An Integrated Dynamic Resource Scheduling Framework in On-Demand Clouds *

Size: px
Start display at page:

Download "An Integrated Dynamic Resource Scheduling Framework in On-Demand Clouds *"

Transcription

1 JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 30, (2014) An Integrated Dynamc Resource Schedulng Framework n On-Demand Clouds * College of Computer Scence and Technology Zhejang Unversty Hangzhou, P.R. Chna E-mal: {lexu; zjuzhwang; chenwz}@zju.edu.cn The bggest advantage of employng vrtualzaton n cloud s the ablty to provson resource flexbly whch makes pay-as-use model possble. However, snce the workload of vrtual machne constantly changes, t s stll a challenge that effcently schedule resource by mgratng vrtual machnes among lots of hosts, especally wth mult-objectve to meet, lke power and QoS constrants. In ths paper, we present a dynamc resource schedulng soluton, named Smart-DRS, whch can well determnes when to schedule, whch to schedule, and where to schedule. It s an ntegrated framework that ams to deal well wth hotspot mtgaton, load balancng and server consoldaton whch are classc problems to solve n on-demand clouds. The expermental results show that our framework strkes a balance between effcency, overhead and nstantanety. Keywords: cloud, dynamc resource schedule, hotspot mtgaton, load balancng, server consoldaton 1. INTRODUCTION The ablty of vrtualzaton technology whch s an mportant enabler for cloud computng brngs mmense benefts n terms of relablty, effcency and scalablty. Vrtualzaton, coupled wth VM (Vrtual Machne) mgraton capablty, enables the cloud data centers to flexbly provson resource whch makes the computng-as-a-servce vson of on-demand clouds possble. Ths pay-as-use model enables user pay only for the resources they really used n whch compute resources are made avalable as a utlty servce an lluson of avalablty of as much resources (e.g., CPU, memory, and network bandwdth) as demanded by the user. It s the most mportant feature of IaaS (Infrastructure as a Servce) and also the bggest push of cloud computng. However, mplementng such an on-demand cloud data center requres an excellent DRS (Dynamc Resource Scheduler), a power and QoS aware scheduler. As we know, the energy consumpton of datacenter s an ncreasngly sharp problem. So we need to power off several PMs (Physcal Machnes) by consoldatng all VMs together to certan selected PMs when the workload s low. Meanwhle, to keep hgh QoS performance of cloud servce, we need to well manage resource to automatcally balance load and mtgate hgh load machne n tme. For achevng aforementoned goals, n ths paper, we present a three-step dynamc resource schedulng strategy, named Smart-DRS, whch s an ntegrated schedulng solu- Receved December 14, 2012; revsed November 8, 2013; accepted November 16, Communcated by Ruay-Shung Chang. * Ths work has been presented n 2012 Internatonal Workshop on Power and QoS Aware Computng (P- QoSCom 12) n conjuncton wth IEEE Cluster 12, Sept. 2012, Bejng, sponsored by IEEE, IEEE Computer Socety, IEEE TCSC, Intel, VMware, NVda, etc. 1537

2 1538 ton that proved to have a good performance by expermentng n the prototype system we developed. We start the schedulng for three steps n general: (1) Frstly, we employ a predcton technque based on Sngle Exponental Smoothng algorthm to determne when to schedule; (2) Then, we use holstc approach (defne a score model) to determne whch VM should be scheduled; (3) At last, n the most mportant process of schedulng, we use a novel methodology based on vector projecton arthmetc to decde where to allocate the potental VMs. The rest of ths paper s organzed as follows: Secton 2 descrbes related work. Secton 3 detals the algorthms usng n Smart-DRS. Fnally, n Secton 4, we mplement a prototype system to evaluate the performance of Smart-DRS. A summary and plan of our future work are descrbed n Secton RELATED WORKS Dynamc resource schedulng s a prmary problem n vrtual envronment management. It needs dynamcally balances computng capacty across a collecton of hardware resources aggregated nto logcal resource pools. When a VM experences an ncreased load, t should automatcally allocated addtonal resources by redstrbutng VM among the PMs n the resource pool. Ths problem has attracted a lot of attenton from the research communty n the last few years. In [1-3], the authors dscusses respectvely about schedulng scheme makng and executng. They treat VM deployment as a Bn Packng problem and use the Bn Packng dynamc programmng algorthm to solve. Ths method traverses dfferent number of backpacks to fnd the least number of backpacks that can accommodate all tems. However, n fact, VM schedulng problem s more complcated than the Bn Packng problem, because VM has more than two-dmenson factors to care about when place t to the target PM. That means deployng ths method on cloud datacenter has several lmtatons. Mnmze the traffc when schedule the VMs has also been studed. [4] makes use of the network footprnts to generate a good mappng for the data center. The goal s to mnmze the communcaton cost between all the VMs. [5] also ams to mnmze the communcaton traffc n a data center. They consder the topology of a data center and set communcaton cost as the number of swtches between two VMs. Reducng power consumpton s also often consdered. The goal s to reduce the number of actve PMs. There are already some approxmate algorthms proposed to solve ths. The most wdely used s frst-ft [6-8], whch can generate a near-optmal placement. But we thnk that frst-ft has many lmtatons. In a producton data center, VMs are enterng and leavng contnuously and redeployment are requred perodcally. Frst-ft gnores the mgraton cost and may generate unnecessary overhead. We need to take the mgraton cost nto account. Besdes above all, Wood et al. [9] develop an ntegrated VM schedulng system named Sandpper whch s a XEN based automated provsonng system for montorng and detectng hotspots. Thus, t detects when a VM s under-provsoned and ether allots more resources locally or mgrates the VM to a new PM whch s capable of supportng the VM. However, Sandpper s easy to choose a wrong target PM, because Sandpper takes VM mgraton decson based on a metrc, whch t refers to as volume. That means

3 AN INTEGRATED DYNAMIC RESOURCE SCHEDULING FRAMEWORK 1539 t converts the three dmensonal resource nformaton of PMs nto a sngle dmenson metrc (whch s volume) and then uses ths sngle dmenson metrc for worst ft n a three dmensonal scenaro. In ths process, the nformaton about the shape of the resource utlzaton s lost. Compared to these works mentoned above, our work has the followng man contrbutons: We present an ntegrated schedulng soluton whch contans the major processng procedures of resource management n cloud, such as when to schedule, whch to schedule and where to schedule. We apply SES algorthm to predct workload and score model to choose the mgrated VM. In addton, more mportantly, we employ a novel algorthm based on vector projecton arthmetc to select the potental target PM whch could retan the shape of the mult-dmensonal object representng VM or PM. We develop a prototype system runnng n a small scale cluster wth 32PMs and 3200VMs to evaluate the performance of Smart-DRS. 3. DETAILS OF SMART-DRS In ths secton, we dscuss n detal about our three-step dynamc resource schedulng framework, Smart-DRS. As we know, there are three classc problems [8] to solve n the feld of cloud resource management whch towards smultaneously mnmzng power consumpton and maxmzng QoS performance: Hotspot mtgaton: The goal of hotspot mtgaton s to avod a condton n whch VM runs n a hot PM whch has nadequate resource to meet ts servce level agreements (SLA) requrements. VM mgraton can mtgate ths hotspot, so as to mprove the QoS performance. Load balancng: The goal of load balancng s to avod a stuaton where there s a large dscrepancy n resource utlzaton levels of the PMs. VMs can be mgrated from the hgh workload host to the low workload one, so as to acheve ths balance and mprove the QoS performance. Server consoldaton: The goal of consoldaton s to avod server sprawl many PMs host low-resource-usage VMs. VMs on lghtly loaded host can be packed onto fewer machnes to meet resource requrements, so as to reduce power consumpton. However, for each of the three goals mentoned above, we need to address three mportant questons: when to schedule, whch to schedule and where to schedule. In the followng, we wll dscuss our methods respectvely. 3.1 When to Schedule? As we know, most of tradtonal DRS strateges are based on the threshold whch means the physcal machne surpasses the upper boundary of threshold, then to trgger a

4 1540 mgraton. Therefore, a VM mgraton could be trggered by nstantaneous peak load, whch leads to the unnecessary waste of mgraton expendture. However, Smart-DRS employs predcton technque based on SES (Sngle Exponental Smoothng) algorthm, whch observes n future load values, f at least k value s above the threshold, we thnk the next load value s above the threshold. SES algorthm s a wdely used forecastng method and an effcent technque that can be appled to tme seres data, ether to produce smooth data for presentaton. SES algorthm s very sutable for our model, because n our system, the tme seres data are a sequence of hstorcal montorng data of resource utlzaton. And the observed phenomenon may be an essentally random process, or t may be an orderly but nosy process. Whereas n the SES the past observatons are weghted equally, exponental smoothng assgns exponentally decreasng weghts over tme. SES removes random perturbatons of tme seres data, then the form of SES s gven by the formula (1), whle the sequence of observatons begns at tme t = 0. y y ( y y ) t1 t t t (1) Where yt 1 s the predcton value at tme t+1, yt s the predcton value at tme t, whle y t s the real value at tme t, s the smoothng factor, and 0 < < 1. Formula (1) can change further to formula (2). y y (1 ) y t1 t t Through the observaton, the above equaton s knd of a recurson equaton, whch can be expanded to formula (3). (2) t1 t y (1 ) y (1 ) y t1 0 t 1 (3) As we can see from formula (3), exponental smoothng forecast value s a weghted sum of all the prevous real observatonal values. That means SES makes use of all the hstorcal data, so t has more stablty and regularty. In addton, we can t avod that there would be some error nformaton n the collected hstorcal montorng data whch wll affect the accuracy of the result. Therefore, we need process the data n advance to delete the nformaton that s not approprate. (4) 1 n 2 1 n n ( y y), y 1 n y 1 y would be arranged n order statstcs. If y satsfes y > k, then y s a bad value, whch should be elmnated, and k s Grubbs crteron coeffcent. Besdes defnng the error nformaton model, we also need to determne the value of and y1 to calculate the predctve value. The Value of : The value of determne the degree of smoothng and how responsve the model s to fluctuaton n the tme seres data. The value of s arbtrary and s de-

5 AN INTEGRATED DYNAMIC RESOURCE SCHEDULING FRAMEWORK 1541 termned both by the nature of the data and the feelng by the forecaster as to what consttutes a good response rate. A smoothng constant close to zero leads to a stable model whle a constant close to one s hghly reactve. To our knowledge, t s good to set a small value of n order to ncrease the weght of hstorcal data when the tme seres data doesn t have fluctuatons. On the contrary, t s good to set a bg value of to ncrease the weght of recent predcton value when the tme seres data has obvous fluctuatons. The Value of y 1 : In fact, the smaller value of, the more senstve our predcton value wll be on the selecton of ths ntal smoother value y 1. In our method, we defne y 1 to be ntalzed to y 1 when the number of seres data s more than an experental value 15 [10]. Whle the number s less than 15, we defne y 1 to be the average value of the seres data. As shown n formula (5). k y / k, k 15 1 y 1 y, k 15 1 (5) 3.2 Whch to Schedule? Selectng one or more VMs for mgraton s a crucal decson of the resource management. The mgraton process not only makes the VM unavalable for a certan amount of tme but also consumes resources lke CPU, memory and network bandwdth on source and destnaton PMs. Performance of other VMs that are hosted on source and destnaton PMs are also affected due to ncreased resource requrements durng mgraton. Many exstng VM selecton approaches are straghtforward that selectng the canddate VM for mgraton because of resource constraned. The VM whose resource requrements can t be locally fulflled s selected. However, Smart-DRS employs a more holstc approach that CPU utlzaton, memory usage and net bandwdth usage are all consdered before selectng the canddate VM. Reducng the memory data quantty that needs to be mgrated s very mportant. Because that lve mgraton mechansm s a way that teratvely copy VM memory pages to the target host as well as trackng those pages that have been modfed n the copy process. It should ntercept all memory access of the mgratng VM whch wll serously affect the applcaton performance runnng n ths VM. By reducng the amount of data copes n the network can mnmze the total mgraton tme and also mtgate the mpact on applcaton performance. Meanwhle, the VM whose CPU utlzaton rate or net bandwdth consumpton s hgh means t has a great help to mtgate hotspot after mgraton. So Smart-DRS defnes a model to balance these factors: Score(cpu, mem, net) = F(workload(cpu, net), cost(mem)). (6) Where Score(cpu, mem, net) takes the three most mportant factors nto consderaton and t s easy to extend to more dmensons. The score of VM s a functon about workload and cost. Meanwhle, the workload s a functon about CPU and network, and the cost s a functon about memory.

6 1542 workload cpu net 1 cost mem (7) Where cpu represents the CPU utlzaton rate of VM, net s ts usage of network bandwdth, mem s ts usage of memory. It s clear that we want to choose the VM wth both the hghest workload value and the cost value. For solvng ths problem, we defne: workload cost Score( cpu, mem, net ). workload cost (8) Score s the harmonc mean of workload and cost, and t would be arranged n descendng order, Smart-DRS chooses the one wth hghest score value as the mgrated VM. From formula (8) we can see that ths VM has both the hghest workload and cost whch means t has the most CPU utlzaton rate and network bandwdth but the least memory usage. In other words, the VM of the hghest score value utlze the most physcal resource but easy to mgrate. 3.3 Where to Schedule? After prevous steps, we should choose out the target PM that has enough resources so that t can support the ncomng mgratng VM. To the best of our knowledge, many exstng methods (ncludng heurstc methods) treat ths problem to be smlar to mult-dmensonal Bn Packng problem. However, whle ths method may seem fne on the surface, certan drawbacks and anomales can be uncovered when we analyze deeper Mult-dmensonal bn packng algorthm Actually the problem of placng resource-aware VMs among PMs looks smlar to the mult-dmensonal bn packng problem. Exstng lteratures [7, 9-11] all tred to solve VM placement problem wth bn packng algorthm and ts varant. However, they are not exactly the same. Exstng methods consder VM as a three dmensonal object and PM as a three dmensonal box. They try to pack as many objects n the boxes as possble, as well as the number of boxes requred are mnmzed. However, whle packng objects nto a gven box, two objects can be placed sde by sde or one on top of the other. But f we consder VMs as the objects, we can t place VMs lke that. Ths s because once a resource s utlzed by a VM, t can t be reused by any other VM. Fg. 1 shows the man dfference. (a) Intal statement (b) Invald placement (c) Invald placement (d) Reasonable placement Fg. 1. The dfference between VM placement and bn packng problem.

7 AN INTEGRATED DYNAMIC RESOURCE SCHEDULING FRAMEWORK 1543 For smplcty, we show ths scenaro n 2D model. As we can see form Fg. 1, when VM2 should be placed n ths PM, the only allowed poston s the one shown n Fg. 1 (d). Whle n bn packng algorthm, the postons as shown n Fgs. 1 (b) and (c) are also allowed. Due to ths shortage of bn packng, we employ a novel algorthm to solve the VM placement problem based on Vector Projecton Vector projecton algorthm Vector Projecton (VP) model, presented n paper [12], s a novel methodology that can make the process of choosng PMs easer. Based on t, we proposed vector projecton algorthm. In ths algorthm, frstly, we choose three major resources avalable wth the PM, namely CPU, MEM and IO. These resources form the three dmensons of an abstract object. We normalze the resources along each axs. Thus, the total avalable resource can be represented as a unt cube whch s called Normalzed Resource Cube (NRC). Secondly, we should express the resource related nformaton of potental VMs and potental target PMs as a vector wthn the NRC, as shown n Fg. 2. The total capacty of PM s expressed as a vector from the orgn of the cube (0, 0, 0) to pont (1, 1, 1). Ths vector s dentfed as Total Capacty Vector (TCV). Resource Utlzaton Vector (RUV) represents the current utlzaton of resources of a PM The vector dfference between TCV and RUV represents the Remanng Capacty Vector (RCV), whch essentally captures how much capacty s left n the PM. The resource requrement of a VM s represented by Resource Requrement Vector (RRV) whch s the vector addton of normalzed resource requrement vectors of each resource type. In order to measure the degree of mbalance of resource utlzaton of a PM, we defne a Resource Imbalance Vector (RIV) of PM. It s the vector dfference between RUV s projecton on TCV and RUV. RIV of a VM s defned n a smlar way: t s the vector dfference between RRV s projecton on TCV and the RRV. Fg. 2. Depcton of vectors n NRC. Table 1. Vectors name of Aronym. Acronym Name TCV Total Capacty of PM Vector RUV Resource Utlzaton of PM Vector RCV Remanng Capacty of PM Vector RRV Resource Requrement of a VM Vector RIV Resource Imbalance of PM/VM Vector Then, we project NRC on a plane perpendcular to the prncpal dagonal of the cube as shown n Fg. 3. It s easy to see that ths would result n a regular hexagon on the sad projecton plane. After that, we take the projecton of the resource vectors onto the projecton plane to make the resource nformaton of 3D space reduce to 2D plane. We refer to ths as Planar Resource Hexagon (PRH). Corners (0, 0, 0) and (1, 1, 1) of the NRC map to the center of the PRH. Other sx corners maps to the sx vertces of the hexagon as marked n the fgure. Then we take the projecton of tp of the resource vectors TCV, RUV (of PM) and RRV (of VM) onto the projecton plane. The projecton

8 1544 ponts of these three vectors would be nsde the regular hexagon. For smplcty, the projecton pont of RUV of PM (RRV of VM) wll be referred to as poston of PM (VM) n the PRH. Note that the projecton of tp of TCV s the center of the PRH. By carefully analyss, we can know the regular hexagon on the projecton plane s dvded nto sx regular trangles. Shown n Fg. 4, we named these trangles lke that: CI defnes the regon where the projecton of tp of RUV whose CPU > MEM > IO. Lkewse, MC represents the regon where the projecton of tp of RUV whose MEM > CPU > IO. Other trangles can be dentfed by smlar nequaltes. Further analyss shows that the VMs and PMs whose resource vectors projected n the same trangle or axs have an affnty. Now, we group all of the potental VMs and potental target PMs based on the rules that the RRV of a VM and RCV of a PM are n the same trangle or the same axs and n the opposte drecton. That means they have the characterstcs of complementary resources. Our ultmate goal s to fnd a best complementary target PM for a potental VM n a group, whose RIV s of same magntude as the VM s RIV and s n the opposte drecton. The groups are shown n Table 2. After comparson, we found Fg. 3. Resource vector projecton. Fg. 4. The planar resource hexagon. Table 2. Vector projecton locaton groupngs. Group Name Group Feature 0 CI Imbalanced n CPU 1 IC Imbalanced n IO 2 IM Imbalanced n IO 3 MI Imbalanced n MEM 4 MC Imbalanced n MEM 5 CM Imbalanced n CPU 6 CPU axs Especally mba n CPU 7 IC/CI Imbalanced n CPU & IO 8 IO axs Especally mba n IO 9 MI/IM Imbalanced n MEM & IO 10 MEM axs Especally mba n MEM 11 MC/CM Imbalanced n mem&cpu 12 Orgn P Resource balancng

9 AN INTEGRATED DYNAMIC RESOURCE SCHEDULING FRAMEWORK 1545 that the trangle s mbalanced n the resource whch the closest resource axs represents. If a trangle has no projecton of tp of vectors, we wll make t wth adjacent trangles nto a new group, because adjacent trangles are wth smlar characterstcs of re-sources. Loop ths process untl every group has resource vectors VP algorthm elaboraton Ths segment detals VP algorthm whch s a VM placement algorthm. In secton 3.2, we have ntroduced how to choose a VM to mgrate. And VM placement problem s how to choose a target PM for ths VM. The target PM for ths VM can be chosen based on one of the three goals: hotspot mtgaton, load balancng, and server consoldaton. Algorthm 1: getmgrated_vm (goal) 1. swtch (goal){ 2. case hotspot mtgaton: 3. for all VMs runnng n ths hotspot PM 4. fnd the hghest value of score(x) 5. return VMx 6. case load balancng: 7. order_allpmlst(); /* accordng to the resource utlzaton of every PM n descendng order. */ 8. P=the PM n the top of AllPMlst 9. fnd the hghest value of score(x) n P 10. return VMx 11. case server consoldaton: 12. order_allpmlst(); /* accordng to the resource utlzaton of every PM n ascendng order. */ 13. P=the PM n the top of AllPMlst 14. fnd the hghest value of score(x) n P 15. return VMx 16. } Algorthm 2: gettarget_pm (VM) 1. named the trangle T whch contans VM.RRV 2. for all PMs do 3. f PM.RCV locates n T 4. add ths PM nto PotentalPMlst 5. end f 6. end for 7. f no PM.RCV locates T then 8. T=T+left trangle of T+rght trangle of T 9. end f 10. Remove PMs from PotentalPMlst whch can t support the VM 11. then 12. choose the most complementary PM n PotentalPMlst whose RIV s of the most closest magntude (s the most slghtly less) as the VM s RIV 13. return ths PM Algorthm 1 gets the VM whch should be mgrated away frstly based on three dfferent goals. Then algorthm 1 would pass ths VM as a parameter to algorthm 2. Algorthm 2 frstly generates the potentalpmlst by addng those PMs whose RCV locates n the same trangle wth the VM s RRV. At last, t chooses the most approprate PM n the potentalpmlst whose RIV s of same magntude as the VM s RIV and s n the opposte drecton. 4. EXPERIMENTAL RESULTS For evaluatng the performance of Smart-DRS, we have mplemented a prototype system as dscussed n Secton 3. Besdes, we buld a small-scale cluster whose detal nformaton s lsted as follows:

10 1546 Table 3. Experments envronment. PM VM CPU VMM Memory Network Numbers Numbers Frequency Verson GHz 4GB 100Mb/s Xen The major goal of our desgn s to evaluate the performance of Smart-DRS strategy whle appled n data center envronment. In order to avod tme-consumng and complcated mplementaton n code wthout knowng potental effects of the modfcaton, we choose some mature open source projects n our system, such as Gangla n montor module and Xen Moton n executon module. The detaled system archtecture s descrbed n Fg. 5. Fg. 5. The archtecture of experment prototype system. 4.1 Evaluaton of When to Schedule In order to evaluate the performance of predctng algorthm, we need some of convctve host load samples. Fortunately, Dnda and O Halloran from Carnege Mellon Unversty offered us plenty of load samples by long-term tracng wth many knds of machnes n a cluster system. We choose the day s collecton of load tme seres on August 18, 2010 as our test samples. In the frst experment, as shown n Fg. 6 (a), we contnuously montor 32 PMs for 21 mnutes and record the real load value. After that, we respectvely calculate the forecastng value for average when = 0.1, 0.3, 0.5, the Mean Relatve Error (MRE) s correspondngly 13.6%, 9.7% and 7.3%. That means SES algorthm has an acceptable predctng accuracy whle appled n a small-scale datacenter and = 0.5 s a better stuaton n our experment envronment. In the second experment, as shown n Fg. 6 (b), we compare Smart-DRS wth other classc forecastng algorthms such as AR (Autoregressve), MA (Movng Average), ARMA (Autoregressve Movng Average), ARIMA (Autoregressve Integrated Movng Average) and BP (Back Propagaton) neural network. We can see that ARMA predcton algorthm has the least mean relatve error, whch means t has the most accuracy, be-

11 AN INTEGRATED DYNAMIC RESOURCE SCHEDULING FRAMEWORK 1547 (a) Compared wth tself of predcton effcency. (b) Compared wth other predcton methods. Fg. 6. Evaluaton of when to schedule performance. cause t combnes AR s advantage wth MA s. Then our Smart-DRS follows, t has more accuracy that BP neural network whch s a knd of contnuous practcng algorthm. Even Smart-DRS s a relatvely smple model, t has a relatvely good performance, and the mean relatve error of t s acceptable. 4.2 Evaluaton of Whch to Schedule SandPper s a Xen based automated provsonng system [6]. It takes VM mgraton decson based on a metrc, whch refers to as sand_volume sand _ volume 1cpu 1mem 1net (9) Where cpu, mem and net are the correspondng normalzed utlzatons of the resources (n ths paper, for smplcty, we just consder PM as a 3D object). Accordng to our experence, n many experments, authors would lke choose the VM whose sand_ volume s the hghest as the mgrated VM. Because sand_volume value represents the resource utlzaton of the VM and mgratng the hghest one helps a lot to mtgate the resource utlzaton rate. However, dfferent VM choosng strategy wll affect the mgraton tme and mgraton numbers. In ths experment, we carry out a seres of 20 groups tests ncludng dfferent ntal statements of our experment cluster, such as hotspot, load unbalancng, low resource utlzaton and so on. We count the mgraton tme and mgraton numbers from ntal state to the fnal state between SandPper model and our score model of Smart-DRS. Fg. 7 (a) tells us that Smart-DRS can save much more mgraton tme whle just a lttle hgh n mgraton numbers compared to SandPper shown n Fg. 7 (b). Because Smart-DRS would lke to choose the VM whch utlzes hgh cpu and net bandwdth but low memory resource. Smart-DRS can reduce much more mgraton tme at the sacrfce of a lttle more mgraton numbers.

12 1548 (a) Mgraton Tme experment. (b) Mgraton Number experment. Fg. 7. Evaluaton of whch to schedule performance. 4.3 Evaluaton of Where to Schedule In ths experment, we try to evaluate the performance of VM placement. Frst of all, we should clam that we have defned Balance to represent the degree of load balancng of each machne and SYS balance to represent the degree of load balancng of the whole system (a hgher value s better). Balance (10) [ ]/2 cpu mem o n cpu cpu / n cpu 1 n mem mem / n mem 1 n o o / n o 1 SYS (11) n balance Balance n (12) 1 Ths experment compares the performance of Smart-DRS wth two classcal Bn Packng algorthms used n dynamc resource schedulng, Best Ft Decreasng (BFD) and Frst Ft Decreasng (FFD). We measure the algorthm executon tme and SYS balance of each algorthm. As shown n Fg. 8, the BFD and FFD consume much more executon Fg. 8. Performance comparsons among varous schedulng algorthms.

13 AN INTEGRATED DYNAMIC RESOURCE SCHEDULING FRAMEWORK 1549 tme and acheve no better performance n SYS balance than Smart-DRS. The reason s that they have to sort all of the potental VMs and target PMs frstly accordng to ther resource utlzaton. 4.4 Overall Evaluaton In the last experment, we assess the performance of Smart-DRS n overall stuaton. By contnuously montorng the resource utlzaton of cpu, mem, net of 32 PMs, we could observe the dfference between before schedulng and after schedulng. In order to carry out ths experment, we have set 0.8 as the upper threshold and 0.2 as the lower threshold. (a) Before schedulng (load balancng). (b) After schedulng (load balancng). (c) Before schedulng (server consoldaton). (d) After schedulng (server consoldaton). Fg. 9. The overall evaluaton between before and after schedulng. Fg. 9 shows that our schedulng algorthm really works. It s clear that the load of system s mbalanced and the resource utlzaton of several machnes exceeds 0.8 before schedulng as shown n Fg. 9 (a). Whle n Fg. 9 (b), after schedulng, the load of system s more balanced and none of resource utlzaton exceeds the upper threshold. Smlarly, n Fg. 9 (c) we can see that the resource utlzaton of several machnes s under 0.2 whle n Fg. 9 (d), these low utlzaton machnes are taken offlne and the others are load balancng. 5. CONCLUSION In ths paper, we studed the resource management method n on-demand cloud envronment and we have presented an ntegrated dynamc resource schedulng framework named Smart-DRS. It employs SES algorthm to predct the resource utlzaton of PMs

14 1550 n order to avod tny and temporary load peak value trggerng unnecessary mgraton. The experment result shows that predcton value s pretty close wth the real value. Then Smart-DRS creates score model to determne whch VM to choose. And the result tells us that t can save much more mgraton tme whch s very mportant to mprove QoS performance. At last, Smart-DRS employs VP algorthm to make the mgraton scheme. The reason why we choose VP algorthm s that t s a novel theory whch can be used to make the process of choosng PMs easer and more approprate. The experment result tells us that t works really well and t s a knd of low cost and effcent method. Nevertheless, our framework has some lmtatons that we plan to address n the future: (1) Frstly, we can t employ Smart-DRS on a large scale cloud data center, so we don t know ts scalablty; (2) Then, we wll comment on the computaton and space complexty of vector projecton algorthm n future work; (3) At last, we don t talk about the user prorty. In the real world, t s very lkely that some hgh prorty users or users who pay more for the resources need to occupy the whole physcal machnes. We can t move other VMs to these machnes even f they are actually dle. In addton, varous other measurements and optmzaton strateges wll need to be explored n the future. REFERENCES 1. F. Hermener, X. Lorca, J. M. Menaud, G. Muller, and J. Lawall, Entropy: a consoldaton manager for clusters, n Proceedngs of ACM SIGPLAN Internatonal Conference on Vrtual Executon Envronments, 2009, pp C. Hyser, B. Mckee, R. Gardner, and B. J. Watson, Autonomc vrtual machne placement n the data center, Hewlett Packard Laboratores, 2007, pp R. Nathuj and K. Schwan, VrtualPower: Coordnated power management n vrtualzed enterprse systems, n Proceedngs of the 21st ACM SIGOPS Symposum on Operatng Systems Prncples, 2007, pp J. D. Sonnek, J. B. S. G. Greensky, R. Reutman, and A. Chandra, Starlng: Mnmzng communcaton overhead n vrtualzed computng platforms usng decentralzed affnty-aware mgraton, n Proceedngs of Internatonal Conference on Parallel Processng, 2010, pp X. Meng, V. Pappas, and L. Zhang, Improvng the scalablty of data center networks wth traffc-aware vrtual machne placement, n Proceedngs of Internatonal Conference on Computer Communcatons, 2010, pp T. Wood, P. J. Shenoy, A. Venkataraman, and M. S. Yousf, Sandpper: Blackbox and gray-box resource management for vrtual machnes, Computer Networks, Vol. 53, 2009, pp M. Cardosa, M. Korupolu, and A. Sngh, Shares and utltes based power consoldaton n vrtualzed server envronments, n Proceedngs of IFIP/IEEE Internatonal Symposum on Integrated Network Management, 2009, pp A. Verma, P. Ahuja, and A. Neog, Power-aware dynamc placement of hpc applcatons, n Proceedngs of the 22nd Annual Internatonal Conference on Super- Computng, 2008, pp T. Wood, P. J. Shenoy, A. Venkataraman, and M. S. Yousf, Black-box and gray-

15 AN INTEGRATED DYNAMIC RESOURCE SCHEDULING FRAMEWORK 1551 box strateges for vrtual machne mgraton, n Proceedngs of the 4th ACM/USE- NIX Symposum on Networked Systems Desgn and Implementaton, 2007, pp M. Mshra, A. Das, P. Kulkarn, and A. Sahoo, Dynamc resource management usng vrtual machne mgratons, Communcatons Magazne, Vol. 50, 2012, pp F. Hermener, S. Demassey, and X. Lorca, Bn repackng schedulng n vrtualzed datacenters, n Proceedngs of the 17th Internatonal Conference on Prncples and Practce of Constrant Programmng, 2011, pp N. Bobroff, A. Kochut, and K. Beaty, Dynamc placement of vrtual machnes for managng sla volatons, n Proceedngs of the 10th Internatonal Symposum on Integrated Network Management, 2007, pp M. Wang, X. Meng, and L. Zhang, Consoldatng vrtual machnes wth dynamc bandwdth demand n data centers, n Proceedngs of IEEE Internatonal Conference on Computer Communcatons, 2011, pp M. Mshra and A. Sahoo, On theory of VM placement: Anomales n exstng methodoges and ther mgraton usng a novel vector based approach, n Proceedngs of IEEE 4th Internatonal Conference on Cloud Computng, 2011, pp J. L, J. Peng, Z. Le, and W. Zhang, An energy-effcent schedulng approach based on prvate clouds, Journal of Informaton Computatonal Scence, Vol. 8, 2011, pp L. Xu, W. Chen, Z. Wang, and S. Yang, Smart-DRS: A strategy of dynamc resource schedulng n cloud data center, n Proceedngs of IEEE Internatonal Conference on Cluster Computng Workshops, 2012, pp T. Imada, M. Sato, and H. Kmura, Power and qos performance characterstcs of vrtualzed servers, n Proceedngs of the 10th Internatonal Conference on Grd Computng, 2009, pp Le Xu () s now a Ph.D. student n the College of Computer Scence and Technology, Zhejang Unversty. Hs current research nterests nclude operatng system, vrtualzaton and cloud nfrastructure. Zonghu Wang ( 总 辉 ), born n March 1979, s a Lecturer n the College of Computer Scence and Engneerng at Zhejang Unversty n Hangzhou, Chna. He receved hs Ph.D. n the College of Computer Scence and Technology at Zhejang Unversty n Hs research nterests focuse on cloud computng, dstrbuted system and computer archtecture.

16 1552 Wenzh Chen ( 陈 ) born n 1969, receved hs MS and Ph.D. degrees from Zhejang Unversty, Hangzhou, Chna. He s now a Professor and Ph.D. supervsor of College of Computer Scence and Technology of Zhejang Unversty. Hs areas of research nclude computer archtecture, system software, embedded system and securty.

Survey on Virtual Machine Placement Techniques in Cloud Computing Environment

Survey on Virtual Machine Placement Techniques in Cloud Computing Environment Survey on Vrtual Machne Placement Technques n Cloud Computng Envronment Rajeev Kumar Gupta and R. K. Paterya Department of Computer Scence & Engneerng, MANIT, Bhopal, Inda ABSTRACT In tradtonal data center

More information

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis The Development of Web Log Mnng Based on Improve-K-Means Clusterng Analyss TngZhong Wang * College of Informaton Technology, Luoyang Normal Unversty, Luoyang, 471022, Chna wangtngzhong2@sna.cn Abstract.

More information

A Replication-Based and Fault Tolerant Allocation Algorithm for Cloud Computing

A Replication-Based and Fault Tolerant Allocation Algorithm for Cloud Computing A Replcaton-Based and Fault Tolerant Allocaton Algorthm for Cloud Computng Tork Altameem Dept of Computer Scence, RCC, Kng Saud Unversty, PO Box: 28095 11437 Ryadh-Saud Araba Abstract The very large nfrastructure

More information

Fault tolerance in cloud technologies presented as a service

Fault tolerance in cloud technologies presented as a service Internatonal Scentfc Conference Computer Scence 2015 Pavel Dzhunev, PhD student Fault tolerance n cloud technologes presented as a servce INTRODUCTION Improvements n technques for vrtualzaton and performance

More information

Open Access A Load Balancing Strategy with Bandwidth Constraint in Cloud Computing. Jing Deng 1,*, Ping Guo 2, Qi Li 3, Haizhu Chen 1

Open Access A Load Balancing Strategy with Bandwidth Constraint in Cloud Computing. Jing Deng 1,*, Ping Guo 2, Qi Li 3, Haizhu Chen 1 Send Orders for Reprnts to reprnts@benthamscence.ae The Open Cybernetcs & Systemcs Journal, 2014, 8, 115-121 115 Open Access A Load Balancng Strategy wth Bandwdth Constrant n Cloud Computng Jng Deng 1,*,

More information

A heuristic task deployment approach for load balancing

A heuristic task deployment approach for load balancing Xu Gaochao, Dong Yunmeng, Fu Xaodog, Dng Yan, Lu Peng, Zhao Ja Abstract A heurstc task deployment approach for load balancng Gaochao Xu, Yunmeng Dong, Xaodong Fu, Yan Dng, Peng Lu, Ja Zhao * College of

More information

Network Aware Load-Balancing via Parallel VM Migration for Data Centers

Network Aware Load-Balancing via Parallel VM Migration for Data Centers Network Aware Load-Balancng va Parallel VM Mgraton for Data Centers Kun-Tng Chen 2, Chen Chen 12, Po-Hsang Wang 2 1 Informaton Technology Servce Center, 2 Department of Computer Scence Natonal Chao Tung

More information

Fair Virtual Bandwidth Allocation Model in Virtual Data Centers

Fair Virtual Bandwidth Allocation Model in Virtual Data Centers Far Vrtual Bandwdth Allocaton Model n Vrtual Data Centers Yng Yuan, Cu-rong Wang, Cong Wang School of Informaton Scence and Engneerng ortheastern Unversty Shenyang, Chna School of Computer and Communcaton

More information

An Interest-Oriented Network Evolution Mechanism for Online Communities

An Interest-Oriented Network Evolution Mechanism for Online Communities An Interest-Orented Network Evoluton Mechansm for Onlne Communtes Cahong Sun and Xaopng Yang School of Informaton, Renmn Unversty of Chna, Bejng 100872, P.R. Chna {chsun,yang}@ruc.edu.cn Abstract. Onlne

More information

Forecasting the Demand of Emergency Supplies: Based on the CBR Theory and BP Neural Network

Forecasting the Demand of Emergency Supplies: Based on the CBR Theory and BP Neural Network 700 Proceedngs of the 8th Internatonal Conference on Innovaton & Management Forecastng the Demand of Emergency Supples: Based on the CBR Theory and BP Neural Network Fu Deqang, Lu Yun, L Changbng School

More information

The Greedy Method. Introduction. 0/1 Knapsack Problem

The Greedy Method. Introduction. 0/1 Knapsack Problem The Greedy Method Introducton We have completed data structures. We now are gong to look at algorthm desgn methods. Often we are lookng at optmzaton problems whose performance s exponental. For an optmzaton

More information

Feature selection for intrusion detection. Slobodan Petrović NISlab, Gjøvik University College

Feature selection for intrusion detection. Slobodan Petrović NISlab, Gjøvik University College Feature selecton for ntruson detecton Slobodan Petrovć NISlab, Gjøvk Unversty College Contents The feature selecton problem Intruson detecton Traffc features relevant for IDS The CFS measure The mrmr measure

More information

A Secure Password-Authenticated Key Agreement Using Smart Cards

A Secure Password-Authenticated Key Agreement Using Smart Cards A Secure Password-Authentcated Key Agreement Usng Smart Cards Ka Chan 1, Wen-Chung Kuo 2 and Jn-Chou Cheng 3 1 Department of Computer and Informaton Scence, R.O.C. Mltary Academy, Kaohsung 83059, Tawan,

More information

Forecasting the Direction and Strength of Stock Market Movement

Forecasting the Direction and Strength of Stock Market Movement Forecastng the Drecton and Strength of Stock Market Movement Jngwe Chen Mng Chen Nan Ye cjngwe@stanford.edu mchen5@stanford.edu nanye@stanford.edu Abstract - Stock market s one of the most complcated systems

More information

A Dynamic Energy-Efficiency Mechanism for Data Center Networks

A Dynamic Energy-Efficiency Mechanism for Data Center Networks A Dynamc Energy-Effcency Mechansm for Data Center Networks Sun Lang, Zhang Jnfang, Huang Daochao, Yang Dong, Qn Yajuan A Dynamc Energy-Effcency Mechansm for Data Center Networks 1 Sun Lang, 1 Zhang Jnfang,

More information

A Programming Model for the Cloud Platform

A Programming Model for the Cloud Platform Internatonal Journal of Advanced Scence and Technology A Programmng Model for the Cloud Platform Xaodong Lu School of Computer Engneerng and Scence Shangha Unversty, Shangha 200072, Chna luxaodongxht@qq.com

More information

Data Broadcast on a Multi-System Heterogeneous Overlayed Wireless Network *

Data Broadcast on a Multi-System Heterogeneous Overlayed Wireless Network * JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 24, 819-840 (2008) Data Broadcast on a Mult-System Heterogeneous Overlayed Wreless Network * Department of Computer Scence Natonal Chao Tung Unversty Hsnchu,

More information

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module LOSSLESS IMAGE COMPRESSION SYSTEMS Lesson 3 Lossless Compresson: Huffman Codng Instructonal Objectves At the end of ths lesson, the students should be able to:. Defne and measure source entropy..

More information

A Load-Balancing Algorithm for Cluster-based Multi-core Web Servers

A Load-Balancing Algorithm for Cluster-based Multi-core Web Servers Journal of Computatonal Informaton Systems 7: 13 (2011) 4740-4747 Avalable at http://www.jofcs.com A Load-Balancng Algorthm for Cluster-based Mult-core Web Servers Guohua YOU, Yng ZHAO College of Informaton

More information

IMPACT ANALYSIS OF A CELLULAR PHONE

IMPACT ANALYSIS OF A CELLULAR PHONE 4 th ASA & μeta Internatonal Conference IMPACT AALYSIS OF A CELLULAR PHOE We Lu, 2 Hongy L Bejng FEAonlne Engneerng Co.,Ltd. Bejng, Chna ABSTRACT Drop test smulaton plays an mportant role n nvestgatng

More information

A Design Method of High-availability and Low-optical-loss Optical Aggregation Network Architecture

A Design Method of High-availability and Low-optical-loss Optical Aggregation Network Architecture A Desgn Method of Hgh-avalablty and Low-optcal-loss Optcal Aggregaton Network Archtecture Takehro Sato, Kuntaka Ashzawa, Kazumasa Tokuhash, Dasuke Ish, Satoru Okamoto and Naoak Yamanaka Dept. of Informaton

More information

INVESTIGATION OF VEHICULAR USERS FAIRNESS IN CDMA-HDR NETWORKS

INVESTIGATION OF VEHICULAR USERS FAIRNESS IN CDMA-HDR NETWORKS 21 22 September 2007, BULGARIA 119 Proceedngs of the Internatonal Conference on Informaton Technologes (InfoTech-2007) 21 st 22 nd September 2007, Bulgara vol. 2 INVESTIGATION OF VEHICULAR USERS FAIRNESS

More information

An Evaluation of the Extended Logistic, Simple Logistic, and Gompertz Models for Forecasting Short Lifecycle Products and Services

An Evaluation of the Extended Logistic, Simple Logistic, and Gompertz Models for Forecasting Short Lifecycle Products and Services An Evaluaton of the Extended Logstc, Smple Logstc, and Gompertz Models for Forecastng Short Lfecycle Products and Servces Charles V. Trappey a,1, Hsn-yng Wu b a Professor (Management Scence), Natonal Chao

More information

An Alternative Way to Measure Private Equity Performance

An Alternative Way to Measure Private Equity Performance An Alternatve Way to Measure Prvate Equty Performance Peter Todd Parlux Investment Technology LLC Summary Internal Rate of Return (IRR) s probably the most common way to measure the performance of prvate

More information

IWFMS: An Internal Workflow Management System/Optimizer for Hadoop

IWFMS: An Internal Workflow Management System/Optimizer for Hadoop IWFMS: An Internal Workflow Management System/Optmzer for Hadoop Lan Lu, Yao Shen Department of Computer Scence and Engneerng Shangha JaoTong Unversty Shangha, Chna lustrve@gmal.com, yshen@cs.sjtu.edu.cn

More information

A New Task Scheduling Algorithm Based on Improved Genetic Algorithm

A New Task Scheduling Algorithm Based on Improved Genetic Algorithm A New Task Schedulng Algorthm Based on Improved Genetc Algorthm n Cloud Computng Envronment Congcong Xong, Long Feng, Lxan Chen A New Task Schedulng Algorthm Based on Improved Genetc Algorthm n Cloud Computng

More information

Load Balancing By Max-Min Algorithm in Private Cloud Environment

Load Balancing By Max-Min Algorithm in Private Cloud Environment Internatonal Journal of Scence and Research (IJSR ISSN (Onlne: 2319-7064 Index Coperncus Value (2013: 6.14 Impact Factor (2013: 4.438 Load Balancng By Max-Mn Algorthm n Prvate Cloud Envronment S M S Suntharam

More information

Vision Mouse. Saurabh Sarkar a* University of Cincinnati, Cincinnati, USA ABSTRACT 1. INTRODUCTION

Vision Mouse. Saurabh Sarkar a* University of Cincinnati, Cincinnati, USA ABSTRACT 1. INTRODUCTION Vson Mouse Saurabh Sarkar a* a Unversty of Cncnnat, Cncnnat, USA ABSTRACT The report dscusses a vson based approach towards trackng of eyes and fngers. The report descrbes the process of locatng the possble

More information

The Load Balancing of Database Allocation in the Cloud

The Load Balancing of Database Allocation in the Cloud , March 3-5, 23, Hong Kong The Load Balancng of Database Allocaton n the Cloud Yu-lung Lo and Mn-Shan La Abstract Each database host n the cloud platform often has to servce more than one database applcaton

More information

On the Optimal Control of a Cascade of Hydro-Electric Power Stations

On the Optimal Control of a Cascade of Hydro-Electric Power Stations On the Optmal Control of a Cascade of Hydro-Electrc Power Statons M.C.M. Guedes a, A.F. Rbero a, G.V. Smrnov b and S. Vlela c a Department of Mathematcs, School of Scences, Unversty of Porto, Portugal;

More information

A hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm

A hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm Avalable onlne www.ocpr.com Journal of Chemcal and Pharmaceutcal Research, 2014, 6(7):1884-1889 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCPRC5 A hybrd global optmzaton algorthm based on parallel

More information

M3S MULTIMEDIA MOBILITY MANAGEMENT AND LOAD BALANCING IN WIRELESS BROADCAST NETWORKS

M3S MULTIMEDIA MOBILITY MANAGEMENT AND LOAD BALANCING IN WIRELESS BROADCAST NETWORKS M3S MULTIMEDIA MOBILITY MANAGEMENT AND LOAD BALANCING IN WIRELESS BROADCAST NETWORKS Bogdan Cubotaru, Gabrel-Mro Muntean Performance Engneerng Laboratory, RINCE School of Electronc Engneerng Dubln Cty

More information

J. Parallel Distrib. Comput.

J. Parallel Distrib. Comput. J. Parallel Dstrb. Comput. 71 (2011) 62 76 Contents lsts avalable at ScenceDrect J. Parallel Dstrb. Comput. journal homepage: www.elsever.com/locate/jpdc Optmzng server placement n dstrbuted systems n

More information

Support Vector Machines

Support Vector Machines Support Vector Machnes Max Wellng Department of Computer Scence Unversty of Toronto 10 Kng s College Road Toronto, M5S 3G5 Canada wellng@cs.toronto.edu Abstract Ths s a note to explan support vector machnes.

More information

PAS: A Packet Accounting System to Limit the Effects of DoS & DDoS. Debish Fesehaye & Klara Naherstedt University of Illinois-Urbana Champaign

PAS: A Packet Accounting System to Limit the Effects of DoS & DDoS. Debish Fesehaye & Klara Naherstedt University of Illinois-Urbana Champaign PAS: A Packet Accountng System to Lmt the Effects of DoS & DDoS Debsh Fesehaye & Klara Naherstedt Unversty of Illnos-Urbana Champagn DoS and DDoS DDoS attacks are ncreasng threats to our dgtal world. Exstng

More information

DBA-VM: Dynamic Bandwidth Allocator for Virtual Machines

DBA-VM: Dynamic Bandwidth Allocator for Virtual Machines DBA-VM: Dynamc Bandwdth Allocator for Vrtual Machnes Ahmed Amamou, Manel Bourguba, Kamel Haddadou and Guy Pujolle LIP6, Perre & Mare Cure Unversty, 4 Place Jusseu 755 Pars, France Gand SAS, 65 Boulevard

More information

Thermal-aware relocation of servers in green data centers

Thermal-aware relocation of servers in green data centers Fronters of Informaton Technology & Electronc Engneerng www.zju.edu.cn/jzus; engneerng.cae.cn; www.sprngerlnk.com ISSN 295-9184 (prnt); ISSN 295-923 (onlne) E-mal: jzus@zju.edu.cn Chaudhry et al. / Front

More information

Luby s Alg. for Maximal Independent Sets using Pairwise Independence

Luby s Alg. for Maximal Independent Sets using Pairwise Independence Lecture Notes for Randomzed Algorthms Luby s Alg. for Maxmal Independent Sets usng Parwse Independence Last Updated by Erc Vgoda on February, 006 8. Maxmal Independent Sets For a graph G = (V, E), an ndependent

More information

Dynamic Fleet Management for Cybercars

Dynamic Fleet Management for Cybercars Proceedngs of the IEEE ITSC 2006 2006 IEEE Intellgent Transportaton Systems Conference Toronto, Canada, September 17-20, 2006 TC7.5 Dynamc Fleet Management for Cybercars Fenghu. Wang, Mng. Yang, Ruqng.

More information

Cloud Auto-Scaling with Deadline and Budget Constraints

Cloud Auto-Scaling with Deadline and Budget Constraints Prelmnary verson. Fnal verson appears In Proceedngs of 11th ACM/IEEE Internatonal Conference on Grd Computng (Grd 21). Oct 25-28, 21. Brussels, Belgum. Cloud Auto-Scalng wth Deadlne and Budget Constrants

More information

APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT

APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT APPLICATION OF PROBE DATA COLLECTED VIA INFRARED BEACONS TO TRAFFIC MANEGEMENT Toshhko Oda (1), Kochro Iwaoka (2) (1), (2) Infrastructure Systems Busness Unt, Panasonc System Networks Co., Ltd. Saedo-cho

More information

Improved SVM in Cloud Computing Information Mining

Improved SVM in Cloud Computing Information Mining Internatonal Journal of Grd Dstrbuton Computng Vol.8, No.1 (015), pp.33-40 http://dx.do.org/10.1457/jgdc.015.8.1.04 Improved n Cloud Computng Informaton Mnng Lvshuhong (ZhengDe polytechnc college JangSu

More information

J. Parallel Distrib. Comput. Environment-conscious scheduling of HPC applications on distributed Cloud-oriented data centers

J. Parallel Distrib. Comput. Environment-conscious scheduling of HPC applications on distributed Cloud-oriented data centers J. Parallel Dstrb. Comput. 71 (2011) 732 749 Contents lsts avalable at ScenceDrect J. Parallel Dstrb. Comput. ournal homepage: www.elsever.com/locate/pdc Envronment-conscous schedulng of HPC applcatons

More information

The OC Curve of Attribute Acceptance Plans

The OC Curve of Attribute Acceptance Plans The OC Curve of Attrbute Acceptance Plans The Operatng Characterstc (OC) curve descrbes the probablty of acceptng a lot as a functon of the lot s qualty. Fgure 1 shows a typcal OC Curve. 10 8 6 4 1 3 4

More information

A Performance Analysis of View Maintenance Techniques for Data Warehouses

A Performance Analysis of View Maintenance Techniques for Data Warehouses A Performance Analyss of Vew Mantenance Technques for Data Warehouses Xng Wang Dell Computer Corporaton Round Roc, Texas Le Gruenwald The nversty of Olahoma School of Computer Scence orman, OK 739 Guangtao

More information

denote the location of a node, and suppose node X . This transmission causes a successful reception by node X for any other node

denote the location of a node, and suppose node X . This transmission causes a successful reception by node X for any other node Fnal Report of EE359 Class Proect Throughput and Delay n Wreless Ad Hoc Networs Changhua He changhua@stanford.edu Abstract: Networ throughput and pacet delay are the two most mportant parameters to evaluate

More information

Vehicle Detection and Tracking in Video from Moving Airborne Platform

Vehicle Detection and Tracking in Video from Moving Airborne Platform Journal of Computatonal Informaton Systems 10: 12 (2014) 4965 4972 Avalable at http://www.jofcs.com Vehcle Detecton and Trackng n Vdeo from Movng Arborne Platform Lye ZHANG 1,2,, Hua WANG 3, L LI 2 1 School

More information

Effective Network Defense Strategies against Malicious Attacks with Various Defense Mechanisms under Quality of Service Constraints

Effective Network Defense Strategies against Malicious Attacks with Various Defense Mechanisms under Quality of Service Constraints Effectve Network Defense Strateges aganst Malcous Attacks wth Varous Defense Mechansms under Qualty of Servce Constrants Frank Yeong-Sung Ln Department of Informaton Natonal Tawan Unversty Tape, Tawan,

More information

A DATA MINING APPLICATION IN A STUDENT DATABASE

A DATA MINING APPLICATION IN A STUDENT DATABASE JOURNAL OF AERONAUTICS AND SPACE TECHNOLOGIES JULY 005 VOLUME NUMBER (53-57) A DATA MINING APPLICATION IN A STUDENT DATABASE Şenol Zafer ERDOĞAN Maltepe Ünversty Faculty of Engneerng Büyükbakkalköy-Istanbul

More information

A Dynamic Load Balancing for Massive Multiplayer Online Game Server

A Dynamic Load Balancing for Massive Multiplayer Online Game Server A Dynamc Load Balancng for Massve Multplayer Onlne Game Server Jungyoul Lm, Jaeyong Chung, Jnryong Km and Kwanghyun Shm Dgtal Content Research Dvson Electroncs and Telecommuncatons Research Insttute Daejeon,

More information

How To Solve An Onlne Control Polcy On A Vrtualzed Data Center

How To Solve An Onlne Control Polcy On A Vrtualzed Data Center Dynamc Resource Allocaton and Power Management n Vrtualzed Data Centers Rahul Urgaonkar, Ulas C. Kozat, Ken Igarash, Mchael J. Neely urgaonka@usc.edu, {kozat, garash}@docomolabs-usa.com, mjneely@usc.edu

More information

How Sets of Coherent Probabilities May Serve as Models for Degrees of Incoherence

How Sets of Coherent Probabilities May Serve as Models for Degrees of Incoherence 1 st Internatonal Symposum on Imprecse Probabltes and Ther Applcatons, Ghent, Belgum, 29 June 2 July 1999 How Sets of Coherent Probabltes May Serve as Models for Degrees of Incoherence Mar J. Schervsh

More information

Study on Model of Risks Assessment of Standard Operation in Rural Power Network

Study on Model of Risks Assessment of Standard Operation in Rural Power Network Study on Model of Rsks Assessment of Standard Operaton n Rural Power Network Qngj L 1, Tao Yang 2 1 Qngj L, College of Informaton and Electrcal Engneerng, Shenyang Agrculture Unversty, Shenyang 110866,

More information

An MILP model for planning of batch plants operating in a campaign-mode

An MILP model for planning of batch plants operating in a campaign-mode An MILP model for plannng of batch plants operatng n a campagn-mode Yanna Fumero Insttuto de Desarrollo y Dseño CONICET UTN yfumero@santafe-concet.gov.ar Gabrela Corsano Insttuto de Desarrollo y Dseño

More information

Network Security Situation Evaluation Method for Distributed Denial of Service

Network Security Situation Evaluation Method for Distributed Denial of Service Network Securty Stuaton Evaluaton Method for Dstrbuted Denal of Servce Jn Q,2, Cu YMn,2, Huang MnHuan,2, Kuang XaoHu,2, TangHong,2 ) Scence and Technology on Informaton System Securty Laboratory, Bejng,

More information

Introduction CONTENT. - Whitepaper -

Introduction CONTENT. - Whitepaper - OneCl oud ForAl l YourCr t c al Bus nes sappl c at ons Bl uew r esol ut ons www. bl uew r e. c o. uk Introducton Bluewre Cloud s a fully customsable IaaS cloud platform desgned for organsatons who want

More information

DEFINING %COMPLETE IN MICROSOFT PROJECT

DEFINING %COMPLETE IN MICROSOFT PROJECT CelersSystems DEFINING %COMPLETE IN MICROSOFT PROJECT PREPARED BY James E Aksel, PMP, PMI-SP, MVP For Addtonal Informaton about Earned Value Management Systems and reportng, please contact: CelersSystems,

More information

Project Networks With Mixed-Time Constraints

Project Networks With Mixed-Time Constraints Project Networs Wth Mxed-Tme Constrants L Caccetta and B Wattananon Western Australan Centre of Excellence n Industral Optmsaton (WACEIO) Curtn Unversty of Technology GPO Box U1987 Perth Western Australa

More information

Cloud-based Social Application Deployment using Local Processing and Global Distribution

Cloud-based Social Application Deployment using Local Processing and Global Distribution Cloud-based Socal Applcaton Deployment usng Local Processng and Global Dstrbuton Zh Wang *, Baochun L, Lfeng Sun *, and Shqang Yang * * Bejng Key Laboratory of Networked Multmeda Department of Computer

More information

Logical Development Of Vogel s Approximation Method (LD-VAM): An Approach To Find Basic Feasible Solution Of Transportation Problem

Logical Development Of Vogel s Approximation Method (LD-VAM): An Approach To Find Basic Feasible Solution Of Transportation Problem INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME, ISSUE, FEBRUARY ISSN 77-866 Logcal Development Of Vogel s Approxmaton Method (LD- An Approach To Fnd Basc Feasble Soluton Of Transportaton

More information

"Research Note" APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES *

Research Note APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES * Iranan Journal of Scence & Technology, Transacton B, Engneerng, ol. 30, No. B6, 789-794 rnted n The Islamc Republc of Iran, 006 Shraz Unversty "Research Note" ALICATION OF CHARGE SIMULATION METHOD TO ELECTRIC

More information

Resource Scheduling in Desktop Grid by Grid-JQA

Resource Scheduling in Desktop Grid by Grid-JQA The 3rd Internatonal Conference on Grd and Pervasve Computng - Worshops esource Schedulng n Destop Grd by Grd-JQA L. Mohammad Khanl M. Analou Assstant professor Assstant professor C.S. Dept.Tabrz Unversty

More information

Multi-timescale Distributed Capacity Allocation and Load Redirect Algorithms for Cloud System

Multi-timescale Distributed Capacity Allocation and Load Redirect Algorithms for Cloud System Mult-tmescale Dstrbuted Capacty Allocaton and Load Redrect Algorthms for Cloud System Danlo Ardagna, Sara Casolar Mchele Colajann, Barbara Pancucc Report n. 2011.23 Mult-tme Scale Dstrbuted Capacty Allocaton

More information

What is Candidate Sampling

What is Candidate Sampling What s Canddate Samplng Say we have a multclass or mult label problem where each tranng example ( x, T ) conssts of a context x a small (mult)set of target classes T out of a large unverse L of possble

More information

Master s Thesis. Configuring robust virtual wireless sensor networks for Internet of Things inspired by brain functional networks

Master s Thesis. Configuring robust virtual wireless sensor networks for Internet of Things inspired by brain functional networks Master s Thess Ttle Confgurng robust vrtual wreless sensor networks for Internet of Thngs nspred by bran functonal networks Supervsor Professor Masayuk Murata Author Shnya Toyonaga February 10th, 2014

More information

Answer: A). There is a flatter IS curve in the high MPC economy. Original LM LM after increase in M. IS curve for low MPC economy

Answer: A). There is a flatter IS curve in the high MPC economy. Original LM LM after increase in M. IS curve for low MPC economy 4.02 Quz Solutons Fall 2004 Multple-Choce Questons (30/00 ponts) Please, crcle the correct answer for each of the followng 0 multple-choce questons. For each queston, only one of the answers s correct.

More information

Methodology to Determine Relationships between Performance Factors in Hadoop Cloud Computing Applications

Methodology to Determine Relationships between Performance Factors in Hadoop Cloud Computing Applications Methodology to Determne Relatonshps between Performance Factors n Hadoop Cloud Computng Applcatons Lus Eduardo Bautsta Vllalpando 1,2, Alan Aprl 1 and Alan Abran 1 1 Department of Software Engneerng and

More information

A Fast Incremental Spectral Clustering for Large Data Sets

A Fast Incremental Spectral Clustering for Large Data Sets 2011 12th Internatonal Conference on Parallel and Dstrbuted Computng, Applcatons and Technologes A Fast Incremental Spectral Clusterng for Large Data Sets Tengteng Kong 1,YeTan 1, Hong Shen 1,2 1 School

More information

Genetic Algorithm Based Optimization Model for Reliable Data Storage in Cloud Environment

Genetic Algorithm Based Optimization Model for Reliable Data Storage in Cloud Environment Advanced Scence and Technology Letters, pp.74-79 http://dx.do.org/10.14257/astl.2014.50.12 Genetc Algorthm Based Optmzaton Model for Relable Data Storage n Cloud Envronment Feng Lu 1,2,3, Hatao Wu 1,3,

More information

Using Elasticity to Improve Inline Data Deduplication Storage Systems

Using Elasticity to Improve Inline Data Deduplication Storage Systems Usng Elastcty to Improve Inlne Data Deduplcaton Storage Systems Yufeng Wang Temple Unversty Phladelpha, PA, USA Y.F.Wang@temple.edu Chu C Tan Temple Unversty Phladelpha, PA, USA cctan@temple.edu Nngfang

More information

Resource Scheduling Based on Dynamic Dependence Injection in Virtualization-based Simulation Grid

Resource Scheduling Based on Dynamic Dependence Injection in Virtualization-based Simulation Grid Proceedngs of the 200 4th Internatonal Conference on Computer Supported Cooperatve Work n Desgn Resource Schedulng Based on Dynamc Dependence Injecton n Vrtualzaton-based Smulaton Grd Hanbng Lu,Hongy Su,

More information

Reliable State Monitoring in Cloud Datacenters

Reliable State Monitoring in Cloud Datacenters Relable State Montorng n Cloud Datacenters Shcong Meng Arun K. Iyengar Isabelle M. Rouvellou Lng Lu Ksung Lee Balaj Palansamy Yuzhe Tang College of Computng, Georga Insttute of Technology, Atlanta, GA

More information

A Resource-trading Mechanism for Efficient Distribution of Large-volume Contents on Peer-to-Peer Networks

A Resource-trading Mechanism for Efficient Distribution of Large-volume Contents on Peer-to-Peer Networks A Resource-tradng Mechansm for Effcent Dstrbuton of Large-volume Contents on Peer-to-Peer Networks SmonG.M.Koo,C.S.GeorgeLee, Karthk Kannan School of Electrcal and Computer Engneerng Krannet School of

More information

Rate Monotonic (RM) Disadvantages of cyclic. TDDB47 Real Time Systems. Lecture 2: RM & EDF. Priority-based scheduling. States of a process

Rate Monotonic (RM) Disadvantages of cyclic. TDDB47 Real Time Systems. Lecture 2: RM & EDF. Priority-based scheduling. States of a process Dsadvantages of cyclc TDDB47 Real Tme Systems Manual scheduler constructon Cannot deal wth any runtme changes What happens f we add a task to the set? Real-Tme Systems Laboratory Department of Computer

More information

NEURO-FUZZY INFERENCE SYSTEM FOR E-COMMERCE WEBSITE EVALUATION

NEURO-FUZZY INFERENCE SYSTEM FOR E-COMMERCE WEBSITE EVALUATION NEURO-FUZZY INFERENE SYSTEM FOR E-OMMERE WEBSITE EVALUATION Huan Lu, School of Software, Harbn Unversty of Scence and Technology, Harbn, hna Faculty of Appled Mathematcs and omputer Scence, Belarusan State

More information

Calculating the high frequency transmission line parameters of power cables

Calculating the high frequency transmission line parameters of power cables < ' Calculatng the hgh frequency transmsson lne parameters of power cables Authors: Dr. John Dcknson, Laboratory Servces Manager, N 0 RW E B Communcatons Mr. Peter J. Ncholson, Project Assgnment Manager,

More information

Enabling P2P One-view Multi-party Video Conferencing

Enabling P2P One-view Multi-party Video Conferencing Enablng P2P One-vew Mult-party Vdeo Conferencng Yongxang Zhao, Yong Lu, Changja Chen, and JanYn Zhang Abstract Mult-Party Vdeo Conferencng (MPVC) facltates realtme group nteracton between users. Whle P2P

More information

How To Understand The Results Of The German Meris Cloud And Water Vapour Product

How To Understand The Results Of The German Meris Cloud And Water Vapour Product Ttel: Project: Doc. No.: MERIS level 3 cloud and water vapour products MAPP MAPP-ATBD-ClWVL3 Issue: 1 Revson: 0 Date: 9.12.1998 Functon Name Organsaton Sgnature Date Author: Bennartz FUB Preusker FUB Schüller

More information

Traffic State Estimation in the Traffic Management Center of Berlin

Traffic State Estimation in the Traffic Management Center of Berlin Traffc State Estmaton n the Traffc Management Center of Berln Authors: Peter Vortsch, PTV AG, Stumpfstrasse, D-763 Karlsruhe, Germany phone ++49/72/965/35, emal peter.vortsch@ptv.de Peter Möhl, PTV AG,

More information

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ).

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ). REVIEW OF RISK MANAGEMENT CONCEPTS LOSS DISTRIBUTIONS AND INSURANCE Loss and nsurance: When someone s subject to the rsk of ncurrng a fnancal loss, the loss s generally modeled usng a random varable or

More information

2. SYSTEM MODEL. the SLA (unlike the only other related mechanism [15] we can compare it is never able to meet the SLA).

2. SYSTEM MODEL. the SLA (unlike the only other related mechanism [15] we can compare it is never able to meet the SLA). Managng Server Energy and Operatonal Costs n Hostng Centers Yyu Chen Dept. of IE Penn State Unversty Unversty Park, PA 16802 yzc107@psu.edu Anand Svasubramanam Dept. of CSE Penn State Unversty Unversty

More information

Efficient Bandwidth Management in Broadband Wireless Access Systems Using CAC-based Dynamic Pricing

Efficient Bandwidth Management in Broadband Wireless Access Systems Using CAC-based Dynamic Pricing Effcent Bandwdth Management n Broadband Wreless Access Systems Usng CAC-based Dynamc Prcng Bader Al-Manthar, Ndal Nasser 2, Najah Abu Al 3, Hossam Hassanen Telecommuncatons Research Laboratory School of

More information

Hosting Virtual Machines on Distributed Datacenters

Hosting Virtual Machines on Distributed Datacenters Hostng Vrtual Machnes on Dstrbuted Datacenters Chuan Pham Scence and Engneerng, KyungHee Unversty, Korea pchuan@khu.ac.kr Jae Hyeok Son Scence and Engneerng, KyungHee Unversty, Korea sonaehyeok@khu.ac.kr

More information

Flexible Distributed Capacity Allocation and Load Redirect Algorithms for Cloud Systems

Flexible Distributed Capacity Allocation and Load Redirect Algorithms for Cloud Systems 2011 IEEE 4th Internatonal Conference on Cloud Computng Flexble Dstrbuted Capacty Allocaton and Load Redrect Algorthms for Cloud Systems Danlo Ardagna, Sara Casolar, Barbara Pancucc Poltecnco d Mlano,

More information

HP Mission-Critical Services

HP Mission-Critical Services HP Msson-Crtcal Servces Delverng busness value to IT Jelena Bratc Zarko Subotc TS Support tm Mart 2012, Podgorca 2010 Hewlett-Packard Development Company, L.P. The nformaton contaned heren s subject to

More information

How To Plan A Network Wide Load Balancing Route For A Network Wde Network (Network)

How To Plan A Network Wide Load Balancing Route For A Network Wde Network (Network) Network-Wde Load Balancng Routng Wth Performance Guarantees Kartk Gopalan Tz-cker Chueh Yow-Jan Ln Florda State Unversty Stony Brook Unversty Telcorda Research kartk@cs.fsu.edu chueh@cs.sunysb.edu yjln@research.telcorda.com

More information

Recurrence. 1 Definitions and main statements

Recurrence. 1 Definitions and main statements Recurrence 1 Defntons and man statements Let X n, n = 0, 1, 2,... be a MC wth the state space S = (1, 2,...), transton probabltes p j = P {X n+1 = j X n = }, and the transton matrx P = (p j ),j S def.

More information

Application of Multi-Agents for Fault Detection and Reconfiguration of Power Distribution Systems

Application of Multi-Agents for Fault Detection and Reconfiguration of Power Distribution Systems 1 Applcaton of Mult-Agents for Fault Detecton and Reconfguraton of Power Dstrbuton Systems K. Nareshkumar, Member, IEEE, M. A. Choudhry, Senor Member, IEEE, J. La, A. Felach, Senor Member, IEEE Abstract--The

More information

P2P/ Grid-based Overlay Architecture to Support VoIP Services in Large Scale IP Networks

P2P/ Grid-based Overlay Architecture to Support VoIP Services in Large Scale IP Networks PP/ Grd-based Overlay Archtecture to Support VoIP Servces n Large Scale IP Networks We Yu *, Srram Chellappan # and Dong Xuan # * Dept. of Computer Scence, Texas A&M Unversty, U.S.A. {weyu}@cs.tamu.edu

More information

Damage detection in composite laminates using coin-tap method

Damage detection in composite laminates using coin-tap method Damage detecton n composte lamnates usng con-tap method S.J. Km Korea Aerospace Research Insttute, 45 Eoeun-Dong, Youseong-Gu, 35-333 Daejeon, Republc of Korea yaeln@kar.re.kr 45 The con-tap test has the

More information

METHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS

METHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS METHODOLOGY TO DETERMINE RELATIONSHIPS BETWEEN PERFORMANCE FACTORS IN HADOOP CLOUD COMPUTING APPLICATIONS Lus Eduardo Bautsta Vllalpando 1,2, Alan Aprl 1 and Alan Abran 1 1 Department of Software Engneerng

More information

An Analysis of Central Processor Scheduling in Multiprogrammed Computer Systems

An Analysis of Central Processor Scheduling in Multiprogrammed Computer Systems STAN-CS-73-355 I SU-SE-73-013 An Analyss of Central Processor Schedulng n Multprogrammed Computer Systems (Dgest Edton) by Thomas G. Prce October 1972 Techncal Report No. 57 Reproducton n whole or n part

More information

An Energy-Efficient Data Placement Algorithm and Node Scheduling Strategies in Cloud Computing Systems

An Energy-Efficient Data Placement Algorithm and Node Scheduling Strategies in Cloud Computing Systems 2nd Internatonal Conference on Advances n Computer Scence and Engneerng (CSE 2013) An Energy-Effcent Data Placement Algorthm and Node Schedulng Strateges n Cloud Computng Systems Yanwen Xao Massve Data

More information

On the Interaction between Load Balancing and Speed Scaling

On the Interaction between Load Balancing and Speed Scaling On the Interacton between Load Balancng and Speed Scalng Ljun Chen and Na L Abstract Speed scalng has been wdely adopted n computer and communcaton systems, n partcular, to reduce energy consumpton. An

More information

Period and Deadline Selection for Schedulability in Real-Time Systems

Period and Deadline Selection for Schedulability in Real-Time Systems Perod and Deadlne Selecton for Schedulablty n Real-Tme Systems Thdapat Chantem, Xaofeng Wang, M.D. Lemmon, and X. Sharon Hu Department of Computer Scence and Engneerng, Department of Electrcal Engneerng

More information

LITERATURE REVIEW: VARIOUS PRIORITY BASED TASK SCHEDULING ALGORITHMS IN CLOUD COMPUTING

LITERATURE REVIEW: VARIOUS PRIORITY BASED TASK SCHEDULING ALGORITHMS IN CLOUD COMPUTING LITERATURE REVIEW: VARIOUS PRIORITY BASED TASK SCHEDULING ALGORITHMS IN CLOUD COMPUTING 1 MS. POOJA.P.VASANI, 2 MR. NISHANT.S. SANGHANI 1 M.Tech. [Software Systems] Student, Patel College of Scence and

More information

Research on Evaluation of Customer Experience of B2C Ecommerce Logistics Enterprises

Research on Evaluation of Customer Experience of B2C Ecommerce Logistics Enterprises 3rd Internatonal Conference on Educaton, Management, Arts, Economcs and Socal Scence (ICEMAESS 2015) Research on Evaluaton of Customer Experence of B2C Ecommerce Logstcs Enterprses Yle Pe1, a, Wanxn Xue1,

More information

SPEE Recommended Evaluation Practice #6 Definition of Decline Curve Parameters Background:

SPEE Recommended Evaluation Practice #6 Definition of Decline Curve Parameters Background: SPEE Recommended Evaluaton Practce #6 efnton of eclne Curve Parameters Background: The producton hstores of ol and gas wells can be analyzed to estmate reserves and future ol and gas producton rates and

More information

Intelligent Method for Cloud Task Scheduling Based on Particle Swarm Optimization Algorithm

Intelligent Method for Cloud Task Scheduling Based on Particle Swarm Optimization Algorithm Unversty of Nzwa, Oman December 9-11, 2014 Page 39 THE INTERNATIONAL ARAB CONFERENCE ON INFORMATION TECHNOLOGY (ACIT2014) Intellgent Method for Cloud Task Schedulng Based on Partcle Swarm Optmzaton Algorthm

More information

Virtual Network Embedding with Coordinated Node and Link Mapping

Virtual Network Embedding with Coordinated Node and Link Mapping Vrtual Network Embeddng wth Coordnated Node and Lnk Mappng N. M. Mosharaf Kabr Chowdhury Cherton School of Computer Scence Unversty of Waterloo Waterloo, Canada Emal: nmmkchow@uwaterloo.ca Muntasr Rahan

More information