-WI) ECMA-373 1 s Ediion / June 2006 Near Field Communicaion Wired Inerface (NFC-WI) Sandard ECMA-373 1 s Ediion / June 2006 Near Field Communicaion Wired Inerface (NFC-WI) Ecma Inernaional Rue du Rhône 114

More information

A Novel Scheme to Reduce Burst-Loss and Provide QoS in Optical Burst Switching Networks

A Novel Scheme to Reduce Burst-Loss and Provide QoS in Optical Burst Switching Networks A Novel Scheme o Reuce Bur-Lo an Provie QoS in Opical Bur Swiching Nework Ahok K. Turuk an Rajeev Kumar Deparmen of Compuer Science an Engineering, Inian Iniue of Technology Kharagpur, Kharagpur, WB 72

More information

Traffic. analysis. The general setting. Example: buffer. Arrival Curves. Cumulative #bits: R(t), R*(t) Instantaneous speeds: r(t), r*(t)

Traffic. analysis. The general setting. Example: buffer. Arrival Curves. Cumulative #bits: R(t), R*(t) Instantaneous speeds: r(t), r*(t) The general seing Traffic Cumulaive #bis: R(), R*() Insananeous speeds: r(), r*() analysis R(): arrivals sysem R*(): deparures Lecure 7 2 Lecure 7 3 Example: buffer R() R*() bi rae c R() = #bis ha arrived

More information

Social-aware Dynamic Router Node Placement in Wireless Mesh Networks

Social-aware Dynamic Router Node Placement in Wireless Mesh Networks Social-aware Dynamic Rouer Node Placemen in Wireless Mesh Neworks Chun-Cheng Lin Pei-Tsung Tseng Ting-Yu Wu Der-Jiunn Deng ** Absrac The problem of dynamic rouer node placemen (dynrnp) in wireless mesh

More information

Investigation and Simulation Model Results of High Density Wireless Power Harvesting and Transfer Method

Investigation and Simulation Model Results of High Density Wireless Power Harvesting and Transfer Method Invesigaion and Simulaion Model Resuls of High Densiy Wireless Power Harvesing and Transfer Mehod Jaber A. Abu Qahouq, Senior Member, IEEE, and Zhigang Dang The Universiy of Alabama Deparmen of Elecrical

More information

Examination Mobile & Wireless Networking ( ) April 12,

Examination Mobile & Wireless Networking ( ) April 12, Page 1 of 5 Examinaion Mobile & Wireless Neworking (192620010) April 12, 2017 13.45 16.45 Noes: Only he overhead shees used in he course, 2 double-sided shees of noes (any fon size/densiy!), and a dicionary

More information

Receiver-Initiated vs. Short-Preamble Burst MAC Approaches for Multi-channel Wireless Sensor Networks

Receiver-Initiated vs. Short-Preamble Burst MAC Approaches for Multi-channel Wireless Sensor Networks Receiver-Iniiaed vs. Shor-Preamble Burs MAC Approaches for Muli-channel Wireless Sensor Neworks Crisina Cano, Boris Bellala, and Miquel Oliver Universia Pompeu Fabra, C/ Tànger 122-140, 08018 Barcelona,

More information

ECMA-373. Near Field Communication Wired Interface (NFC-WI) 2 nd Edition / June Reference number ECMA-123:2009

ECMA-373. Near Field Communication Wired Interface (NFC-WI) 2 nd Edition / June Reference number ECMA-123:2009 ECMA-373 2 nd Ediion / June 2012 Near Field Communicaion Wired Inerface (NFC-WI) Reference number ECMA-123:2009 Ecma Inernaional 2009 COPYRIGHT PROTECTED DOCUMENT Ecma Inernaional 2012 Conens Page 1 Scope...

More information

Dynamic Hybrid Topology Design for Integrated Traffic Support in WDM Mesh Networks 1 (Invited Paper)

Dynamic Hybrid Topology Design for Integrated Traffic Support in WDM Mesh Networks 1 (Invited Paper) Dynamic Hybrid Topology Design for Inegraed Traffic Suppor in WDM Mesh Neworks 1 (Invied Paper) Mina Youssef Elecrical and Compuer Engineering Kansas Sae Universiy Manhaan KS, 66502 USA Baek-Young Choi

More information

Pulse Train Controlled PCCM Buck-Boost Converter Ming Qina, Fangfang Lib

Pulse Train Controlled PCCM Buck-Boost Converter Ming Qina, Fangfang Lib 5h Inernaional Conference on Environmen, Maerials, Chemisry and Power Elecronics (EMCPE 016 Pulse Train Conrolled PCCM Buck-Boos Converer Ming Qina, Fangfang ib School of Elecrical Engineering, Zhengzhou

More information

P. Bruschi: Project guidelines PSM Project guidelines.

P. Bruschi: Project guidelines PSM Project guidelines. Projec guidelines. 1. Rules for he execuion of he projecs Projecs are opional. Their aim is o improve he sudens knowledge of he basic full-cusom design flow. The final score of he exam is no affeced by

More information

EXPERIMENT #9 FIBER OPTIC COMMUNICATIONS LINK

EXPERIMENT #9 FIBER OPTIC COMMUNICATIONS LINK EXPERIMENT #9 FIBER OPTIC COMMUNICATIONS LINK INTRODUCTION: Much of daa communicaions is concerned wih sending digial informaion hrough sysems ha normally only pass analog signals. A elephone line is such

More information

Direct Analysis of Wave Digital Network of Microstrip Structure with Step Discontinuities

Direct Analysis of Wave Digital Network of Microstrip Structure with Step Discontinuities Direc Analysis of Wave Digial Nework of Microsrip Srucure wih Sep Disconinuiies BILJANA P. SOŠIĆ Faculy of Elecronic Engineering Universiy of Niš Aleksandra Medvedeva 4, Niš SERBIA MIODRAG V. GMIROVIĆ

More information

Mobile Communications Chapter 3 : Media Access

Mobile Communications Chapter 3 : Media Access Moivaion Can we apply media access mehods from fixed neworks? Mobile Communicaions Chaper 3 : Media Access Moivaion SDMA, FDMA, TDMA Aloha Reservaion schemes Collision avoidance, MACA Polling CDMA SAMA

More information

Variation Aware Cross-Talk Aggressor Alignment by Mixed Integer Linear Programming

Variation Aware Cross-Talk Aggressor Alignment by Mixed Integer Linear Programming ariaion Aware Cross-alk Aggressor Alignmen by Mixed Ineger Linear Programming ladimir Zoloov IBM. J. Wason Research Cener, Yorkown Heighs, NY zoloov@us.ibm.com Peer Feldmann D. E. Shaw Research, New York,

More information

Performance Analysis of High-Rate Full-Diversity Space Time Frequency/Space Frequency Codes for Multiuser MIMO-OFDM

Performance Analysis of High-Rate Full-Diversity Space Time Frequency/Space Frequency Codes for Multiuser MIMO-OFDM Performance Analysis of High-Rae Full-Diversiy Space Time Frequency/Space Frequency Codes for Muliuser MIMO-OFDM R. SHELIM, M.A. MATIN AND A.U.ALAM Deparmen of Elecrical Engineering and Compuer Science

More information

Using Box-Jenkins Models to Forecast Mobile Cellular Subscription

Using Box-Jenkins Models to Forecast Mobile Cellular Subscription Open Journal of Saisics, 26, 6, 33-39 Published Online April 26 in SciRes. hp://www.scirp.org/journal/ojs hp://dx.doi.org/.4236/ojs.26.6226 Using Box-Jenkins Models o Forecas Mobile Cellular Subscripion

More information

Will my next WLAN work at 1 Gbps?

Will my next WLAN work at 1 Gbps? Will my nex WLAN work a 1 Gbps? Boris Bellala boris.bellala@upf.edu hp://www.dic.upf.edu/ bbellal/ Deparmen of Informaion and Communicaion Technologies (DTIC) Universia Pompeu Fabra (UPF) 2013 Ouline Moivaion

More information

Collaborative communication protocols for wireless sensor networks

Collaborative communication protocols for wireless sensor networks Collaboraive communicaion proocols for wireless sensor neworks Sefan Dulman, Lodewijk v. Hoesel, Tim Nieberg, Paul Havinga Faculy of Elecrical Engineering, Mahemaics and Compuer Science Universiy of Twene

More information

Delay improved Media Access Control for Passive Optical Networks

Delay improved Media Access Control for Passive Optical Networks Copyrigh Noice c 2015 IEEE. Personal use of his maerial is permied. However, permission o reprin/republish his maerial for adverising or promoional purposes or for creaing new collecive works for resale

More information

Lecture September 6, 2011

Lecture September 6, 2011 cs294-p29 Seminar on Algorihmic Game heory Sepember 6, 2011 Lecure Sepember 6, 2011 Lecurer: Chrisos H. Papadimiriou Scribes: Aloni Cohen and James Andrews 1 Game Represenaion 1.1 abular Form and he Problem

More information

DAGSTUHL SEMINAR EPIDEMIC ALGORITHMS AND PROCESSES: FROM THEORY TO APPLICATIONS

DAGSTUHL SEMINAR EPIDEMIC ALGORITHMS AND PROCESSES: FROM THEORY TO APPLICATIONS DAGSTUHL SEMINAR 342 EPIDEMIC ALGORITHMS AND PROCESSES: FROM THEORY TO APPLICATIONS A Sysems Perspecive Pascal Felber Pascal.Felber@unine.ch hp://iiun.unine.ch/! Gossip proocols Inroducion! Decenralized

More information

Performance Analysis of A Burst-Frame-Based MAC Protocol for Ultra-Wideband Ad Hoc Networks

Performance Analysis of A Burst-Frame-Based MAC Protocol for Ultra-Wideband Ad Hoc Networks Performance Analysis of A Burs-Frame-Based MAC Proocol for Ulra-Wideband Ad Hoc Neworks Kejie Lu, Dapeng Wu, Yuguang Fang Deparmen of Elecrical and Compuer Engineering Universiy Of Florida Gainesville,

More information

On the Efficiency of Shaping Live Video Streams

On the Efficiency of Shaping Live Video Streams On he Efficiency of Shaping Live Video Sreams György Dán and Vikoria Fodor KTH, Royal Insiue of Technology Elecrum 9, 6440 Kisa, Sweden gyuri,vikoria @i.kh.se Absrac In his work he efficiency of shaping

More information

Motion-blurred star image acquisition and restoration method based on the separable kernel Honglin Yuana, Fan Lib and Tao Yuc

Motion-blurred star image acquisition and restoration method based on the separable kernel Honglin Yuana, Fan Lib and Tao Yuc 5h Inernaional Conference on Advanced Maerials and Compuer Science (ICAMCS 206) Moion-blurred sar image acquisiion and resoraion mehod based on he separable kernel Honglin Yuana, Fan Lib and Tao Yuc Beihang

More information

Avoid link Breakage in On-Demand Ad-hoc Network Using Packet's Received Time Prediction

Avoid link Breakage in On-Demand Ad-hoc Network Using Packet's Received Time Prediction Avoid link Breakage in On-Demand Ad-hoc Nework Using acke's Received Time redicion Naif Alsharabi, Ya ping Lin, Waleed Rajeh College of Compuer & Communicaion Hunan Universiy ChangSha 182 Sharabi28@homail.com,

More information

Network Design and Optimization for Quality of Services in Wireless Local Area Networks using Multi-Objective Approach

Network Design and Optimization for Quality of Services in Wireless Local Area Networks using Multi-Objective Approach Chuima Prommak and Naruemon Waanapongsakorn Nework Design and Opimizaion for Qualiy of Services in Wireless Local Area Neworks using Muli-Objecive Approach CHUTIMA PROMMAK, NARUEMON WATTANAPONGSAKORN *

More information

Integrated Scheduling of Multimedia and Hard Real-Time Tasks

Integrated Scheduling of Multimedia and Hard Real-Time Tasks Inegraed Scheduling of Mulimedia and Hard Real-Time Tasks Hiroyuki Kaneko, John A. Sankovic, Subhabraa Sen and Krihi Ramamriham Compuer Science Deparmen Universiy of Massachuses LGRC, Box 346 Amhers MA

More information

An off-line multiprocessor real-time scheduling algorithm to reduce static energy consumption

An off-line multiprocessor real-time scheduling algorithm to reduce static energy consumption An off-line muliprocessor real-ime scheduling algorihm o reduce saic energy consumpion Firs Workshop on Highly-Reliable Power-Efficien Embedded Designs Shenzhen, China Vincen Legou, Mahieu Jan, Lauren

More information

Energy Efficient Data Fragmentation for Ubiquitous Computing

Energy Efficient Data Fragmentation for Ubiquitous Computing The Compuer Journal Advance Access published Sepember 1, 2013 The Briish Compuer Sociey 2013. All righs reserved. For Permissions, please email: journals.permissions@oup.com doi:10.1093/comjnl/bx080 Energy

More information

ECE-517 Reinforcement Learning in Artificial Intelligence

ECE-517 Reinforcement Learning in Artificial Intelligence ECE-517 Reinforcemen Learning in Arificial Inelligence Lecure 11: Temporal Difference Learning (con.), Eligibiliy Traces Ocober 8, 2015 Dr. Iamar Arel College of Engineering Deparmen of Elecrical Engineering

More information

Chapter 2. The Physical Layer

Chapter 2. The Physical Layer Chaper 2 The Physical Layer The Physical Layer Defines he mechanical, elecrical and iming inerfaces o he nework Transmission media - guided (copper and fiber opics) - wireless (radio erresrial) - saellie

More information

EE 330 Lecture 24. Amplification with Transistor Circuits Small Signal Modelling

EE 330 Lecture 24. Amplification with Transistor Circuits Small Signal Modelling EE 330 Lecure 24 Amplificaion wih Transisor Circuis Small Signal Modelling Review from las ime Area Comparison beween BJT and MOSFET BJT Area = 3600 l 2 n-channel MOSFET Area = 168 l 2 Area Raio = 21:1

More information

Anti-Jamming Schedules for Wireless Data Broadcast Systems

Anti-Jamming Schedules for Wireless Data Broadcast Systems Ani-Jamming Schedules for Wireless Daa Broadcas Sysems Paolo Codenoi 1, Alexander Sprinson, and Jehoshua Bruck Absrac Modern sociey is heavily dependen on wireless neworks for providing voice and daa communicaions.

More information

A WIDEBAND RADIO CHANNEL MODEL FOR SIMULATION OF CHAOTIC COMMUNICATION SYSTEMS

A WIDEBAND RADIO CHANNEL MODEL FOR SIMULATION OF CHAOTIC COMMUNICATION SYSTEMS A WIDEBAND RADIO CHANNEL MODEL FOR SIMULATION OF CHAOTIC COMMUNICATION SYSTEMS Kalle Rui, Mauri Honanen, Michael Hall, Timo Korhonen, Veio Porra Insiue of Radio Communicaions, Helsini Universiy of Technology

More information

On the Scalability of Ad Hoc Routing Protocols

On the Scalability of Ad Hoc Routing Protocols On he Scalabiliy of Ad Hoc Rouing Proocols César A. Saniváñez Bruce McDonald Ioannis Savrakakis Ram Ramanahan Inerne. Research Dep. Elec. & Comp. Eng. Dep. Dep. of Informaics Inerne. Research Dep. BBN

More information

Two-area Load Frequency Control using IP Controller Tuned Based on Harmony Search

Two-area Load Frequency Control using IP Controller Tuned Based on Harmony Search Research Journal of Applied Sciences, Engineering and Technology 3(12): 1391-1395, 211 ISSN: 24-7467 Maxwell Scienific Organizaion, 211 Submied: July 22, 211 Acceped: Sepember 18, 211 Published: December

More information

A Segmentation Method for Uneven Illumination Particle Images

A Segmentation Method for Uneven Illumination Particle Images Research Journal of Applied Sciences, Engineering and Technology 5(4): 1284-1289, 2013 ISSN: 2040-7459; e-issn: 2040-7467 Maxwell Scienific Organizaion, 2013 Submied: July 17, 2012 Acceped: Augus 15, 2012

More information

Foreign Fiber Image Segmentation Based on Maximum Entropy and Genetic Algorithm

Foreign Fiber Image Segmentation Based on Maximum Entropy and Genetic Algorithm Journal of Compuer and Communicaions, 215, 3, 1-7 Published Online November 215 in SciRes. hp://www.scirp.org/journal/jcc hp://dx.doi.org/1.4236/jcc.215.3111 Foreign Fiber Image Segmenaion Based on Maximum

More information

4.5 Biasing in BJT Amplifier Circuits

4.5 Biasing in BJT Amplifier Circuits 4/5/011 secion 4_5 Biasing in MOS Amplifier Circuis 1/ 4.5 Biasing in BJT Amplifier Circuis eading Assignmen: 8086 Now le s examine how we C bias MOSFETs amplifiers! f we don bias properly, disorion can

More information

Signaling Cost Analysis for Handoff Decision Algorithms in Femtocell Networks

Signaling Cost Analysis for Handoff Decision Algorithms in Femtocell Networks Signaling Cos Analysis for Hoff Decision Algorihms in Femocell Neworks Wahida Nasrin Jiang Xie The Universiy of Norh Carolina a Charloe Charloe, NC 28223-1 Email: {wnasrin, Linda.Xie}uncc.edu Absrac Femocells

More information

DS CDMA Scheme for WATM with Errors and Erasures Decoding

DS CDMA Scheme for WATM with Errors and Erasures Decoding DS CDMA Scheme for WATM wih Errors and Erasures Decoding Beaa J. Wysocki*, Hans-Jürgen Zepernick*, and Tadeusz A. Wysocki** * Ausralian Telecommunicaions Research Insiue Curin Universiy of Technology GPO

More information

Starvation Mitigation Through Multi-Channel Coordination in CSMA Multi-hop Wireless Networks

Starvation Mitigation Through Multi-Channel Coordination in CSMA Multi-hop Wireless Networks Sarvaion Miigaion Through Muli-Channel Coordinaion in CSMA Muli-hop Wireless Neworks Jingpu Shi, Theodoros Salonidis, and Edward W. Knighly Deparmen of Elecrical and Compuer Engineering Rice Universiy,

More information

The design of an improved matched filter in DSSS-GMSK system

The design of an improved matched filter in DSSS-GMSK system Journal of Physics: Conference Series PAPER OPEN ACCESS The design of an improved mached filer in DSSS-GMSK sysem To cie his aricle: Mao Wei-ong e al 16 J. Phys.: Conf. Ser. 679 1 View he aricle online

More information

Jitter Analysis of Current-Mode Logic Frequency Dividers

Jitter Analysis of Current-Mode Logic Frequency Dividers Universiy of California a Davis, Deparmen of Elecrical and Compuer Engineering Marko Aleksic Jier Analysis of Curren-Mode Logic Frequency Dividers Ph.D. Research Proposal able of Conens Secion : Inroducion

More information

Key Issue. 3. Media Access. Hidden and Exposed Terminals. Near and Far Terminals. FDD/FDMA General Scheme, Example GSM. Access Methods SDMA/FDMA/TDMA

Key Issue. 3. Media Access. Hidden and Exposed Terminals. Near and Far Terminals. FDD/FDMA General Scheme, Example GSM. Access Methods SDMA/FDMA/TDMA Key Issue Can we apply media access mehods from fixed neworks? 3. Media Access SDMA, FDMA, TDMA Aloha and Reservaion Schemes Avoidance and Polling MACA, CDMA, SAMA Example CSMA/CD: Carrier Sense Muliple

More information

A new image security system based on cellular automata and chaotic systems

A new image security system based on cellular automata and chaotic systems A new image securiy sysem based on cellular auomaa and chaoic sysems Weinan Wang Jan 2013 Absrac A novel image encrypion scheme based on Cellular Auomaa and chaoic sysem is proposed in his paper. The suggesed

More information

A New Voltage Sag and Swell Compensator Switched by Hysteresis Voltage Control Method

A New Voltage Sag and Swell Compensator Switched by Hysteresis Voltage Control Method Proceedings of he 8h WSEAS Inernaional Conference on ELECTRIC POWER SYSTEMS, HIGH VOLTAGES, ELECTRIC MACHINES (POWER '8) A New Volage Sag and Swell Compensaor Swiched by Hyseresis Volage Conrol Mehod AMIR

More information

State Space Modeling, Simulation and Comparative Analysis of a conceptualised Electrical Control Signal Transmission Cable for ROVs

State Space Modeling, Simulation and Comparative Analysis of a conceptualised Electrical Control Signal Transmission Cable for ROVs Sae Space Modeling, Simulaion and omparaive Analysis of a concepualised Elecrical onrol Signal ransmission able for ROVs James Naganda, Deparmen of Elecronic Engineering, Konkuk Universiy, Seoul, Korea

More information

Transmit Beamforming with Reduced Feedback Information in OFDM Based Wireless Systems

Transmit Beamforming with Reduced Feedback Information in OFDM Based Wireless Systems Transmi Beamforming wih educed Feedback Informaion in OFDM Based Wireless Sysems Seung-Hyeon Yang, Jae-Yun Ko, and Yong-Hwan Lee School of Elecrical Engineering and INMC, Seoul Naional Universiy Kwanak

More information

Communication Systems. Department of Electronics and Electrical Engineering

Communication Systems. Department of Electronics and Electrical Engineering COMM 704: Communicaion Lecure : Analog Mulipliers Dr Mohamed Abd El Ghany Dr. Mohamed Abd El Ghany, Mohamed.abdel-ghany@guc.edu.eg nroducion Nonlinear operaions on coninuous-valued analog signals are ofen

More information

Specification of a Collision-Free CSMA MAC Protocol for Wireless LANs: the CANlike protocol

Specification of a Collision-Free CSMA MAC Protocol for Wireless LANs: the CANlike protocol Specificaion of a Collision-Free CSMA MAC Proocol for Wireless LANs: he CANlike proocol Guy Juanole, Xuan Hung Nguyen, Gérard Mouney To cie his version: Guy Juanole, Xuan Hung Nguyen, Gérard Mouney. Specificaion

More information

Optical phase locked loop for transparent inter-satellite communications

Optical phase locked loop for transparent inter-satellite communications Opical phase locked loop for ransparen iner-saellie communicaions F. Herzog 1, K. Kudielka 2,D.Erni 1 and W. Bächold 1 1 Communicaion Phoonics Group, Laboraory for Elecromagneic Fields and Microwave Elecronics,

More information

Installing remote sites using TCP/IP

Installing remote sites using TCP/IP v dc Keypad from nework Whie/ 3 Whie/ 4 v dc Keypad from nework Whie/ 3 Whie/ 4 v dc Keypad from nework Whie/ 3 Whie/ 4 +v pu +v pu +v pu v dc Keypad from nework Whie/ 3 Whie/ 4 v dc Keypad from nework

More information

Lecture 4. EITN Chapter 12, 13 Modulation and diversity. Antenna noise is usually given as a noise temperature!

Lecture 4. EITN Chapter 12, 13 Modulation and diversity. Antenna noise is usually given as a noise temperature! Lecure 4 EITN75 2018 Chaper 12, 13 Modulaion and diversiy Receiver noise: repeiion Anenna noise is usually given as a noise emperaure! Noise facors or noise figures of differen sysem componens are deermined

More information

Surveillance System with Object-Aware Video Transcoder

Surveillance System with Object-Aware Video Transcoder MITSUBISHI ELECTRIC RESEARCH LABORATORIES hp://www.merl.com Surveillance Sysem wih Objec-Aware Video Transcoder Toshihiko Haa, Naoki Kuwahara, Toshiharu Nozawa, Derek Schwenke, Anhony Vero TR2005-115 April

More information

Network Performance Metrics

Network Performance Metrics Fundamenals of Compuer Neworks ECE 478/578 Lecure #3 Insrucor: Loukas Lazos Dep of Elecrical and Compuer Engineering Universiy of rizona Nework Performance Merics andwidh moun of daa ransmied per uni of

More information

Proceedings of International Conference on Mechanical, Electrical and Medical Intelligent System 2017

Proceedings of International Conference on Mechanical, Electrical and Medical Intelligent System 2017 on Mechanical, Elecrical and Medical Inelligen Sysem 7 Consan On-ime Conrolled Four-phase Buck Converer via Saw-oohwave Circui and is Elemen Sensiiviy Yi Xiong a, Koyo Asaishi b, Nasuko Miki c, Yifei Sun

More information

PREVENTIVE MAINTENANCE WITH IMPERFECT REPAIRS OF VEHICLES

PREVENTIVE MAINTENANCE WITH IMPERFECT REPAIRS OF VEHICLES Journal of KONES Powerrain and Transpor, Vol.14, No. 3 2007 PEVENTIVE MAINTENANCE WITH IMPEFECT EPAIS OF VEHICLES Józef Okulewicz, Tadeusz Salamonowicz Warsaw Universiy of Technology Faculy of Transpor

More information

Multiuser Interference in TH-UWB

Multiuser Interference in TH-UWB Ouline Roman Merz, Cyril Boeron, Pierre-André Farine Insiue of Microechnology Universiy of Neuchâel 2000 Neuchâel Workshop on UWB for Sensor Neworks, 2005 Ouline Ouline 1 Inroducion Moivaions and Goals

More information

AN303 APPLICATION NOTE

AN303 APPLICATION NOTE AN303 APPLICATION NOTE LATCHING CURRENT INTRODUCTION An imporan problem concerning he uilizaion of componens such as hyrisors or riacs is he holding of he componen in he conducing sae afer he rigger curren

More information

HIGH THROUGHPUT EVALUATION OF SHA-1 IMPLEMENTATION USING UNFOLDING TRANSFORMATION

HIGH THROUGHPUT EVALUATION OF SHA-1 IMPLEMENTATION USING UNFOLDING TRANSFORMATION VOL., NO. 5, MARCH 26 ISSN 89-668 26-26 Asian Research Publishing Nework (ARPN). All righs reserved. HIGH THROUGHPUT EVALUATION OF SHA- IMPLEMENTATION USING UNFOLDING TRANSFORMATION Shamsiah Bini Suhaili

More information

ACTIVITY BASED COSTING FOR MARITIME ENTERPRISES

ACTIVITY BASED COSTING FOR MARITIME ENTERPRISES ACTIVITY BASED COSTING FOR MARITIME ENTERPRISES 1, a 2, b 3, c 4, c Sualp Omer Urkmez David Sockon Reza Ziarai Erdem Bilgili a, b De Monfor Universiy, UK, c TUDEV, Insiue of Mariime Sudies, Turkey 1 sualp@furrans.com.r

More information

Forward-link Throughput Optimization for Wireless Cellular Packet Data Networks

Forward-link Throughput Optimization for Wireless Cellular Packet Data Networks JCS& Vol. 11 No. Ocober 11 Forward-link hroughpu Opimizaion for Wireless Cellular Packe Daa Neworks Moses E. Ekpenyong Deparmen of Compuer Science Universiy of Uyo, PMB. 117, 3, Uyo, NIGERIA (ekpenyong_moses@yahoo.com,

More information

Memorandum on Impulse Winding Tester

Memorandum on Impulse Winding Tester Memorandum on Impulse Winding Teser. Esimaion of Inducance by Impulse Response When he volage response is observed afer connecing an elecric charge sored up in he capaciy C o he coil L (including he inside

More information

Multiple Load-Source Integration in a Multilevel Modular Capacitor Clamped DC-DC Converter Featuring Fault Tolerant Capability

Multiple Load-Source Integration in a Multilevel Modular Capacitor Clamped DC-DC Converter Featuring Fault Tolerant Capability Muliple Load-Source Inegraion in a Mulilevel Modular Capacior Clamped DC-DC Converer Feauring Faul Toleran Capabiliy Faisal H. Khan, Leon M. Tolber The Universiy of Tennessee Elecrical and Compuer Engineering

More information

Accurate Tunable-Gain 1/x Circuit Using Capacitor Charging Scheme

Accurate Tunable-Gain 1/x Circuit Using Capacitor Charging Scheme Accurae Tunable-Gain 1/x Circui Using Capacior Charging Scheme Byung-Do Yang and Seo Weon Heo This paper proposes an accurae unable-gain 1/x circui. The oupu volage of he 1/x circui is generaed by using

More information

PRM and VTM Parallel Array Operation

PRM and VTM Parallel Array Operation APPLICATION NOTE AN:002 M and V Parallel Array Operaion Joe Aguilar VI Chip Applicaions Engineering Conens Page Inroducion 1 High-Level Guidelines 1 Sizing he Resisor 4 Arrays of Six or More Ms 5 Sysem

More information

Automatic Power Factor Control Using Pic Microcontroller

Automatic Power Factor Control Using Pic Microcontroller IDL - Inernaional Digial Library Of Available a:www.dbpublicaions.org 8 h Naional Conference on Advanced Techniques in Elecrical and Elecronics Engineering Inernaional e-journal For Technology And Research-2017

More information

Development of Temporary Ground Wire Detection Device

Development of Temporary Ground Wire Detection Device Inernaional Journal of Smar Grid and Clean Energy Developmen of Temporary Ground Wire Deecion Device Jing Jiang* and Tao Yu a Elecric Power College, Souh China Universiy of Technology, Guangzhou 5164,

More information

Chapter 2 Introduction: From Phase-Locked Loop to Costas Loop

Chapter 2 Introduction: From Phase-Locked Loop to Costas Loop Chaper 2 Inroducion: From Phase-Locked Loop o Cosas Loop The Cosas loop can be considered an exended version of he phase-locked loop (PLL). The PLL has been invened in 932 by French engineer Henri de Belleszice

More information

Shortest and Efficient Multipath Routing in Mobile ad hoc Network (MANET)

Shortest and Efficient Multipath Routing in Mobile ad hoc Network (MANET) Indian Journal of Science and Technology, Vol 9(45), DOI: 10.17485/ijs/2016/v9i45/102597, December 2016 ISSN (Prin) : 0974-6846 ISSN (Online) : 0974-5645 Shores and Efficien Mulipah Rouing in Mobile ad

More information

Mode-Changes in COTS Time-Triggered Network Hardware without Online Reconfiguration

Mode-Changes in COTS Time-Triggered Network Hardware without Online Reconfiguration Mode-Changes in COTS Time-Triggered Nework Hardware wihou Online Reconfiguraion Florian Heilmann, Ali Syed and Gerhard Fohler Chair of Real-ime Sysems Technische Universiä Kaiserslauern, Germany {heilmann,syed,fohler}@ei.uni-kl.de

More information

Process-Stacking Multiplexing Access for 60 GHz Millimeter-Wave WPANs

Process-Stacking Multiplexing Access for 60 GHz Millimeter-Wave WPANs Process-Sacking Muliplexing Access for 6 GHz Millimeer-Wave WPANs Claudio Esevez, David Fuenealba, Aravind Kailas Deparmen of Elecrical Engineering, Universidad de Chile, Saniago, Chile {cesevez, dfuenea}@ing.uchile.cl

More information

The Influence of Propagation Environment in a Live GSM Network

The Influence of Propagation Environment in a Live GSM Network The Influence of Propagaion Environmen in a ive GSM Nework Yu-Huei Tseng, Wen-Shyang Hwang, *Ce-Kuen Shieh Deparmen of Elecrical Engineering, Naional Kaohsiung Universiy of Applied Sciences, Kaohsiung,

More information

Pointwise Image Operations

Pointwise Image Operations Poinwise Image Operaions Binary Image Analysis Jana Kosecka hp://cs.gmu.edu/~kosecka/cs482.hml - Lookup able mach image inensiy o he displayed brighness values Manipulaion of he lookup able differen Visual

More information

4 20mA Interface-IC AM462 for industrial µ-processor applications

4 20mA Interface-IC AM462 for industrial µ-processor applications Because of he grea number of indusrial buses now available he majoriy of indusrial measuremen echnology applicaions sill calls for he sandard analog curren nework. The reason for his lies in he fac ha

More information

Photo Modules for PCM Remote Control Systems

Photo Modules for PCM Remote Control Systems Phoo Modules for PCM Remoe Conrol Sysems Descripion The HS38B series are miniaurized receivers for infrared remoe conrol sysems. PIN diode and preamplifier are assembled on lead frame, he epoxy package

More information

A New Design of Private Information Retrieval for Storage Constrained Databases

A New Design of Private Information Retrieval for Storage Constrained Databases A ew Design of Privae Informaion Rerieval for Sorage Consrained Daabases icholas Woolsey, Rong-Rong Chen, and Mingyue Ji Deparmen of Elecrical and Compuer Engineering, Universiy of Uah Sal Lake Ciy, UT,

More information

Lecture 11. Digital Transmission Fundamentals

Lecture 11. Digital Transmission Fundamentals CS4/MSc Compuer Neworking Lecure 11 Digial Transmission Fundamenals Compuer Neworking, Copyrigh Universiy of Edinburgh 2005 Digial Transmission Fundamenals Neworks consruced ou of Links or ransmission

More information

A NEW DUAL-POLARIZED HORN ANTENNA EXCITED BY A GAP-FED SQUARE PATCH

A NEW DUAL-POLARIZED HORN ANTENNA EXCITED BY A GAP-FED SQUARE PATCH Progress In Elecromagneics Research Leers, Vol. 21, 129 137, 2011 A NEW DUAL-POLARIZED HORN ANTENNA EXCITED BY A GAP-FED SQUARE PATCH S. Ononchimeg, G. Ogonbaaar, J.-H. Bang, and B.-C. Ahn Applied Elecromagneics

More information

Auto-Tuning of PID Controllers via Extremum Seeking

Auto-Tuning of PID Controllers via Extremum Seeking 25 American Conrol Conference June 8-, 25. Porland, OR, USA ThA7.2 Auo-Tuning of PID Conrollers via Exremum Seeking Nick illingsworh* and Miroslav rsić Deparmen of Mechanical and Aerospace Engineering

More information

Communications II Lecture 7: Performance of digital modulation

Communications II Lecture 7: Performance of digital modulation Communicaions II Lecure 7: Performance of digial modulaion Professor Kin K. Leung EEE and Compuing Deparmens Imperial College London Copyrigh reserved Ouline Digial modulaion and demodulaion Error probabiliy

More information

How to Shorten First Order Unit Testing Time. Piotr Mróz 1

How to Shorten First Order Unit Testing Time. Piotr Mróz 1 How o Shoren Firs Order Uni Tesing Time Pior Mróz 1 1 Universiy of Zielona Góra, Faculy of Elecrical Engineering, Compuer Science and Telecommunicaions, ul. Podgórna 5, 65-246, Zielona Góra, Poland, phone

More information

A Novel Routing Algorithm for Power Line Communication over a Low-voltage Distribution Network in a Smart Grid

A Novel Routing Algorithm for Power Line Communication over a Low-voltage Distribution Network in a Smart Grid Energies 0, 6, 4-48; doi:0.90/en604 Aricle OPEN ACCESS energies ISSN 996-07 www.mdpi.com/journal/energies A Novel Rouing Algorihm for Power Line Communicaion over a Low-volage Disribuion Nework in a Smar

More information

Chapter 2 Summary: Continuous-Wave Modulation. Belkacem Derras

Chapter 2 Summary: Continuous-Wave Modulation. Belkacem Derras ECEN 44 Communicaion Theory Chaper Summary: Coninuous-Wave Modulaion.1 Modulaion Modulaion is a process in which a parameer of a carrier waveform is varied in accordance wih a given message (baseband)

More information

Adaptive Antenna Array for Reliable OFDM Transmission

Adaptive Antenna Array for Reliable OFDM Transmission Adapive Anenna Array for Reliable OFDM Transmission Shinsuke Hara Deparmen of Elecronics, Informaion Sysem And Energy Engineering, Graduae School of Engineering, Osaka Universiy Conens of This Presenaion

More information

5 Spatial Relations on Lines

5 Spatial Relations on Lines 5 Spaial Relaions on Lines There are number of useful problems ha can be solved wih he basic consrucion echniques developed hus far. We now look a cerain problems, which involve spaial relaionships beween

More information

On Eliminating the Exposed Terminal Problem Using Signature Detection

On Eliminating the Exposed Terminal Problem Using Signature Detection 1 On Eliminaing he Exposed Terminal Problem Using Signaure Deecion Junmei Yao, Tao Xiong, Jin Zhang and Wei Lou Deparmen of Compuing, The Hong Kong Polyechnic Universiy, Hong Kong {csjyao, csxiong, csjzhang,

More information

Demodulation Based Testing of Off Chip Driver Performance

Demodulation Based Testing of Off Chip Driver Performance Demodulaion Based Tesing of Off Driver Performance Wilfried Daehn Hochschule Magdeburg-Sendahl Fachbereich Elekroechnik Posfach 368 39 Magdeburg Phone: ++49 39 886 4673 Fa: ++49 39 886 426 Email: wilfried.daehn@compuer.org

More information

TSOP12.. IR Receiver Modules for Remote Control Systems VISHAY. Vishay Semiconductors

TSOP12.. IR Receiver Modules for Remote Control Systems VISHAY. Vishay Semiconductors TSOP1.. IR Receiver Modules for Remoe Conrol Sysems Descripion The TSOP1.. - series are miniaurized receivers for infrared remoe conrol sysems. PIN diode and preamplifier are assembled on lead frame, he

More information

Fuzzy Inference Model for Learning from Experiences and Its Application to Robot Navigation

Fuzzy Inference Model for Learning from Experiences and Its Application to Robot Navigation Fuzzy Inference Model for Learning from Experiences and Is Applicaion o Robo Navigaion Manabu Gouko, Yoshihiro Sugaya and Hiroomo Aso Deparmen of Elecrical and Communicaion Engineering, Graduae School

More information

Radio Resource Management in Beyond 3G Systems

Radio Resource Management in Beyond 3G Systems IEEE MELECON 2006, May 16-19, Benalmádena (Málaga), Spain Radio Resource Managemen in Beyond 3G Sysems Absrac.- Beyond 3G sysems is usually he erm used o refer o he new scenarios in he wireless arena where

More information

Efficient Data Encoding and Decoding for Network-On-Chip Application

Efficient Data Encoding and Decoding for Network-On-Chip Application Inernaional Journal of Curren Engineering and Technology E-ISSN 2277 4106, P-ISSN 2347 5161 2015 INPRESSCO, All Righs Reserved Available a hp://inpressco.com/caegory/ijce Research Aricle Efficien Daa Encoding

More information

A Perspective on Radio Resource Management in B3G

A Perspective on Radio Resource Management in B3G A Perspecive on Radio Resource Managemen in B3G Oriol Sallen Universia Poliècnica de Caalunya (UPC) E-mail: sallen@sc.upc.edu Absrac.- Beyond 3G usually refers o heerogeneous scenarios where differen Radio

More information

HS0038B5. IR Receiver Modules for Remote Control Systems. Vishay Semiconductors

HS0038B5. IR Receiver Modules for Remote Control Systems. Vishay Semiconductors IR Receiver Modules for Remoe Conrol Sysems Descripion The - series are miniaurized receivers for infrared remoe conrol sysems. PIN diode and preamplifier are assembled on lead frame, he epoxy package

More information

DURING last decade, many communication technologies

DURING last decade, many communication technologies 1 TeamPhone: Neworking Smarphones for isaser Recovery Zongqing Lu, Guohong ao, and Thomas La Pora and suppors daa communicaions among rescue workers in infrasrucure-consrained and infrasrucure-less scenarios.

More information