Hybrid Job Scheduling Mechanism Using a Backfill-based Multi-queue Strategy in Distributed Grid Computing

Size: px
Start display at page:

Download "Hybrid Job Scheduling Mechanism Using a Backfill-based Multi-queue Strategy in Distributed Grid Computing"

Transcription

1 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September Hybrd Job Schedulng Mechansm Usng a Backfll-based Mult-queue Strategy n Dstrbuted Grd Computng Ken Park Changhoon Kang and Sungsook Km Aou Unversty, Suwon, Gyeongg-do, South Korea Gangdong College, Icheon, Gyeongg-do, South Korea Anyang Unversty, Anyang, Gyeongg-do, South Korea Summary In a dstrbuted computng envronment, effectve ob schedulng s a crtcal challenge. In ths paper, a hybrd ob schedulng mechansm s proposed that consders both the meta-schedulng scheme for dstrbutng obs to overall nodes and the local ob schedulng scheme for assgnng obs wthn a local node at a smultaneous tme. Dependng on the number of requred processors and the expected executon tme of obs, the order of prortes s establshed. Jobs wth hgh prorty are then allocated to a ob queue whereas those wth low prorty are assgned to a backfll queue wth remote obs that are sent from other nodes. Experments conducted to evaluate the performance of the proposed mechansm show that the utlzaton of a grd computng system becomes more effcent and watng tmes are consderably reduced. Key words: Meta-schedulng, Job schedulng, Backfll schedulng, Dstrbuted Grd computng 1. Introducton As modern technologes advance, hgh-performance computng power s more desred to solve complex and large-scale computaton problems. To meet ths need, a sgnfcant amount of research has been conducted to explot grd computng, whch bnds up heterogeneous computng resources nto a sngle hgh performance system [1]. Recently, a varety of grd tools have been developed to support the basc grd servces such as 1) grd resource dscovery servce to locate avalable resources that are connected to grds [2], 2) grd ob schedulng servce to dstrbute and schedule obs for an effcent processng, 3) grd securty servce for the system protecton, and 4) grd accountng servce to assess costs ncurred by usng the computng resources [3]. In addton, the research on the grd mddleware lke Globus [4] and Legon [5] have made mplementatons of grd applcaton programs much easer. In a grd computng envronment, effectve ob schedulng s a crtcal challenge. The functons of the scheduler n a grd computng envronment can bascally be classfed nto two categores that nclude: 1) meta-schedulng functon regardng the dstrbuted processng of the complex problem among nodes and 2) the local ob schedulng functon wthn a local node [6]. The metaschedulng can mprove system utlzaton by dspatchng obs to local nodes n an effcent manner, and the local ob schedulng can reduce the slowdown of ob processng by schedulng obs wthn a node. So far, a maorty of researchers have consdered these two types of functons separately. Therefore, t s desrable to develop a hybrd ob schedulng mechansm that consders both the metaschedulng scheme and the local ob schedulng scheme n order to mprove the overall performance of the grd computng system. For example, Lawson et. al.[7] presented a schedulng for both remote obs that s mgrated from another node and local obs by use of the mult-queue concept n a dstrbuted parallel processng envronment. Although a varety of dstrbuted computng schedulng algorthms have been studed [8], they tend to consder only the problem s complexty, and they may be unsutable to be drectly used n a geographcally dstrbuted grd computng envronment. A backfll algorthm [9] that fts for a heterogeneous dstrbuted parallel computng envronment was ntroduced to acheve a faster response tme n grd computng. However, there exst trade-offs among performance metrcs because the algorthm may yeld dfferent ob dspatchng results dependng on the characterstcs and prortes of obs. More specfcally, prortes of a ob that are determned to ncrease the system utlzaton can ncur ob slowdown, and as a consequence may result n a decrease of system utlzaton. It s well known that NP-completeness results when the utlzaton of the grd computng system and ob slowdown s smultaneously mproved by adustng the prorty of obs based on ob characterstcs [10]. Motvated by ths, we propose a hybrd ob schedulng mechansm that can explot both the meta-schedulng scheme and the local ob schedulng scheme. Ths paper s organzed as follows. In secton 2, related work s presented, and the hybrd ob schedulng mechansm s then presented n secton 3. Secton 4 dscusses the Manuscrpt receved September 5, 2012 Manuscrpt revsed September 20, 2012

2 40 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September 2012 backfll-based mult-queues schedulng scheme. In secton 5, the experments and the results are provded to llustrate the performance of the proposed mechansm, and the research summary and future drectons for possble research are dscussed n Secton Related Work The meta-schedulng schemes can be classfed nto centralzed schedulng and de-centralzed schedulng. In the centralzed schedulng, all grd obs are submtted to the meta-scheduler that dstrbutes them to nodes based on the status of the nodes n whch ths status nformaton s managed by the scheduler. Although ths approach can be satsfactory, t may not be scalable because the metascheduler needs to manage all nformaton and dstrbute the obs. On the other hand, the de-centralzed schedulng does not requre a separate meta-scheduler snce all the obs are submtted drectly to the schedulers of the relevant local node. Dependng on a schedulng polcy, the scheduler of each node then assgns obs to tself or sends them to another avalable node. Although hgh scalablty can be realzed by ths approach, ths scheme s consderably more dffcult to mplement, and synchronzaton s hard to acheve [11]. Most of the research on grd schedulng s conducted based on the decentralzed schedulng scheme where submtted obs n a certan node are assgned to other nodes when the obs cannot be processed n that node [12]-[14]. Wth regard to the ob processng types, there are three types: batch processng, nteractve processng, and parallel processng. Snce the batch processng and the nteractve processng are executed n a sngle node, submtted obs are usually allocated to the least-loaded node. If a node s overloaded, the obs beng executed n that node are sent to another node. On the other hand, the parallel processng requres multple concurrent processors to process obs. Therefore, heavy delays n communcatng among nodes or long ob executon tmes can be ncurred f the obs that are beng executed on multple nodes do not run n a synchronous manner [15]. Varable parttonng scheme s one of the most frequently used schedulng schemes for parallel processng ob schedulng. In ths scheme, assgnments of obs can be descrbed as rectangular shapes n a graph wth the requred resource space on the x-axs and the ob executon tme on the y-axs. As Fg. 1 shows, the occuped area represents the system utlzaton of a parallel system. It s worth notng that a ob wth a low prorty cannot be executed even when resources are avalable snce the varable parttonng follows the frst-come frst-served (FCFS) polcy. As shown n Fg. 1, t s obvous that ob (4) that arrves after ob (3) has been completed cannot be executed snce the resource s not enough due to the two obs (1 and 2) that are beng currently processed. Nor can ob (5) be executed even though there exsts suffcent avalable resources to process t. Ths s because ob (5) has a lower prorty than ob (4) accordng to FCFS polcy. Ths causes fragmentaton that decreases the system performance as a result of beng unable to use dle resources. Therefore, t can be nferred that the response tme for ob (5) can be reduced f the ob can be assgned wthout nfluencng ob (4). Tme Job 5 arrves Job 4 arrves Runnng Job (J1) Runnng Job (J2) Watng Job (5) Watng Job (4) Runnng Job (J3) Fg. 1 Varable parttonng scheme. Shortage System resource The backfll scheme has been proposed to address the fragmentaton problem [9]. Ths scheme changes the prorty of grd obs n the ob queue n such a way that a ob can be moved to the top of the ob queue f t can be executed by currently avalable resources wthout delayng the processng of hgher prorty obs. A number of schemes based on backfll schedulng have been developed to enhance performance by: 1) reducng ob slowdown by assgnng a hgh prorty to a ob wth short executon tme [16], 2) reducng ob slowdown by sortng ob orders accordng to the watng tme snce ther submsson [17], and 3) ncreasng system utlzaton by allowng ob slowdowns ncurred by obs wth short executon tme or low prorty [18]. In spte of the research, the trade-off between system utlzaton and ob slowdowns remans to be addressed [19]. Although gang schedulng [20] and dynamc partton [21] have also been proposed for the purpose of addressng the fragmentaton problem, several lmtatons such as communcaton overhead or clock synchronzaton n a node have restrcted ther extensve use. In the backfll schedulng schemes, the conservatve backfll schemes allow executon of a low prorty ob only

3 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September when t does not cause any slowdowns of all the other obs wth hgher prorty. The EASY backfll schemes execute low prorty obs n a queue when the slowdown of only the frst ob n the queue s not caused. The EASY backfll can ncrease the system utlzaton of the entre system by backfllng more obs than the conservatve backfll. However, t may not be able to guarantee ob completon tmes to users due to possble unbounded delays caused by the ob slowdowns. In ths paper, a hybrd ob schedulng mechansm s presented to ncrease system utlzaton that s not easy to acheve by the conservatve backfll and to prevent the unbounded delay ncurred by the EASY backfll. The mechansm dstrbutes hgh-prorty obs to the ob queue where reserved obs are stored. Also, t allocates obs wth low prorty or remote obs that are sent from other nodes to the backfll queues. Several experments are conducted to evaluate the proposed mechansm wth real data, and the results show that the utlzaton of a grd computng system ncreases whle ob slowdown decreases at the same tme. meta-scheduler that keeps track of processng requests for remote obs, and t then allocates the ob n the ob queue f the meta-schedulng polcy s not volated (7. Allocate Job). After the recevng node completes the ob n accordance wth a local ob schedulng strategy, t notfes the user (8. Reply Job). 3. Structure of Dstrbuted Grd Computng System Fg. 2 shows the archtecture of the grd computng system that enables the hybrd ob schedulng mechansm. In the archtecture, grd workers (.e. computers) consttutng a grd computng node are lnked as a sngle logcal group through the Internet, and they operate as a sngle computer system. Each node conssts of grd workers, a metascheduler, and a local ob scheduler. The hybrd ob schedulng mechansm proposed n ths paper supports the nteractons between a meta-scheduler and a local ob scheduler to mnmze the schedulng load wthn a local node and to facltate cooperaton among nodes. As such, t can mprove the effcency and the avalablty of a grd computng system by allowng both the meta-schedulng and the local ob schedulng to be smultaneously utlzed. When a user submts a ob to a node (0. Intate Job), the scheduler of the node ether allocates the ob to the ob queue for the submtted ob to be processed by a local ob schedulng polcy, or sends t to one of the backfll queues where the ob wats to be assgned to the ob queue (1. Request Job). If t s found to be mpossble to execute the ob at the node due to excessve load, the scheduler requests other nodes for processng t (2. Request Remote Job). If resources are avalable n other node(s) by broadcastng and replyng among nodes (3. Dscover & 4. Volunteer), the ob s sent to the other node(s) (5. Negotate & 6. Request Job). At ths tme, the local scheduler n the recevng node communcates wth the Fg. 2 Archtecture of dstrbuted grd computng system. To put t another way, the proposed archtecture for the hybrd ob schedulng can ncorporate centralzed herarchcal schedulng strateges that can be exercsed by meta-schedulers and de-centralzed schedulng polces mplemented by local ob schedulers smultaneously. Thus, communcaton bottlenecks, SPOF (Sngle Pont of Falure) problems, or scalablty problems can be addressed wth the grd computng archtecture. Fg. 3 Mult-queue scheduler of computng nodes. Fg. 3 shows the structure of the mult-queue scheduler of

4 42 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September 2012 local nodes n a dstrbuted grd computng system. Each local node conssts of workers partcpatng n the grd computng and a scheduler that supports both the metaschedulng and the local ob schedulng by means of the mult-queue backfll mechansm. The scheduler agan conssts of two types of queues: The frst s the ob queue that stores local obs that are executed by FCFS polcy, and the second are the two backfll queues where obs are watng for backfllng. It should be noted that the scheduler consders both the states of all the workers and the ob characterstcs when t employs the schedulng strategy. In other words, a ob that s submtted to an arbtrary node s assgned to one of the three queues wthn the node by the local ob scheduler. To whch queue a ob s assgned s determned based on the number of requred processors and the expected executon tme of the ob. Jobs that requre a large number of processors are assgned to the ob queue where other obs are reserved dependng on a ob prorty. On the other hand, local obs requrng a small number of processors and obs that are receved from other nodes are sent to the backfll queues. If the node s overloaded, t s sent to other node wth the least-loaded backfll queues by the meta-scheduler. The ob s then stored n one of the backfll queues of the recevng node. 4. Mult-queue Schedulng Mechansm 4.1 Mult-queue Management Strategy The mult-queue management strategy (MMS) s related to the process that allocates submtted obs to ether the ob queue or the backfll queues based on ther characterstcs. In backfll schedulng, t s mportant to determne how many obs are to be backflled n an effcent way. Jobs that requre a small number of processors may have hgh possblty of beng backflled snce t s easer for them to acqure avalable processors wthn a node. Ths approach can ncrease the lkelhood of mprovng the utlzaton of the processors. In addton, whle obs n the ob queue are executed n FCFS fashon, obs n the backfll queue can be backflled whenever a backfll s possble even when a ob n the ob queue s beng executed. Thus, MMS strategy can be appled when 1) a ob s submtted to a local node, 2) a ob on the local node s completed, or 3) a remote ob s dspatched from other nodes. It s worth remarkng that prortes of remote obs are set to be lower than those of the local obs. Ths s appled by allocatng them to the backfll queues rather than to the ob queue. It can contrbute to guaranteeng the completon tme of local obs that are submtted drectly to a local node. More specfcally, n order to prevent a local ob submtted to a certan node from beng slowed down by remote obs, flags are gven to obs of specfed ob types such as a local ob or a remote ob Job Classfcaton and Prorty Assgnment Strateges for Local Jobs For the meta-schedulers and the local ob schedulers to apply MMS, they reference basc nformaton that ncludes ob number (), ob arrval tme ( TA ), and node number ( k ) on the grd. In addton, obs submtted to a certan local node contan the parameters wth regard to the number of the requred processors ( P ) for executon and the expected executon tme ( TES ). Thus, a ob can be represented by the followng expresson (1). JOB( TA, TES, P, k ) (1) Based on nformaton about obs, MMS classfes and prortzes obs through the followng steps. In step 1, obs are classfed nto three groups dependng on the total number of requred processors n the node to whch they are drectly submtted or remotely dspatched. In step 2, each group s further classfed nto two subgroups accordng to the expected executon tme. For those obs that are classfed n one of sx groups that are formed through these steps, MMS s appled n such a way that 1) obs requrng a large number of processors are allocated to the ob queue for fast completon, and 2) those requrng a small number of processors are assgned to the backfll queues for the hgher system utlzaton through more frequent backflls (refer to Fgure 4). In step 3, prortes of obs are assgned. n.mult.dstrbute () f (.req <= n.processor * 1/3) n.mult.enqueue(, backfll2); else f (.req <= n.processor * 2/3) n.mult.enqueue(, backfll1); else n.mult.enqueue(, obqueue); n.mult.mgrate(n, ); n.mult.enqueue(, backfll1); Fg. 4 Mult-queue management strategy pseudo-code. Step 1 Job classfcaton by the number of processors

5 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September In ths step, obs are classfed nto three groups based on the number of requred processors as expressed by condtons (2) (4), where TP s the number of processors n node. After beng classfed, obs are allocated to the queues n a node. Large obs are stored n the ob queue, and medum obs and small obs are stored n the frst backfll queue and the second backfll queue, respectvely. L(Large Job) : TP 2 P 3 TP M(Medum Job) : TP 1 2 P 3 < TP 3 S(Small Job) : 1 P 1 < TP 3 Step 2 Job classfcaton by expected executon tme In ths step, obs n each group formed by the step 1 are further dvded nto two types based on the expected executon tme. If the expected executon tme of a ob s longer than the average executon tme of obs that have been completed n the recevng node up to a current pont of tme, the ob s classfed as L-type (Long Job). Otherwse, t s characterzed as S-type (Short Job). Ths classfcaton s expressed n condtons (5)-(6): (2) (3) (4) L(Long Job) : TES TMEAN (5) S(Short Job) : TES < TMEAN ( 6) where TES s the expected executon tme of the ob, and TMEAN s the average executon tme of obs that have been executed on the recevng node. It should be noted that the average executon tme n a node vares as the node contnues to process multple obs. Therefore, the classfcaton crteron changes n a dynamcal manner. SS (Small Short) Low SL (Small Long) ML MS (Medum Long) (Medum Short) prorty Fg. 5 Prorty order of local obs. LL (Large Long) LS (Large Short) Step 3- Prorty assgnment by ob class A ob type that s obtaned through the above mentoned ob classfcaton steps s used to determne the prorty of ob processng n the ob queue and the sequence of the backfll n the backfll queues. More specfcally, S-type obs have hgher prortes n processng or backfllng than L-type obs. Fg. 5 shows the prortes of local obs that Hgh are determned by the two steps descrbed above. Jobs n the rghtmost cells have the hghest prorty, and those n the leftmost have lowest prorty Job Classfcaton and Prorty Assgnment Strateges for Remote Jobs As aforementoned, remote obs sent from other nodes are allocated to the backfll queues of a recevng node. The second part n Fg. 4 shows the step that processes the remote obs. Any remote obs are assgned to the backfll queue (e.g. Backfll1) of a recevng node by the dstrbuton strategy for remote obs. Ths s n order to gve the local obs hgher prorty over remote obs so that they can be processed earler than the remote obs. The classfcaton of the remote obs s done n a smlar way as the local obs. In the frst step, remote obs are classfed accordng to the number of the requred processors ( P ) for executon, and they are labeled as ether Remote Large (RL) obs or Remote Small (RS) obs dependng on condtons (7) (8). RL (Remote Large Job) : TP 1 P 2 TP RS (Remote Small Job) : 1 P 1 < TP 2 In the second step, they are further classfed based on the executon tme as n condtons (9)-(10). When a ob s sent from another node, the recevng node references nformaton concernng ob executon tme and compares t wth the average executon tme up to a current pont of tme to determne the type of the remote ob. (7) (8) RL (Remote Long Job) : TES TMEAN (9) RS (Remote Short Job) : TES < TMEAN (10) The Fg. 6 shows the prorty orders of remote obs that are determned by the two steps mentoned above. RSS (Remote Small Short) Low RSL (Remote Small Long) RLL (Remote Large Long) prorty Fg. 6 Prorty order of remote obs. 4.2 Hybrd Job Schedulng Strategy RLS (Remote Large Short) Hgh The meta-schedulng strategy ams at mprovng the system utlzaton of the entre grd system and the local ob

6 44 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September 2012 schedulng strategy helps acheve an effcent ob dstrbuton wthn a node. Jobs n the backfll queues can be delayed unboundedly when requred processors for executon of a ob are not obtaned, and the obs are not reserved n the backfll queue. In partcular, the probablty of unbounded delay of remote obs can be larger than that of local obs. The drawbacks of both the conservatve and the EASY backfll mechansms that were descrbed n secton 2 can be allevated by the mechansm proposed n ths paper, whch prevents unbounded delays and mproves system utlzaton. n.global.dscover () V = volunteer lst wth dle resources; for each v V v.global.negotate(n, ); n.global.negotate (n, ) runnable = n.local.create(); f (runnable) n.global.commt(n, ); else n.global.wat(n, ); n.global.commt (n, ) n.mult.mgrate(n, ); n.global.wat (n, ) f (n.mult..resv_counterg++ > Threshold){ fx ob s prorty nto the urgent level; n.global.reserve(); } (a) meta scheduler n.local.create () span = some runs durng an admtted tme; f (n.aval >=.req) n.local.execute (, span); else { q = n.mult.dstrbute(); n.local.wat(); } n.local.remote () n.global.dscover(); n.local.execute (, span).met =.met span; f (.met == ) n.mult.depart(); else { span = some runs durng an admtted tme; n.local.execute(, span); } n.local.wat ( ) f (n.mult..rcounterl++> Threshold){ fx ob s prorty nto the urgent level; n.local.reserve(); } n.local.reserve () span = thereserved executon slotsfor.req; n.local.backfll(, span); n.local.backfll (, span).met =.met span; f (.met == ) n.mult.depart(); else n.local.execute(); n.local.volunteer () newly = current dle resources; n.aval = n.aval + newly; (b) local scheduler Fg. 7 The meta- and the local ob schedulng procedure. Meta-schedulng Strategy: When a ob s submtted to a node, the meta-scheduler determnes whether the ob s to be processed n the recevng node or to be sent to another node based on the status of recevng node and grd system. In ths stage, the meta-scheduler negotates wth meta-schedulers of other nodes for better ob allocaton. Fg. 7 (a) shows the procedures of the meta- schedulng. When a ob () s submtted to a node (n), the meta-scheduler of node (n) searches for volunteer nodes for processng a ob () (n. global.dscover ()). At ths tme, the meta-scheduler of a volunteer node (e.g., node v) communcates wth the local ob scheduler to check resource avalablty for processng the ob (). If a resource s avalable, node (v) responds wth a commt message (n.global.commt (v, )). Then, ob () s mgrated from node (n) to node (v) (n.global.mgrate (v, )). Local Job schedulng Strategy: The local ob schedulers can manage nformaton about a node status regardng the number of avalable processors, the system utlzaton (memory usage, network usage, etc.), system slowdowns, and expected response tme. Based on ths nformaton, the scheduler makes a decson wth regard to ob executon and dspatchng to other node (refer to Fg. 7 (b)). To prevent unbounded delay of obs n the ob queue, a certan number of obs are reserved to be executed so that they are not slowed down due to other obs n the backfll queue. By dong ths, the system utlzaton can also be ncreased more than the EASY backfll whereas only the frst ob s reserved n the EASY backfll mechansm. The number of the reserved obs ( J ) n the ob queue s r determned by (11): J r = α J, 0.1 α 0.5 (11) n where J s the number of queued obs n the ob queue, n and α s the threshold value used to adust the number of the reserved obs. To prevent unbounded delays of obs n the backfll queue, a reservaton technque s also appled to obs n the backfll queue. More specfcally, when the number of obs that are not backflled ( J ) s found to be larger than a NB threshold value ( β ) at a certan pont of tme, those obs are dspatched to the ob queue, or sent to other node f they can be executable n that node. It can be expressed as (12). J β (12) NB When the slowdowns of obs n the backfll queues occur even f the obs are reserved by condton 12, those obs need to be sent to another node. In ths case, those obs become remote obs n the other node that receves them. To determne whch node the obs are to be sent to, the mechansm proposed n ths paper checks whether the obs can be executed n the recevng node n consderaton of executon and transfer tme of the obs. By dong ths, the nfluence of remote obs n a recevng node can be mnmzed. The number of reserved obs plays a key role n the proposed mechansm. It s set to greater than n the EASY backfll and less than n the conservatve backfll.

7 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September Performance Evaluaton In order to evaluate the performance of the hybrd ob schedulng mechansm proposed n ths paper, several experments were conducted. In ths type of experment, nput data play a crtcal role n evaluatng the performance of dfferent mechansms because of ther nfluence on the ob load. Accordng to the followng two steps, the nput data are generated. In the frst step, parameters of ndependent varables (e.g., mean arrval tme, mean estmated executon tme, mean number of processors, mean wdth, etc.) are generated from the set of workload logs of Fetelson Archve [22]. Table 1 shows the parameters that are generated from CTC trace (CTC: the Cornell Theory Center 512-node IBM SP2, 79,296 obs) and Fetelson workload. The nodes of the CTC computer are not all dentcal and they dffer n type and memory. Table 1: Workload data Month Total A Type B Type July Aug Sep Oct Nov Dec Jan Feb Mar Apr May Average compared by means of the average response tme metrcs. The average slowdown tme of the ob can be obtaned usng (13): [16] N resp( ) = wdth( ) * max( S, exec( )) AverageSlowdownTme = 1 N (13) where N s the total number of obs, S s the processng tme of the ob wth the shortest executon tme, resp() s the tme elapsed between ob submsson and completon, exec() s the executon tme of the ob, and wdth() s the number of the processors requred for the ob to be executed. To compare the effectveness of the mult-queue (AverageSlowdownTme m ) and the sngle queue (AverageSlowdownTme 1 ) strateges, the slowdown rato s used. SlowdownRato (14) AverageSlowdownTme1 AverageSlowdownTmem = mn( AverageSlowdownTme, 1 AverageSlowdownTme ) As can be observed n (14), a postve value of SlowdownRato means that the average slowdown of the sngle queue strategy s greater than that of the mult-queue strategy. On the other hand, negatve value of SlowdownRato ndcates that the mult-queue strategy causes greater slowdowns due to the management overheads. m A Type: Jobs requrng at most 256 nodes, B Type: Jobs requrng at most 128 nodes In the second step, wth the parameters that are generated n the frst step, other parameters for the experments are then obtaned based on the probablty functons whch are descrbed as follows. The average ob arrval, the ob request tme, and the average number of requred processors are assumed to follow an exponental dstrbuton wth the rate of 0.167/mn. ( 7200 obs/month), an exponental dstrbuton wth the average of 100 mn, and a unform dstrbuton from 1 to 64, respectvely [9]. Fnally, a grd system wth 8 nodes of 64 processors each s used. The backfllng ratos that s the rato of the number of backflled obs to the number of queued obs wth regard to the system utlzaton are measured, and the mpact of the hybrd mult-queue and the remote ob processng s assessed n terms of the slowdown rato of the obs (SlowdownRato, refer to equaton 14). In addton, the slowdowns caused by dfferent schedulng strateges are Backfllng Rato CONS EASY HYBRID System Utlzaton Fg. 8 Backfllng ratos dependng on the system utlzaton. Fg. 8 shows the changes of the backfllng ratos of three dfferent strateges dependng on the system utlzaton that s calculated by dvdng the ob arrval tme by the average

8 46 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September CONS 0.6 EASY HYBRID 0.4 Slowdown Rato dspatchng obs to remote nodes plays an mportant role n mprovng the response tme of ob executons f a dedcated network s not used to connect nodes of a grd. Note that the ncreased system utlzaton does not always mean reduced ob response tme. 24 Average Response Tme (n 10 3 sec) processng tme. Therefore, more frequent ob requests can ncrease the system utlzaton. As can be observed n Fg. 8, the EASY backfll method has more backflls than the conservatve backfll method. Furthermore, backflls occur the most frequently when the hybrd mult-queue schedulng mechansm s appled, whch means that hybrd mult-queue schedulng mechansm mproves the system utlzaton of the entre grd system. 0.2 CONS EASY 22 HYBRID SS SL MS ML LS 10 LL System Utlzaton -0.4 Fg. 10 Average response tme accordng to system utlzaton Job Classfcaton Fg. 9 The changes of the slowdown rato. 20 nse Tme (n 10 3 sec) Average Respo α Fg. 9 shows the changes of the slowdown rato of obs accordng to the number of requred processors and the expected executon tme for whch both are used ob classfcaton. It can be observed that the hybrd multqueue s more effectve to decrease the ob slowdown rato than the sngle queue regardless of schedulng strategy. In addton, the slowdowns occur for the obs (LS and LL) that requre a large number of processors due to the backflls of the obs wth low prorty. However, the hybrd mult-queue schedulng mechansm causes fewer slowdowns than both the conservatve and the EASY backfll methods even f a sngle queue s used. From ths observaton, t can be nferred that the hybrd ob schedulng method can have advantages that nclude the unbounded delay preventon and the ncreased backfll rato whch are the drawbacks of the conservatve backfll method and of the EASY backfll methods, respectvely. Fg. 10 shows the changes of the average response tme wth regard to the system utlzaton. In case of the EASY backfll method, hgher system utlzaton ncurs a steep ncrease n the average response tme due to ncreased number of backflls. In the case of the hybrd method proposed n ths paper, the average response tme s greater than the conservatve method n spte of smaller ob slowdown. Ths s because the hybrd method dspatches obs to other nodes more frequently than the conservatve method. Therefore, t can be nferred that the overhead of β Fg. 11 Average response tme wth parameters ( α, β ) used n the reservaton polcy. Fg. 11 shows the graph of the average response tmes accordng to the value of α (the threshold value that s used to adust the rato of the reserved obs to queued obs n ob queue) and β (the threshold value for those obs that are backflled for executon) that are ntroduced to mplement the reservaton polcy of the conservatve and the EASY backflll methods and to prevent unbounded delay of the EASY backfll method, respectvely. As α ncreases, the average response tme also ncreases. Ths s because more obs are moved to the ob queue and backflled, resultng n longer watng tme of obs that are

9 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September not backflled. Also, the decrease of the threshold value of β that restrans the backfll of obs that can otherwse be backflled causes the ncrease n the average response tme. However, the mpact of the value of β on the average response tme s observed to decrease for somewhat large values of β (e.g., β 64). Ths s more obvous n case of small values of α. In partcular, because there can be more avalable processors f each node has more processors, the probablty of the slowdowns caused by the backflls decreases. 6. Concluson Although a sgnfcant amount of research has been conducted on the developments of schedulng mechansms to reduce ob slowdown, mprove the response tme and the system utlzaton of a grd system, most of the research has been pursued n a separate manner to solve problems of the meta-schedulng methods that dstrbute obs to each node and the local ob schedulng methods wthn a node. In order to address ths stuaton, we have proposed the hybrd ob schedulng mechansm that consders both the meta-schedulng and the local ob schedulng approaches at the same tme. Jobs that are submtted to grd computng nodes are classfed based on the number of requred processors and the expected ob executon tme. Ths classfcaton schema s then used to determne whether obs are processed n a local node or dspatched to other node(s) accordng to the ob load. As the backfllng rato and ob slowdowns are mproved wth the hybrd mult-queue schedulng mechansm, the utlzaton of a grd computng system becomes more effcent and watng tmes are reduced. Future work wll nclude the nvestgaton of the mpact of the network communcaton overhead on overall performance of a grd system and wll be checkng for the mplementaton ssues wth varous nput data sets. Acknowledgments Ths work (Grants No. C ) was supported by Busness for Cooperatve R&D between Industry, Academy, and Research Insttute funded Korea Small and Medum Busness Admnstraton n References [1] I. Foster, et. al., Grd Servces for Dstrbuted System Integraton," Computer, Vol. 35, No. 6, pp , [2] Ken Park, et. al., Credble Worker Selecton Mechansm for Grd Computng, IEICE Transactons on Informaton and Systems, Vol. E89-D, No.2, pp , Feb [3] K. Krauter, et. al., A Taxonomy and Survey of Grd Resource Management Systems for Dstrbuted Computng, Software Practce and Experence Journal, Vol. 32, No. 2, pp , Feb Summary: Grd computng s concerned wth the sharng and coordnated use of dverse resources n dstrbuted "vrtual organzatons." The dynamc and multnsttutonal nature of these envronments ntroduces challengng securty ssues that demand new techn... [4] I. Foster, et. al., Globus: A Metacomputng Infrastructure Toolkt, The Internatonal Journal of Supercomputer Applcatons and Performance Computng, Vol. 11, No. 2, pp , Oct [5] A. Grmshaw, et. al., Legon: Lessons Learned Buldng a Grd Operatng System, Proceedngs of the IEEE, Vol. 93, Issue 3, pp , Mar [6] H. Shan, et. al., Job Superscheduler Archtecture and Performance n Computatonal Grd Envronments, In SC 2003 Conference, [7] B. Lawson, et. al., "Multple-queue Backfllng Schedulng wth Prortes and Reservatons for Parallel Systems," The 8th Internatonal Workshop, JSSPP 2002 Ednburgh, Scotland, UK, pp , July [8] T. Braun, et. al., A Comparson of Eleven Statc Heurstcs for Mappng a Class of Independent Tasks onto Heterogeneous Dstrbuted Computng Systems, Journal of Parallel and Dstrbuted Computng, Vol. 61, pp , [9] A. Mualem, et. al., Utlzaton, Predctablty, Workloads and User Run tme Estmates n Schedulng the IBM SP2 wth Backfllng, IEEE Trans. Parallel and Dstrbuted System, Vol. 12, No. 6, pp , June [10] O. Ibarra, et. al., Heurstc Algorthm for Schedulng Independent Tasks on Nondentcal Processors, Journal of ACM, Vol. 24, No. 2, pp , Apr [11] V. Hamscher, et. al., "Evaluaton of Job-Schedulng Strateges for Grd Computng," The 1st IEEE/ACM Internatonal Workshop on Grd Computng (Grd 2000) at the 7th Internatonal Conference on Hgh Performance Computng (HPC-2000), LNCS 1971, pp , [12] V. Subraman, et. al., Dstrbuted Job Schedulng on Computatonal Grds Usng Multple Smultaneous Requests, The 11th IEEE Internatonal Symposum on Hgh Performance Dstrbuted Computng (HPDC ), pp , July [13] Q. Wang, et. al., De-centralzed Job Schedulng on Computatonal Grds Usng Dstrbuted Backfllng, Grd and Cooperatve Computng - GCC 2004: Thrd Internatonal Conference, LNCS 3251, pp , Oct [14] K. L, Job Schedulng for Grd Computng on Metacomputers, The 19th Internatonal Parallel and Dstrbuted Processng Symposum (IPDPS 2005), Abstracts Proceedngs, Apr [15] D. Fetelson, et. al., Theory and Practce n Parallel Job Schedulng, Job Schedulng Strateges for Parallel Processng, IPPS'97 Workshop, Geneva, Swtzerland, LNCS 1291, pp. 1-34, Apr. 5, [16] D. Zotkn, et al., Job-Length Estmaton and Performance n Backfllng Schedulers, The 8th IEEE Internatonal

10 48 IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.12 No.9, September 2012 Symposum on Hgh Performance Dstrbuted Computng (HPDC'99), Aug [17] S. Srnvasan, et al., Characterzaton of Backfllng Strateges for Parallel Jobs Schedulng, 31st Internatonal Conference on Parallel Processng Workshops (ICPP 2002 Workshops), pp , Aug [18] W. A. Ward Jr., et al., Schedulng Jobs on Parallel Systems Usng a Relaxed Backfll Strategy,, 8th Internatonal Workshop on Job Schedulng Strateges for Parallel Processng(JSSPP 2002), pp , July [19] B. G. Lawson, et al., Self-Adaptng Backfllng Schedulng for Parallel Systems, 31st Internatonal Conference on Parallel Processng (ICPP 2002), pp , Aug [20] D. Fetelson, et al., Improved Utlzaton and Responsveness wth Gang Schedulng, Job Schedulng Strateges for Parallel Processng, IPPS'97 Workshop, Geneva, Swtzerland, LNCS 1291, pp , Apr [21] R. McCann, et. al., A Dynamc Processor Allocaton Polcy for Multprogrammed Sharedmemory Multprocessors, ACM Trans. on Computer System, Vol. 11, No. 2, pp , May [22] D. Fetelson, Logs of Real Parallel Workloads from Producton Systems, Avalable: Ken Park receved the B.S. degree n ndustral engneerng from Hanyang Unversty, Seoul, Korea, n 1989, the M.S. degree n ndustral engneerng from Pohang Unversty of Scence and Technology, Pohang, Korea, n 1991, and the Ph.D. degree from the Department of Computer Engneerng, Graduate School, Aou Unversty, Suwon, Korea, n From 1991 to 1997, he was wth the Software Research and Development Center, Samsung Electroncs Company Ltd., Suwon, as a Senor Researcher. From 2001 to 2002, he was wth the Network Equpment Test Center, Electroncs and Telecommuncatons Research Insttute, Daeeon, Korea, as a Senor Researcher. From 2002 to 2004, he was wth the Department of Computer Engneerng, Anyang Unversty, Anyang, Korea, as a Professor. Snce 2004, he has been an Assocate Professor wth the Dvson of Industral and Informaton Systems Engneerng, Aou Unversty. From 2010 to 2011, he was wth Rutgers, The State Unversty of New Jersey, Pscataway, as a Vstng Professor. Hs research nterests nclude n-vehcle network, fault-tolerant computng, and cloud computng. Changhoon Kang was born n Daeeon, Korea. He receved the B.S. and M.S. degrees n computer scence from Chungnam Natonal Unversty, Korea, n 1986 and 1988, respectvely, and Ph. D. degree n Department of Computer Engneerng, Graduate School of Aou Unversty n Korea n He s currently an Assocate Professor n Department of Vsual Broadcastng Meda, Kangdong College. From 1990 to 1993, he worked n the Computer Center of Chungnam Natonal Unversty, Korea, as an Assstant Teacher. Hs research nterests nclude cluster computng, and grd computng. Sung sook Km receved the B.S degree n Educatonal technology from Hanyang Unversty, Seoul, Korea, n 1991, the M.S. degree n E-Learnng from Aou Unversty, Suwon, Korea, n 2009, respectvely, where she s currently workng toward the Ph.D. degree n the Department of Computer Scence & Engneerng n Anyang Unversty. Her research nterests nclude real-tme data mnng from moble log data and task schedulng for MapReduce software framework n cloud computng.

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism for Nested Loops with Non-uniform and Flow Dependences Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr

More information

Distributed Resource Scheduling in Grid Computing Using Fuzzy Approach

Distributed Resource Scheduling in Grid Computing Using Fuzzy Approach Dstrbuted Resource Schedulng n Grd Computng Usng Fuzzy Approach Shahram Amn, Mohammad Ahmad Computer Engneerng Department Islamc Azad Unversty branch Mahallat, Iran Islamc Azad Unversty branch khomen,

More information

Scheduling Remote Access to Scientific Instruments in Cyberinfrastructure for Education and Research

Scheduling Remote Access to Scientific Instruments in Cyberinfrastructure for Education and Research Schedulng Remote Access to Scentfc Instruments n Cybernfrastructure for Educaton and Research Je Yn 1, Junwe Cao 2,3,*, Yuexuan Wang 4, Lanchen Lu 1,3 and Cheng Wu 1,3 1 Natonal CIMS Engneerng and Research

More information

A Frame Packing Mechanism Using PDO Communication Service within CANopen

A Frame Packing Mechanism Using PDO Communication Service within CANopen 28 A Frame Packng Mechansm Usng PDO Communcaton Servce wthn CANopen Mnkoo Kang and Kejn Park Dvson of Industral & Informaton Systems Engneerng, Ajou Unversty, Suwon, Gyeongg-do, South Korea Summary The

More information

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory Background EECS. Operatng System Fundamentals No. Vrtual Memory Prof. Hu Jang Department of Electrcal Engneerng and Computer Scence, York Unversty Memory-management methods normally requres the entre process

More information

Virtual Machine Migration based on Trust Measurement of Computer Node

Virtual Machine Migration based on Trust Measurement of Computer Node Appled Mechancs and Materals Onlne: 2014-04-04 ISSN: 1662-7482, Vols. 536-537, pp 678-682 do:10.4028/www.scentfc.net/amm.536-537.678 2014 Trans Tech Publcatons, Swtzerland Vrtual Machne Mgraton based on

More information

Cluster Analysis of Electrical Behavior

Cluster Analysis of Electrical Behavior Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School

More information

Video Proxy System for a Large-scale VOD System (DINA)

Video Proxy System for a Large-scale VOD System (DINA) Vdeo Proxy System for a Large-scale VOD System (DINA) KWUN-CHUNG CHAN #, KWOK-WAI CHEUNG *# #Department of Informaton Engneerng *Centre of Innovaton and Technology The Chnese Unversty of Hong Kong SHATIN,

More information

AADL : about scheduling analysis

AADL : about scheduling analysis AADL : about schedulng analyss Schedulng analyss, what s t? Embedded real-tme crtcal systems have temporal constrants to meet (e.g. deadlne). Many systems are bult wth operatng systems provdng multtaskng

More information

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

More information

Evaluation of an Enhanced Scheme for High-level Nested Network Mobility

Evaluation of an Enhanced Scheme for High-level Nested Network Mobility IJCSNS Internatonal Journal of Computer Scence and Network Securty, VOL.15 No.10, October 2015 1 Evaluaton of an Enhanced Scheme for Hgh-level Nested Network Moblty Mohammed Babker Al Mohammed, Asha Hassan.

More information

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments Comparson of Heurstcs for Schedulng Independent Tasks on Heterogeneous Dstrbuted Envronments Hesam Izakan¹, Ath Abraham², Senor Member, IEEE, Václav Snášel³ ¹ Islamc Azad Unversty, Ramsar Branch, Ramsar,

More information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,

More information

Related-Mode Attacks on CTR Encryption Mode

Related-Mode Attacks on CTR Encryption Mode Internatonal Journal of Network Securty, Vol.4, No.3, PP.282 287, May 2007 282 Related-Mode Attacks on CTR Encrypton Mode Dayn Wang, Dongda Ln, and Wenlng Wu (Correspondng author: Dayn Wang) Key Laboratory

More information

A Genetic Algorithm Based Dynamic Load Balancing Scheme for Heterogeneous Distributed Systems

A Genetic Algorithm Based Dynamic Load Balancing Scheme for Heterogeneous Distributed Systems Proceedngs of the Internatonal Conference on Parallel and Dstrbuted Processng Technques and Applcatons, PDPTA 2008, Las Vegas, Nevada, USA, July 14-17, 2008, 2 Volumes. CSREA Press 2008, ISBN 1-60132-084-1

More information

Efficient Distributed File System (EDFS)

Efficient Distributed File System (EDFS) Effcent Dstrbuted Fle System (EDFS) (Sem-Centralzed) Debessay(Debsh) Fesehaye, Rahul Malk & Klara Naherstedt Unversty of Illnos-Urbana Champagn Contents Problem Statement, Related Work, EDFS Desgn Rate

More information

Load Balancing for Hex-Cell Interconnection Network

Load Balancing for Hex-Cell Interconnection Network Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,

More information

Simulation Based Analysis of FAST TCP using OMNET++

Simulation Based Analysis of FAST TCP using OMNET++ Smulaton Based Analyss of FAST TCP usng OMNET++ Umar ul Hassan 04030038@lums.edu.pk Md Term Report CS678 Topcs n Internet Research Sprng, 2006 Introducton Internet traffc s doublng roughly every 3 months

More information

Real-time Fault-tolerant Scheduling Algorithm for Distributed Computing Systems

Real-time Fault-tolerant Scheduling Algorithm for Distributed Computing Systems Real-tme Fault-tolerant Schedulng Algorthm for Dstrbuted Computng Systems Yun Lng, Y Ouyang College of Computer Scence and Informaton Engneerng Zheang Gongshang Unversty Postal code: 310018 P.R.CHINA {ylng,

More information

Evaluation of Parallel Processing Systems through Queuing Model

Evaluation of Parallel Processing Systems through Queuing Model ISSN 2278-309 Vkas Shnde, Internatonal Journal of Advanced Volume Trends 4, n Computer No.2, March Scence - and Aprl Engneerng, 205 4(2), March - Aprl 205, 36-43 Internatonal Journal of Advanced Trends

More information

Concurrent Apriori Data Mining Algorithms

Concurrent Apriori Data Mining Algorithms Concurrent Apror Data Mnng Algorthms Vassl Halatchev Department of Electrcal Engneerng and Computer Scence York Unversty, Toronto October 8, 2015 Outlne Why t s mportant Introducton to Assocaton Rule Mnng

More information

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task

Term Weighting Classification System Using the Chi-square Statistic for the Classification Subtask at NTCIR-6 Patent Retrieval Task Proceedngs of NTCIR-6 Workshop Meetng, May 15-18, 2007, Tokyo, Japan Term Weghtng Classfcaton System Usng the Ch-square Statstc for the Classfcaton Subtask at NTCIR-6 Patent Retreval Task Kotaro Hashmoto

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty

More information

Improved Resource Allocation Algorithms for Practical Image Encoding in a Ubiquitous Computing Environment

Improved Resource Allocation Algorithms for Practical Image Encoding in a Ubiquitous Computing Environment JOURNAL OF COMPUTERS, VOL. 4, NO. 9, SEPTEMBER 2009 873 Improved Resource Allocaton Algorthms for Practcal Image Encodng n a Ubqutous Computng Envronment Manxong Dong, Long Zheng, Kaoru Ota, Song Guo School

More information

Maintaining temporal validity of real-time data on non-continuously executing resources

Maintaining temporal validity of real-time data on non-continuously executing resources Mantanng temporal valdty of real-tme data on non-contnuously executng resources Tan Ba, Hong Lu and Juan Yang Hunan Insttute of Scence and Technology, College of Computer Scence, 44, Yueyang, Chna Wuhan

More information

Wishing you all a Total Quality New Year!

Wishing you all a Total Quality New Year! Total Qualty Management and Sx Sgma Post Graduate Program 214-15 Sesson 4 Vnay Kumar Kalakband Assstant Professor Operatons & Systems Area 1 Wshng you all a Total Qualty New Year! Hope you acheve Sx sgma

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Proceedngs of the Wnter Smulaton Conference M E Kuhl, N M Steger, F B Armstrong, and J A Jones, eds A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Mark W Brantley Chun-Hung

More information

Queueing Network-based Optimisation Techniques for Workload Allocation in Clusters of Computers *

Queueing Network-based Optimisation Techniques for Workload Allocation in Clusters of Computers * Queueng Network-based Optmsaton Technques for Workload Allocaton n Clusters of Computers * Lgang He, Stephen A. Jarvs, Davd Bacgalupo, Danel P. Spooner, Xnuo Chen and Graham R. Nudd Department of Computer

More information

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster

More information

A Model Based on Multi-agent for Dynamic Bandwidth Allocation in Networks Guang LU, Jian-Wen QI

A Model Based on Multi-agent for Dynamic Bandwidth Allocation in Networks Guang LU, Jian-Wen QI 216 Jont Internatonal Conference on Artfcal Intellgence and Computer Engneerng (AICE 216) and Internatonal Conference on etwork and Communcaton Securty (CS 216) ISB: 978-1-6595-362-5 A Model Based on Mult-agent

More information

Load-Balanced Anycast Routing

Load-Balanced Anycast Routing Load-Balanced Anycast Routng Chng-Yu Ln, Jung-Hua Lo, and Sy-Yen Kuo Department of Electrcal Engneerng atonal Tawan Unversty, Tape, Tawan sykuo@cc.ee.ntu.edu.tw Abstract For fault-tolerance and load-balance

More information

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 A mathematcal programmng approach to the analyss, desgn and

More information

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z.

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z. TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS Muradalyev AZ Azerbajan Scentfc-Research and Desgn-Prospectng Insttute of Energetc AZ1012, Ave HZardab-94 E-mal:aydn_murad@yahoocom Importance of

More information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,

More information

Performance Study of Parallel Programming on Cloud Computing Environments Using MapReduce

Performance Study of Parallel Programming on Cloud Computing Environments Using MapReduce Performance Study of Parallel Programmng on Cloud Computng Envronments Usng MapReduce Wen-Chung Shh, Shan-Shyong Tseng Department of Informaton Scence and Applcatons Asa Unversty Tachung, 41354, Tawan

More information

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers

Content Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth

More information

Burst Round Robin as a Proportional-Share Scheduling Algorithm

Burst Round Robin as a Proportional-Share Scheduling Algorithm Burst Round Robn as a Proportonal-Share Schedulng Algorthm Tarek Helmy * Abdelkader Dekdouk ** * College of Computer Scence & Engneerng, Kng Fahd Unversty of Petroleum and Mnerals, Dhahran 31261, Saud

More information

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices Steps for Computng the Dssmlarty, Entropy, Herfndahl-Hrschman and Accessblty (Gravty wth Competton) Indces I. Dssmlarty Index Measurement: The followng formula can be used to measure the evenness between

More information

Dynamic Bandwidth Provisioning with Fairness and Revenue Considerations for Broadband Wireless Communication

Dynamic Bandwidth Provisioning with Fairness and Revenue Considerations for Broadband Wireless Communication Ths full text paper was peer revewed at the drecton of IEEE Communcatons Socety subject matter experts for publcaton n the ICC 008 proceedngs. Dynamc Bandwdth Provsonng wth Farness and Revenue Consderatons

More information

An Efficient Garbage Collection for Flash Memory-Based Virtual Memory Systems

An Efficient Garbage Collection for Flash Memory-Based Virtual Memory Systems S. J and D. Shn: An Effcent Garbage Collecton for Flash Memory-Based Vrtual Memory Systems 2355 An Effcent Garbage Collecton for Flash Memory-Based Vrtual Memory Systems Seunggu J and Dongkun Shn, Member,

More information

Constructing Minimum Connected Dominating Set: Algorithmic approach

Constructing Minimum Connected Dominating Set: Algorithmic approach Constructng Mnmum Connected Domnatng Set: Algorthmc approach G.N. Puroht and Usha Sharma Centre for Mathematcal Scences, Banasthal Unversty, Rajasthan 304022 usha.sharma94@yahoo.com Abstract: Connected

More information

Scheduling Independent Tasks in Heterogeneous Environments under Communication Constraints

Scheduling Independent Tasks in Heterogeneous Environments under Communication Constraints Schedulng Independent Tasks n Heterogeneous Envronments under Communcaton Constrants Petros Lampsas 1 Thanass Loukopoulos 2 Fedon Dmopoulos 1 Marouso Athanasou 1 1 Department of Informatcs and Computer

More information

Parallel matrix-vector multiplication

Parallel matrix-vector multiplication Appendx A Parallel matrx-vector multplcaton The reduced transton matrx of the three-dmensonal cage model for gel electrophoress, descrbed n secton 3.2, becomes excessvely large for polymer lengths more

More information

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION 24 CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION The present chapter proposes an IPSO approach for multprocessor task schedulng problem wth two classfcatons, namely, statc ndependent tasks and

More information

A Semi-Distributed Load Balancing Architecture and Algorithm for Heterogeneous Wireless Networks

A Semi-Distributed Load Balancing Architecture and Algorithm for Heterogeneous Wireless Networks A Sem-Dstrbuted oad Balancng Archtecture and Algorthm for Heterogeneous reless Networks Md. Golam Rabul Ala Choong Seon Hong * Kyung Hee Unversty, Korea rob@networkng.khu.ac.kr, cshong@khu.ac.kr Abstract

More information

A New Transaction Processing Model Based on Optimistic Concurrency Control

A New Transaction Processing Model Based on Optimistic Concurrency Control A New Transacton Processng Model Based on Optmstc Concurrency Control Wang Pedong,Duan Xpng,Jr. Abstract-- In ths paper, to support moblty and dsconnecton of moble clents effectvely n moble computng envronment,

More information

Security Enhanced Dynamic ID based Remote User Authentication Scheme for Multi-Server Environments

Security Enhanced Dynamic ID based Remote User Authentication Scheme for Multi-Server Environments Internatonal Journal of u- and e- ervce, cence and Technology Vol8, o 7 0), pp7-6 http://dxdoorg/07/unesst087 ecurty Enhanced Dynamc ID based Remote ser Authentcaton cheme for ult-erver Envronments Jun-ub

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana

More information

Chapter 1. Introduction

Chapter 1. Introduction Chapter 1 Introducton 1.1 Parallel Processng There s a contnual demand for greater computatonal speed from a computer system than s currently possble (.e. sequental systems). Areas need great computatonal

More information

A Binarization Algorithm specialized on Document Images and Photos

A Binarization Algorithm specialized on Document Images and Photos A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a

More information

DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT

DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT Bran J. Wolf, Joseph L. Hammond, and Harlan B. Russell Dept. of Electrcal and Computer Engneerng, Clemson Unversty,

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning the Kernel Parameters in Kernel Minimum Distance Classifier Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department

More information

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour 6.854 Advanced Algorthms Petar Maymounkov Problem Set 11 (November 23, 2005) Wth: Benjamn Rossman, Oren Wemann, and Pouya Kheradpour Problem 1. We reduce vertex cover to MAX-SAT wth weghts, such that the

More information

Positive Semi-definite Programming Localization in Wireless Sensor Networks

Positive Semi-definite Programming Localization in Wireless Sensor Networks Postve Sem-defnte Programmng Localzaton n Wreless Sensor etworks Shengdong Xe 1,, Jn Wang, Aqun Hu 1, Yunl Gu, Jang Xu, 1 School of Informaton Scence and Engneerng, Southeast Unversty, 10096, anjng Computer

More information

ARTICLE IN PRESS. Signal Processing: Image Communication

ARTICLE IN PRESS. Signal Processing: Image Communication Sgnal Processng: Image Communcaton 23 (2008) 754 768 Contents lsts avalable at ScenceDrect Sgnal Processng: Image Communcaton journal homepage: www.elsever.com/locate/mage Dstrbuted meda rate allocaton

More information

ELEC 377 Operating Systems. Week 6 Class 3

ELEC 377 Operating Systems. Week 6 Class 3 ELEC 377 Operatng Systems Week 6 Class 3 Last Class Memory Management Memory Pagng Pagng Structure ELEC 377 Operatng Systems Today Pagng Szes Vrtual Memory Concept Demand Pagng ELEC 377 Operatng Systems

More information

HIERARCHICAL SCHEDULING WITH ADAPTIVE WEIGHTS FOR W-ATM *

HIERARCHICAL SCHEDULING WITH ADAPTIVE WEIGHTS FOR W-ATM * Copyrght Notce c 1999 IEEE. Personal use of ths materal s permtted. However, permsson to reprnt/republsh ths materal for advertsng or promotonal purposes or for creatng new collectve wors for resale or

More information

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points; Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features

More information

Learning-Based Top-N Selection Query Evaluation over Relational Databases

Learning-Based Top-N Selection Query Evaluation over Relational Databases Learnng-Based Top-N Selecton Query Evaluaton over Relatonal Databases Lang Zhu *, Wey Meng ** * School of Mathematcs and Computer Scence, Hebe Unversty, Baodng, Hebe 071002, Chna, zhu@mal.hbu.edu.cn **

More information

Two-Stage Data Distribution for Distributed Surveillance Video Processing with Hybrid Storage Architecture

Two-Stage Data Distribution for Distributed Surveillance Video Processing with Hybrid Storage Architecture Two-Stage Data Dstrbuton for Dstrbuted Survellance Vdeo Processng wth Hybrd Storage Archtecture Yangyang Gao, Hatao Zhang, Bngchang Tang, Yanpe Zhu, Huadong Ma Bejng Key Lab of Intellgent Telecomm. Software

More information

Efficient Content Distribution in Wireless P2P Networks

Efficient Content Distribution in Wireless P2P Networks Effcent Content Dstrbuton n Wreless P2P Networs Qong Sun, Vctor O. K. L, and Ka-Cheong Leung Department of Electrcal and Electronc Engneerng The Unversty of Hong Kong Pofulam Road, Hong Kong, Chna {oansun,

More information

Classifier Selection Based on Data Complexity Measures *

Classifier Selection Based on Data Complexity Measures * Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

CMPS 10 Introduction to Computer Science Lecture Notes

CMPS 10 Introduction to Computer Science Lecture Notes CPS 0 Introducton to Computer Scence Lecture Notes Chapter : Algorthm Desgn How should we present algorthms? Natural languages lke Englsh, Spansh, or French whch are rch n nterpretaton and meanng are not

More information

Solution Brief: Creating a Secure Base in a Virtual World

Solution Brief: Creating a Secure Base in a Virtual World Soluton Bref: Creatng a Secure Base n a Vrtual World Soluton Bref: Creatng a Secure Base n a Vrtual World Abstract The adopton rate of Vrtual Machnes has exploded at most organzatons, drven by the mproved

More information

WIRELESS communication technology has gained widespread

WIRELESS communication technology has gained widespread 616 IEEE TRANSACTIONS ON MOBILE COMPUTING, VOL. 4, NO. 6, NOVEMBER/DECEMBER 2005 Dstrbuted Far Schedulng n a Wreless LAN Ntn Vadya, Senor Member, IEEE, Anurag Dugar, Seema Gupta, and Paramvr Bahl, Senor

More information

A GENETIC ALGORITHM FOR PROCESS SCHEDULING IN DISTRIBUTED OPERATING SYSTEMS CONSIDERING LOAD BALANCING

A GENETIC ALGORITHM FOR PROCESS SCHEDULING IN DISTRIBUTED OPERATING SYSTEMS CONSIDERING LOAD BALANCING A GENETIC ALGORITHM FOR PROCESS SCHEDULING IN DISTRIBUTED OPERATING SYSTEMS CONSIDERING LOAD BALANCING M. Nkravan and M. H. Kashan Department of Electrcal Computer Islamc Azad Unversty, Shahrar Shahreqods

More information

Modelling a Queuing System for a Virtual Agricultural Call Center

Modelling a Queuing System for a Virtual Agricultural Call Center 25-28 July 2005, Vla Real, Portugal Modellng a Queung System for a Vrtual Agrcultural Call Center İnc Şentarlı, a, Arf Orçun Sakarya b a, Çankaya Unversty, Department of Management,06550, Balgat, Ankara,

More information

Real-time Scheduling

Real-time Scheduling Real-tme Schedulng COE718: Embedded System Desgn http://www.ee.ryerson.ca/~courses/coe718/ Dr. Gul N. Khan http://www.ee.ryerson.ca/~gnkhan Electrcal and Computer Engneerng Ryerson Unversty Overvew RTX

More information

Priority-Based Scheduling Algorithm for Downlink Traffics in IEEE Networks

Priority-Based Scheduling Algorithm for Downlink Traffics in IEEE Networks Prorty-Based Schedulng Algorthm for Downlnk Traffcs n IEEE 80.6 Networks Ja-Mng Lang, Jen-Jee Chen, You-Chun Wang, Yu-Chee Tseng, and Bao-Shuh P. Ln Department of Computer Scence Natonal Chao-Tung Unversty,

More information

RAP. Speed/RAP/CODA. Real-time Systems. Modeling the sensor networks. Real-time Systems. Modeling the sensor networks. Real-time systems:

RAP. Speed/RAP/CODA. Real-time Systems. Modeling the sensor networks. Real-time Systems. Modeling the sensor networks. Real-time systems: Speed/RAP/CODA Presented by Octav Chpara Real-tme Systems Many wreless sensor network applcatons requre real-tme support Survellance and trackng Border patrol Fre fghtng Real-tme systems: Hard real-tme:

More information

Fault tolerant workflow scheduling based on replication and resubmission of tasks in Cloud Computing

Fault tolerant workflow scheduling based on replication and resubmission of tasks in Cloud Computing Fault tolerant workflow schedulng based on replcaton and resubmsson of tasks n Cloud Computng Jayadvya S K* Department of Computer Scence & Engneerng Natonal Insttute of Technology Truchrappall, Taml Nadu,

More information

The Codesign Challenge

The Codesign Challenge ECE 4530 Codesgn Challenge Fall 2007 Hardware/Software Codesgn The Codesgn Challenge Objectves In the codesgn challenge, your task s to accelerate a gven software reference mplementaton as fast as possble.

More information

The Research of Support Vector Machine in Agricultural Data Classification

The Research of Support Vector Machine in Agricultural Data Classification The Research of Support Vector Machne n Agrcultural Data Classfcaton Le Sh, Qguo Duan, Xnmng Ma, Me Weng College of Informaton and Management Scence, HeNan Agrcultural Unversty, Zhengzhou 45000 Chna Zhengzhou

More information

Hybrid Non-Blind Color Image Watermarking

Hybrid Non-Blind Color Image Watermarking Hybrd Non-Blnd Color Image Watermarkng Ms C.N.Sujatha 1, Dr. P. Satyanarayana 2 1 Assocate Professor, Dept. of ECE, SNIST, Yamnampet, Ghatkesar Hyderabad-501301, Telangana 2 Professor, Dept. of ECE, AITS,

More information

Towards Autonomous Service Composition in A Grid Environment

Towards Autonomous Service Composition in A Grid Environment Towards Autonomous Servce Composton n A Grd Envronment Wllam K. Cheung +, Jmng Lu +, Kevn H. Tsang +, Raymond K. Wong ++ Department of Computer Scence + Hong Kong Baptst Unversty Hong Kong {wllam,jmng,hhtsang}@comp.hkbu.edu.hk

More information

The Data Warehouse in a Distributed Utility Environment

The Data Warehouse in a Distributed Utility Environment The Data Warehouse n a Dstrbuted Utlty Envronment Charles A. Mllgan Dstngushed Engneer, Sun Mcrosystems Charles.mllgan@sun.com Abstract Utlty provsonng, Grd resource management, nstant copy kosks, and

More information

Dynamic Bandwidth Allocation Schemes in Hybrid TDM/WDM Passive Optical Networks

Dynamic Bandwidth Allocation Schemes in Hybrid TDM/WDM Passive Optical Networks Dynamc Bandwdth Allocaton Schemes n Hybrd TDM/WDM Passve Optcal Networks Ahmad R. Dhan, Chad M. Ass, and Abdallah Sham Concorda Insttue for Informaton Systems Engneerng Concorda Unversty, Montreal, Quebec,

More information

Private Information Retrieval (PIR)

Private Information Retrieval (PIR) 2 Levente Buttyán Problem formulaton Alce wants to obtan nformaton from a database, but she does not want the database to learn whch nformaton she wanted e.g., Alce s an nvestor queryng a stock-market

More information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,

More information

A QoS-aware Scheduling Scheme for Software-Defined Storage Oriented iscsi Target

A QoS-aware Scheduling Scheme for Software-Defined Storage Oriented iscsi Target A QoS-aware Schedulng Scheme for Software-Defned Storage Orented SCSI Target Xanghu Meng 1,2, Xuewen Zeng 1, Xao Chen 1, Xaozhou Ye 1,* 1 Natonal Network New Meda Engneerng Research Center, Insttute of

More information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for

More information

A fair buffer allocation scheme

A fair buffer allocation scheme A far buffer allocaton scheme Juha Henanen and Kalev Klkk Telecom Fnland P.O. Box 228, SF-330 Tampere, Fnland E-mal: juha.henanen@tele.f Abstract An approprate servce for data traffc n ATM networks requres

More information

Internet Traffic Managers

Internet Traffic Managers Internet Traffc Managers Ibrahm Matta matta@cs.bu.edu www.cs.bu.edu/faculty/matta Computer Scence Department Boston Unversty Boston, MA 225 Jont work wth members of the WING group: Azer Bestavros, John

More information

Wireless Sensor Network Localization Research

Wireless Sensor Network Localization Research Sensors & Transducers 014 by IFSA Publshng, S L http://wwwsensorsportalcom Wreless Sensor Network Localzaton Research Lang Xn School of Informaton Scence and Engneerng, Hunan Internatonal Economcs Unversty,

More information

Adaptive Power-Aware Prefetch in Wireless Networks

Adaptive Power-Aware Prefetch in Wireless Networks Adaptve Power-Aware Prefetch n Wreless Networks Langzhong Yn, Student Member, IEEE, and Guohong Cao, Member, IEEE Abstract Most of the prefetch technques used n the current cache management schemes do

More information

Reliability and Performance Models for Grid Computing

Reliability and Performance Models for Grid Computing Relablty and Performance Models for Grd Computng Yuan-Shun Da,2, Jack Dongarra,3,4 Department of Electrcal Engneerng and Computer Scence, Unversty of Tennessee, Knoxvlle 2 Department of Industral and Informaton

More information

Fast Computation of Shortest Path for Visiting Segments in the Plane

Fast Computation of Shortest Path for Visiting Segments in the Plane Send Orders for Reprnts to reprnts@benthamscence.ae 4 The Open Cybernetcs & Systemcs Journal, 04, 8, 4-9 Open Access Fast Computaton of Shortest Path for Vstng Segments n the Plane Ljuan Wang,, Bo Jang

More information

An efficient iterative source routing algorithm

An efficient iterative source routing algorithm An effcent teratve source routng algorthm Gang Cheng Ye Tan Nrwan Ansar Advanced Networng Lab Department of Electrcal Computer Engneerng New Jersey Insttute of Technology Newar NJ 7 {gc yt Ansar}@ntedu

More information

A Fair Access Mechanism Based on TXOP in IEEE e Wireless Networks

A Fair Access Mechanism Based on TXOP in IEEE e Wireless Networks 11 Internatonal Journal of Communcaton Networks and Informaton Securty (IJCNIS) Vol. 8, No. 1, Aprl 16 A Far Access Mechansm Based on TXOP n IEEE 8.11e Wreless Networks Marjan Yazdan 1, Maryam Kamal, Neda

More information

Lecture 5: Multilayer Perceptrons

Lecture 5: Multilayer Perceptrons Lecture 5: Multlayer Perceptrons Roger Grosse 1 Introducton So far, we ve only talked about lnear models: lnear regresson and lnear bnary classfers. We noted that there are functons that can t be represented

More information

Use of Genetic Algorithms in Efficient Scheduling for Multi Service Classes

Use of Genetic Algorithms in Efficient Scheduling for Multi Service Classes Use of Genetc Algorthms n Effcent Schedulng for Mult Servce Classes Shyamale Thlakawardana and Rahm Tafazoll Centre for Communcatons Systems Research (CCSR), Unversty of Surrey, Guldford, GU27XH, UK Abstract

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

An Entropy-Based Approach to Integrated Information Needs Assessment Dstrbuton Statement A: Approved for publc release; dstrbuton s unlmted. An Entropy-Based Approach to ntegrated nformaton Needs Assessment June 8, 2004 Wllam J. Farrell Lockheed Martn Advanced Technology

More information

Analysis of Collaborative Distributed Admission Control in x Networks

Analysis of Collaborative Distributed Admission Control in x Networks 1 Analyss of Collaboratve Dstrbuted Admsson Control n 82.11x Networks Thnh Nguyen, Member, IEEE, Ken Nguyen, Member, IEEE, Lnha He, Member, IEEE, Abstract Wth the recent surge of wreless home networks,

More information

A Heuristic for Mining Association Rules In Polynomial Time*

A Heuristic for Mining Association Rules In Polynomial Time* Complete reference nformaton: Ylmaz, E., E. Trantaphyllou, J. Chen, and T.W. Lao, (3), A Heurstc for Mnng Assocaton Rules In Polynomal Tme, Computer and Mathematcal Modellng, No. 37, pp. 9-33. A Heurstc

More information

an assocated logc allows the proof of safety and lveness propertes. The Unty model nvolves on the one hand a programmng language and, on the other han

an assocated logc allows the proof of safety and lveness propertes. The Unty model nvolves on the one hand a programmng language and, on the other han UNITY as a Tool for Desgn and Valdaton of a Data Replcaton System Phlppe Quennec Gerard Padou CENA IRIT-ENSEEIHT y Nnth Internatonal Conference on Systems Engneerng Unversty of Nevada, Las Vegas { 14-16

More information

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms Course Introducton Course Topcs Exams, abs, Proects A quc loo at a few algorthms 1 Advanced Data Structures and Algorthms Descrpton: We are gong to dscuss algorthm complexty analyss, algorthm desgn technques

More information

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 An Iteratve Soluton Approach to Process Plant Layout usng Mxed

More information

Fibre-Optic AWG-based Real-Time Networks

Fibre-Optic AWG-based Real-Time Networks Fbre-Optc AWG-based Real-Tme Networks Krstna Kunert, Annette Böhm, Magnus Jonsson, School of Informaton Scence, Computer and Electrcal Engneerng, Halmstad Unversty {Magnus.Jonsson, Krstna.Kunert}@de.hh.se

More information

BIN XIA et al: AN IMPROVED K-MEANS ALGORITHM BASED ON CLOUD PLATFORM FOR DATA MINING

BIN XIA et al: AN IMPROVED K-MEANS ALGORITHM BASED ON CLOUD PLATFORM FOR DATA MINING An Improved K-means Algorthm based on Cloud Platform for Data Mnng Bn Xa *, Yan Lu 2. School of nformaton and management scence, Henan Agrcultural Unversty, Zhengzhou, Henan 450002, P.R. Chna 2. College

More information