Dynamic Processor Scheduling with Client Resources for Fast Multi-resolution WWW Image Browsing

Size: px
Start display at page:

Download "Dynamic Processor Scheduling with Client Resources for Fast Multi-resolution WWW Image Browsing"

Transcription

1 Dynamic Pocesso Scheduling with Resouces fo Fast Multi-esolution WWW Image Bowsing Daniel Andesen, Tao Yang, David Watson, and Athanassios Poulakidas Depatment of Compute Science Univesity of Califonia Santa Babaa, CA fdandese, tyang, david, Abstact WWW-based Intenet infomation sevice has gown enomously duing the last few yeas, and majo pefomance bottlenecks have been caused by WWW seve and Intenet bandwidth inadequacies. Utilizing client-site computing powe and also multi-pocesso suppot at the seve site can substantially impove the system esponse time. In this pape, we examine the use of scheduling techniques in monitoing and adapting to wokload vaiation at client and seve sites fo suppoting fast WWW image bowsing. We povide both analytic and expeimental esults to identify the impact of system loads and netwok bandwidth on esponse times and demonstate the effectiveness of ou scheduling stategy. 1 Intoduction One of the fundamental oles fo Wold-Wide Web (WWW) bowses and seves in today' s envionment is to povide a unifom inteface fo the on-line access of thousands of digitized documents, such as images. The main pefomance bottlenecks ae seve computing capability and Intenet bandwidth. We examine these two issues unde the scope of WWW-based digital libay (DL) systems [12]. While multi-pocesso suppot fo a seve is citical fo a popula WWW site [3], tansfeing pat of the seve's wokload to the client is also possible since the cuent Web bowses have achieved the ability to download executable content. Taking advantage of multipocesso suppot with client esouces can lead to significantly impoved use intefaces and esponse times. An application of the seve-client load shifting is in digital image bowsing. The cuent collections of the Alexandia Digital Libay (ADL) poject at UCSB [1] involve geogaphicallyefeenced mateials, such as maps, satellite images, digitized aeial photogaphs, and associated metadata. The image size is lage, fom MB. In such a case, it is obviously impactical to send all the images matching a quey in thei entiety. The seve must send lowesolution images fo the use to futhe cull befoe highesolution final vesions ae deliveed. A technique fo educing the netwok bandwidth equiements in viewing high-esolution images is to let the client fist bowse lowesolution thumbnails, then econstuct high-esolution images fom the existing thumbnails with a small amount of additional data deliveed fom the seve. This technique is called pogessive image delivey and equies cetain computing esouces at client sites. The main eseach challenge is the effective management and utilization of esouces fom multi-pocesso WWW seves and client-site machines. Blindly tansfeing load onto clients may not be advisable, since the bytecode pefomance of Java is usually 5-10 times slowe than a client machine' s potential. Also a numbe of commecial copoations ae developing netwok computes, with little o no had dive, a minimal pocesso, but with Java and the Intenet netwoking built in. A caeful design of the scheduling stategy is needed to avoid imposing too much buden on these Net PCs. In this pape we pesent a scheduling model fo patitioning and mapping client-seve computation based on dynamically changing seve load and client-site compute capabilities. Ou model incopoates multiple factos including disk, netwok, and computational abilities. We show how pope scheduling can lead to significantly impoved esponse times ove most cuent implementations in suppoting WWW image bowsing. The pape is oganized as follows: Section 2 gives the backgound on pogessive image bowsing. Section 3 pesents the patitioning of wavelet-based image accessing opeations. Section 4 discusses the scheduling stategy fo a multi-pocesso WWW seve and the cost modeling fo a wavelet task. Section 5 gives an analysis of the esponse time pefomance fo pocessing a fixed numbe of equests. Section 6 pesents the expeimental esults. Section 7 discusses elated wok and conclusions. 1

2 2 Multi-esolution image bowsing We focus on bowsing lage digitized data objects, e.g. images in the ADL system. With cuent netwok speeds, it is quite infeasible to conside sending the full contents of an image file to uses fo the bowsing puposes. The ADL has adopted pogessive multi-esolution image delivey and subegion bowsing as stategies to educe Intenet taffic when accessing map images [1]. This appoach is based on the idea that uses often bowse lage images via a thumbnail (coase esolution), and desie to apidly view highe-esolution vesions and subegions of those images aleady being viewed. We biefly descibe the techniques of wavelet image data etieval and tansfomation fo multi-esolution bowsing. Given an image, a fowad wavelet tansfom poduces a sub-sampled image of lowe esolution called a thumbnail, and thee additional coefficient data sets. Moe fomally, fo the given quantized image I 1 of esolution R R 1, we specify the input and output of the fowad wavelet tansfom as follows. (I 2 ; C 1 ; C 2 ; C 3 ) = F owad W avelet(i 1 ) I 2 is the thumbnail of esolution R 2 R, C 2 1; C 2 and C 3 ae of esolution R 2 R. Figue 1 depicts the esult of wavelet 2 tansfom. be econstucted at the client site. The image econstuction is not time consuming, taking about 1.5 seconds fo a image on a SUN SPARC 5. The size of compessed data C 1 ; C 2 ; C 3 to be tansfeed is geneally in the ange of 10 to 100KBytes, which takes less than 1 second ove a T1 link. If a use wishes to access subegions of an image I 1, then the coesponding subegions in thumbnail I 2 ; C 1 ; C 2 ; C 3 can be etieved and the econstuction pefomed accodingly. We model such a pocess as follows. subegion(i1) = Inv W(subegion(I2); subegion(c1); subegion(c2); subegion(c3)): A detailed definition of fowad and invese wavelet functions can be found in [8]. The time complexity of wavelet tansfoms is popotional to the image size. The wavelet tansfom can be applied ecusively, namely the thumbnail I 2 can be decomposed futhe to poduce smalle thumbnails. The actual wavelet implementation fo data access and image econstuction is a combination of Java and Computational Gateway Inteface (CGI) pogams [13]. Namely, an HTTP equest activates a CGI pogam at the seve, which computes the seve' s data using the pedefined pocedue. The esults ae sent to the client and then pocessed by the Java client-side application. Oiginal view subimage selection wavelet decomposition 3 Task patitioning fo pogessive subegion image delivey Reconstuction Seve side seve disk image data extact subegion eceate coefficients Thumbnail econs. pictue D 1 D 2 D 3 D 4 memoy view pictue side highe-esolution subimage Figue 1: Reconstucting a high-esolution subegion fom the thumbnail and coefficients. The invese wavelet tansfom (Inv W ) can be pefomed to e-constuct the oiginal image on-the-fly fom the coefficient data sets and the thumbnail. I 1 = Inv W (I 2 ; C 1 ; C 2 ; C 3 ): If image thumbnail I 2 is available at the client site, then by equesting that the seve sends C 1 ; C 2 ; C 3, image I 1 can 1 Rectangula shapes can also be suppoted while squae images ae used hee fo demonstation. Figue 2: A task chain fo wavelet image enhancement. Fou cutoff points ae depicted fo possible computation patitioning between client and seve. If the client pefoms the image econstuction, the thumbnail is aleady available fom the client memoy and does need to be tansmitted fom the seve. The computation involved in multi-esolution image constuction can be patially executed at a seve and at a client also. Based on an implementation in [15], we model the computation and communication involved using a chain of subtasks depicted in Figue 2: 1) Fetching compessed wavelet data and extacting the subegion. The wavelet image data is stoed in a combined quadtee/huffman encoded fom on a disk. These compessed files must be fetched. Then the appopiate subtee of a quadtee with its associated compessed coefficient data must be extacted in its compessed fom. The compessed

3 coefficient data is sent on to the next stage. 2) Receating the coefficients. The compessed coefficients must be expanded to thei oiginal fom. 3) Reconstucting the pixels. Afte the coefficients ae available, the invese wavelet function is called to ceate the new highe-esolution image fom the thumbnail image. Notice that the thumbnail image needs to be fetched fom the seve disk if the econstuction is conducted on the seve. Othewise, the thumbnail image is aleady available on the memoy of the client machine. 4) Viewing the image. Fo ou puposes, we assume the viewing of the image takes no computation time, and must be done on the client. Figue 2 depicts the above pocessing steps and fou possible cutoff points fo patitioning this chain fo the seve and client. We discuss the possible computation and communication scenaios fo fou patitioning points below. Notice that we also need to conside that the data sent fom the seve to the client may be compessed fist fo tansmission, then decompessed at the client site. D1: The client stats fom subegion extaction. The entie compessed image data needs to be tansfeed, but the image thumbnail does not need to be tansmitted. The tansmitted wavelet data is not futhe compessible. D2: The client stats fom coefficient data eceation. A pat of compessed image data is etieved on the seve based on the subegion position. The image thumbnail does not need to be tansmitted. The tansmitted subimage data is not futhe compessible. D3: The client stats with image econstuction. Coefficient econstuction is conducted at the seve site. But the deived subegion coefficient data must be futhe compessed othewise the size of uncompessed coefficient data is simila to that of the oiginal subegion image and it would be moe efficient to send the oiginal image. Thus the ovehead of seve compession and client decompession must be incopoated. The image thumbnail does not need to be tansmitted. D4: The client does not do any computation. The image thumbnail needs to be etieved fom the seve disk. The esult of image econstuction is not compessible futhe. 4 Pocesso scheduling fo WWW equests and cost modeling fo wavelet tasks Ou WWW seve consists of a set of wokstation nodes connected with a fast local netwok as shown in Figue 3 and is pesented as a single logical seve to the Intenet. The use equests ae fist evenly outed to pocessos via DNS otation [13]. The seve nodes in the system communicate with each othe and ediect equests to the pope node by actively monitoing the usages of CPU, I/O channels, and the inteconnection netwok. Ou scheduling stategy fo ediecting a equest to an appopiate seve node is to find a pocesso that minimizes the oveall esponse time. In [3], we have poposed Intenet HTTP equests DNS Disks WWW seve Fast netwok Figue 3: The achitectue of a multi-pocesso WWW seve. a pediction model fo appoximating the pocessing time of a given equest on a pocesso. In this model, we conside seveal factos that affect the esponse time. These include loads on CPU, disk, and netwok esouces. The load of a pocessing unit must be monitoed so that equests can be distibuted to elatively lightly loaded pocessos. Disk channel usage must also be obseved and simultaneous use equests accessing diffeent disks can utilize paallel I/O to achieve highe thoughput. The local inteconnection netwok bandwidth affects the pefomance of file etieval since many files may not eside on the local disk of a pocesso, so emote file etieval though the netwok file system will be involved. Local netwok taffic congestion could damatically slow the equest pocessing. Fo a wavelet task chain, we need to not only select an appopiate seve node fo pocessing, but also patition this chain into two pats. One pat is executed on an appopiate seve node, anothe pat is executed on the client. The optimum patitioning point is the one which minimizes the oveall pocessing time, which equies the dynamic consideation of client/seve machine loads and capabilities, and the netwok bandwidth between client and seve. We assume that the local communication delay between subtasks within the same machine (client o seve) is zeo while client-seve communication delay is detemined by the cuent available bandwidth and latency between them. The cost fo pocessing a wavelet chain is modeled as: t s = t ediection + t data + t seve + t net + t client : t ediection is the cost to ediect the equest to anothe pocesso, if equied. t seve is the time fo seve computation equied. t client is the time fo any client computation equied. The values of t seve and t client depend on how a task chain is patitioned. t data is the seve time to tansfe the equied data fom the seve disk dive, o fom the emote disk if the file is not local. t net is the cost fo tansfeing the pocessing esults ove the Intenet. We discuss t data and t net in moe detail. ( tlstatup + data size b t data = disk 1 If local, t statup + data size min(b disk 1;b lnet 2) If emote. The equested file is the compessed image data. If image econstuction is conducted at the seve then the thumbnail image should also be included. If econstuction is

4 conducted at the client then the thumbnail image is aleady available and thus it is not included. If the file is local, the time equied to fetch the data is simply the file size divided by the available bandwidth of the local stoage system, b disk, plus some statup ovehead t lstatup. We also measue the disk channel load 1. If thee ae many concuent equests, the disk tansmission pefomance degades accodingly. If the data is emote, then the file must be etieved though the inteconnection netwok. The local netwok bandwidth, b lnet, and load 2 must be incopoated, plus the statup ovehead t stat. Expeimentally, we found on the Meiko appoximately a 10% penalty fo a emote NFS access, and on the SUN wokstations connected by Spac/Ethenet the cost inceases by 50%-70%. In ou cuent implementation, t lstatup and t statup ae neglected because if the files ae lage, these costs ae elatively small and if the files ae small, ediection t ediection and netwok ovehead will dominate. t net = t nstat + client-seve communication size : Net bandwidth This tem is used to estimate the time necessay to etun the esults back to the client ove the netwok. The numbe of bytes equied again depends on how the patitioning is conducted. If the seve does image econstuction, then the entie subegion image needs to be shipped. If the client only does image econstuction, the seve only needs to send compessed the coefficient data. t nstat is the statup time fo netwok connection and is ignoed in the cuent setting fo simila easons to those given fo t lstat and t stat. Given this cost pediction fo an image bowsing equest, the system compaes all seve nodes and enumeate all possible patitioning choices fo the coesponding chain, then selects a patitioning and a seve node to each the minimum esponse time. 5 An analysis fo homogeneous client-seve systems It is difficult to analyze the pefomance of this scheme fo geneal cases. We make a numbe of assumptions to examine the impact of system esouces on the selection of patitioning points. While ou scheme woks fo pocessing a sequence of equests, we study the esponse time in simultaneously pocessing a fixed numbe of equests. Ou esult eflects the scheduling pefomance fo esponding to a bust of equests, which occus fequently in many WWW sites [5, 9]. In the following analysis, we assume that the system is homogeneous in the sense that all nodes have the same CPU speed and the initial load, and each node has a local disk with the same bandwidth. We assume that all clients ae unifomly loaded with the same machine capabilities. We assume that all equests have the same type of image opeations fo econstucting an n n subimage. Each seve node pocesso eceives a unifom numbe of equests, and poduces a stable thoughput of infomation equested. We define the following tems: R equests eceived fo the entie system. p the numbe of seve nodes. equests eceived pe pocesso. = R. Notice that p we assume that the equests aiving at each node afte the division via DNS ae unifom. A the aveage ovehead in pepocessing a equest, and deciding a ediection. the aveage bandwidth of local disk access. the aveage bandwidth of emote disk access. c the aveage pobability of a pocessed equest accessing a local disk. S the aveage slowdown atio of the client CPU compaed to the seve node. S = CP U seve speed CP U client speed. d the aveage ediection pobability. O the aveage ovehead of ediection. B the aveage netwok bandwidth available between the seve cluste and the client. H the aveage pocessing time fo each equest. F 1 the aveage size of the compessed wavelet data (quadtee and coefficients). k the aveage faction of F 1 actually needed fo a subegion. g the constant atio fo the seve cost of sending disk files, e.g., gf 1 is the time fo sending data F 1. F 2 the size of the thumbnail. The image size is n=2n=2. E 1 the aveage seve CPU time fo extacting the subegion infomation. E 2 the aveage seve CPU time fo ceating the subegion coefficient data. E 3 the aveage seve CPU time fo image econstuction. E c the aveage seve CPU time fo coefficient data compession. E d the aveage seve CPU time fo coefficient data decompession. Among the equests aiving at each node, we assume the pobability of accessing one of the seve disks is equal to 1/p. Then 1=p equests ae accessing the local disk. Among those equests, d of them will be ediected to othe nodes but d equests will be ediected fom othe nodes to this node (we also assume that ediection is unifomly distibuted because of the homogeneous system). Ou expeiments show that in such cases, the ediected equests tend to follow file locality. Thus the total numbe of equests pocessed at each pocesso afte ediection is equests pe second. Among them, the total numbe of equests accessing the local disk is =p fom the oiginal aival tasks, plus an additional d ediected equests. Then the pobability of accessing a local disk fo those equests is: p c + d = = 1 p + d: Let H be the aveage esponse time fo each equest (the time fom when the client launches a equest to the time the client eceives desied data). Then H = t sys + t data + W 1 + W 2 + W 3 + W 4 + W 5 + t net :

5 whee t sys is the ovehead fo possible ediection, HTTP connection and pasing. t data is seve time spent fo eading the compessed data and thumbnail file if needed. W 1 is time spent fo extacting a subegion. W 2 is time spent fo eceating the wavelet coefficient data. W 3 is fo the wavelet image econstuction. W 4 is the ovehead fo compessing/uncompessing data tansmitted between client and seve. W 5 is the seve time equied to send data to the client. t net is the netwok time fo client-seve data tansmission. Fo the case of accessing local disk, the disk bandwidth is shaed by equests with pobability of c. Fo the case of accessing a emote disk, the netwok bandwidth is shaed by equests with pobability of (1? c). Thus we have t data = c F + (1? c) F whee F = F 1 + F 2 if the seve does the econstuction, o F = F 1 othewise. The computation cost is the non-ovelapped CPU cycles fo pocessing a equest. Notice that CPU cycles ae shaed by equests. t sys = (A + d(a + O)): And, E1 if seve extacts a subegion, W 1 = S E 1 if client does it. E2 if seve eceates coefficient data, W 2 = S E 2 if client does it. E3 if seve does image econstuction, W 3 = S E 3 if client does it. Ec + E W d S if cutoff point is D3, 4 = 0 else. W 5 = g F n, and t net = F n =B whee F n = 8 >< >: F 1 if cutoff point is D1, k F 1 if cutoff point is D2, k F 1 if cutoff point is D3, n 2 if cutoff point is D4. Thee ae 4 possible patitions and we mak the esponse time fo these patitions as H 1 ; H 2 ; H 3 and H 4. We choose the one with the minimum pocessing time. Patition D1: H 1 = Patition D2: H 2 = Patition D3: H 3 = c F1 + (1? c) F1 + (A + d(a + O)) + gf 1 + S E 1 + S E 2 + S E 3 + F 1 =B: c F1 + (1? c) F1 + (A + d(a + O)) + (E 1 + gkf 1 ) + S (E 2 + E 3 ) + k F 1 =B: c F1 + (1? c) F1 + (E 1 + E 2 + E c + gkf 1 ) +S (E 3 + E d ) + k F 1 =B: + (A + d(a + O)) Patition D4: H 4 = c F1+F2 + (1? c) F1+F2 + (A + d(a + O)) + (E 1 + E 2 + E 3 + gn 2 ) + n 2 =B: We can futhe detemine the ediection atio d fo diffeent patitions in ode to minimize the esponse time. A detailed analysis can be found in [2]. Due to space estictions, the esults ae summaized as follows: Case a: When A + O (F 1 + F 2 )( 1? 1 ), d = 0 fo all H 1 ; H 2 ; H 3 and H 4. Case b: When A + O < F 1 ( 1? 1 ), d = 1? 1=p fo all H 1 ; H 2 ; H 3 and H 4. Case c: When F 1 ( 1 b? 1 2 b ) 1 A+O < (F 1 +F 2 )( 1 b? 2 1 ), d = 0 fo H 1 ; H 2 and H 3, and 1? 1=p fo H 4. Then, H = min(h 1 ; H 2 ; H 3 ; H 4 ): The above fomula can help us to undestand the patitioning selection. Fo example in Figue 4, we illustate the modeled esponse times and thei elationship to seve load, client capabilities, and netwok bandwidth using the following paametes based on ou expeimental esults: n=512, k = 0.25, R = 6, p = 6, E 1 = 1.4 sec, E 5 = 0.4 sec, E 6 = 2.6 sec, = byte pe second, = bytes pe second, A = sec, O = 0.1 sec, g = 2:5 10?7 sec/byte, F 1 = bytes, F 2 = (n=2) 2, E c = 0.9 sec, E d = 0.9 sec. Figue 4 (a) plots the esults fo H 1 ; H 2 ; H 3 ; and H 4, when S is set to ange fom 0.5 to 3. Fo (b), B anges fom to bytes/sec while S is 2. Fom (a) and (b), we can see that if the seve is vey fast o the Intenet communication is vey fast, then D4 is the best patition and the seve does eveything. Othewise, the best patition is at D2, and it is advisable to send ove the compessed data fo the client to pocess. 6 Expeimental Results We have implemented a pototype of ou dynamic scheduling scheme with client esouces on a Meiko CS- 2 distibuted memoy paallel machine based on ou pevious SWEB wok [3]. The Meiko CS-2 is essentially a wokstation cluste with a fast netwok inteconnect. Each node has a scala pocessing unit (a 40Mhz SupeSpac chip) with 32MB of RAM unning SUN Solais 2.3. Ou pimay expeimental testbed consists of six Meiko CS-2 nodes as ou seve, each of which is connected to a dedicated 1GB had dive on which the test files eside. Disk sevice is available to all othe nodes via NSF mounts. The client machines ae loaded with ou custom Java applet libay implementing some of the basic opeations, including wavelet econstuction. To avoid Intenet bandwidth fluctuations, clients ae located within the campus netwok

6 16 512x512 Zoom x512 Zoom * Fo H1 x Fo H o Fo H3 + Fo H4 Seconds 10 8 Seconds Fo H4 o Fo H3 x Fo H2 * Fo H (a) S: CPU Ratio of Seve to (b) Netwok Bandwidth x 10 4 Figue 4: Impact of system esouces on patitioning. (a) powe vaying, (b) Netwok bandwidth vaying. to assist in expeimental stability ove multiple uns. We pimaily examine the pefomance of wavelet subimage etieval. We select an opeation which extacts a subegion at full esolution fom a 2K 2K map image, epesenting the use zooming in on a point of inteest at a highe esolution afte examining at an image thumbnail. The ovehead of ou scheduling and load monitoing is quite small fo all expeiments. Analyzing a equest takes about the ange of 2-4ms, and load monitoing takes about 0.1% of CPU esouces. These esults ae vey simila to those epoted in [3]. Aveage Response time (seconds) Seves with Resouces 6 nodes 5 nodes 4 nodes 3 nodes 2 nodes 1 node Requests pe second Figue 5: Request esponse times with client esouces as the numbe of seves and RPS change. Pefomance impovement using seve and client esouces. We examine the pefomance of the multipocesso seve with client esouces. We un a test fo a peiod of 30 seconds, at each second R equests ae launched fom clients. Figue 5 shows the aveage esponse times fo 1, 2, 3, 4, 5 and 6 seve nodes with client computing esouces. We can see that the esponse times dop significantly by using multiple seve nodes. Fo example, with RPS=1, the aveage esponse time fo the 6- node seves with client esouces is less than 5 seconds while the time fo the 1-node seve is about 25 seconds. With and without client esouces. We also compae the impovement atio of the esponse time with client esouces ove the esponse time without client esouces in Table 1. As the seve load inceases steadily, the esponse time impovement atio inceases damatically. We also note a significant incease in the numbe of equests pe second (RPS) a seve system can complete by using client esouces. The detailed esults ae in [2]. RPS= p=6 142 % 164% 192% 679% 4029% p=5 96% 183% 191% 1838% 7893% Table 1: Response time impovement atio with and without client esouce. Load balancing with hotspots. Hotspots ae a typical poblem with DNS otation, whee a single seve node exhibits a highe load than its pees [13, 14]. We examine how ou system deals with hotspots. We diected a fixed numbe of equests to a subset of nodes in ou seve cluste, giving a wide ange of load dispaities. Without ou schedule, those nodes would have to pocess all of those equests. Ou scheduling algoithm can effectively deal with tempoay hotspots at vaious nodes by ediecting equests to othe nodes in the goup. The esult is shown in Figue 6. The X axis shows the ange of pocessos which eceive equests. The uppe cuve of Figue 6 shows the aveage esponse time in seconds when no scheduling is pefomed and the equest pocessing is limited to the fixed subset of nodes. The lowe cuve shows the aveage esponse time with pesence of ou schedule. In that case the load is evenly distibuted among all pocessos by ediection, and esponse times ae vey unifom. Dynamic scheduling with vaying client capabilities. Vaious clients will have diffeing capabilities in tems of netwok bandwidth and pocessing capabilities. We illustate the effect of these on the scheduling decisions of ou scheme and examine if the theoetical esults pesented in Section 5 match the system decisions unde the specified assumptions. Figue 7 shows how the system makes the decision in tems of cutoff points fo the subimage extac-

7 Aveage Response time (seconds) With Scheduling Without Scheduling Numbe of Nodes Figue 6: System pefomance with equest concentation at fixed seve subsets. Tests fo a peiod of 30 sec., 4 RPS. tion when pocessing R concuent equests whee R=18, and when atificially adjusting the seve/client bandwidth and CPU atio epoted by the client. Each coodinate enty in Figue 7 is maked with the decision of schedule. We ovelaid the theoetical pedictions fom the fomula of Section 5 on the system esults. Fo each enty, if the choices fo all equests agee with the theoetical pediction, we mak the actual selected cutoff decision, othewise we mak the pecentage of disageement. Fo example, In Figue 7, when the bandwidth is 100,000 bytes/sec. and S = 4, the pecentage of disageement with the theoetical model is 6% among all pocessed equests. We can see the theoetical model closely matches the system' s selections. As bandwidth deceases, the schedule would incease the pecentage of equests who make use of client CPU fo data decompession and image econstuction, minimizing the data sent ove the netwok. On the othe hand, as client CPU speeds decease, we expect the seve to do moe pocessing. -seve bandwidth Seve/ CPU Speed Ratio D2 D2 D2 D2 D2 D2 D2 D2 500 D2 D2 D2 D2 D2 D2 D2 D2 1,000 D2 D2 D2 D2 D2 D2 D2 D2 10,000 D2 D2 D2 D2 D2 D2 D4 D4 100,000 D2 D2 D D4 D4 D4 150,000 D2 D2 D2 77 D4 D4 D4 D4 250,000 D2 D2 D2 93 D4 D4 D4 D4 500,000 D2 D2 D2 95 D4 D4 D4 D4 750,000 D2 D2 D2 88 D4 D4 D4 D4 1,000,000 D2 D2 D2 98 D4 D4 D4 D4 bytes/sec Pedicted D2 egion Pedicted D4 egion Figue 7: Effects of CPU speed and netwok bandwidth on decisions. 7 Related wok and concluding emaks We have pesented a dynamic scheduling and cost model fo pocessing image bowsing equests by utilizing both client and WWW seve esouces. We have demonstated the effectiveness of ou scheme in adapting to diffeent client-seve capabilities fo minimizing esponse times. Shifting computation fom a seve to its clients essentially scattes the wokload aound the wold. This elates the global computing and application-level scheduling pojects [6, 7, 10]. Those pojects deal with an integation of diffeent machines as one vitual machine. Ou expeience in using bandwidth and load infomation fo scheduling could be useful to this eseach. Addessing client configuation vaiation is discussed in [11] fo filteing multi-media data in ode to educe netwok bandwidth equiements but does not conside the use of client esouces fo integated computing. Ou cuent wok is to genealize this wok to suppot othe applications on the WWW fo adaptive seve/client scheduling [2]. Acknowledgments This wok was suppoted in pat by funding fom NSF IRI , NSF CCR , NSF CDA and a gant fom NRaD. We would like to thank Ome Egecioglu, Osca Ibaa, Tey Smith, Cong Fu, Nobet Stubel, and the ADL image pocessing team fo many valuable discussions and suggestions. Refeences [1] D.Andesen, L.Cave, R.Dolin, C.Fische, J.Few, M.Goodchild, O.Ibaa, R.Kothui, M.Lasgaad, B.Manjunath, D.Nebet, J.Simpson, T.Smith, T.Yang, Q.Zheng, The WWW Pototype of the Alexandia Digital Libay, Poceedings of ISDL'95: Intenational Symposium on Digital Libaies, Japan August 22-25, [2] D.Andesen, T.Yang, ' Adaptive Scheduling with Resouces to Impove WWW Seve Scalability', Dept. of Compute Science Tech Rpt. TRCS96-27, U.C. Santa Babaa, [3] D.Andesen, T.Yang, V.Holmedahl, O.Ibaa, SWEB: Towads a Scalable Wold Wide Web Seve on Multicomputes, Poc. of 10th IEEE Intenational Symp. on Paallel Pocessing (IPPS'96), Apil, 1996, Hawaii. pp [4] D. Andesen, T. Yang, O. Egecioglu, O.H. Ibaa, T.R. Smith, Scalability Issues fo High Pefomance Digital Libaies on the Wold Wide Web, Poc. of the 3d Foum on Reseach and Tech. Advances in Digital Libaies (ADL96), pp , May, [5] M. Alitt, C. Williamson, Web Seve Wokload Chaacteization: The Seach fo Invaiants, Poc. SIGMETRICS Confeence, Philadelphia, PA, May, [6] F. Beman, R. Wolski, S. Figueia, J. Schopf, G. Shao, Application-Level Scheduling on Distibuted Heteogeneous Netwoks, Poc. of Supecomputing ' 96, [7] H. Casanova, J. Dongaa, NetSolve: A netwok seve fo solving computation science poblems, Poc. of Supecomputing' 96, ACM/IEEE, Nov., [8] E.C.K. Chui, Wavelets: A Tutoial in Theoy and Applications, Academic Pess, [9] M. Covella, A. Bestavos, Self-Similaity in Wold Wide Web Taffic Evidence and Possible Causes, Poc. SIGMETRICS96, Philadelphia, May, [10] K. Dince, and G. C. Fox, Building a wold-wide vitual machine based on Web and HPCC technologies. Poc. of Supecomputing'96, ACM/IEEE, Novembe, [11] A. Fox, E. Bewe, Reducing WWW Latency and Bandwidth Requiements by Real-Time Distillation, Compute Netwoks and ISDN Systems, Volume 28, issues 711, p May, [12] E. Fox, Akscyn, R., Fuuta, R. and Leggett, J. (Eds), Special issue on digital libaies, CACM, Apil [13] E.D. Katz, M. Butle, R. McGath, A Scalable HTTP Seve: the NCSA Pototype, Compute Netwoks and ISDN Systems. vol. 27, 1994, pp [14] D. Mosedale, W. Foss, R. McCool, Administeing Vey High Volume Intenet Sevices, 1995 LISA IX, Monteey, CA, Septembe, [15] A. Poulakidas, A. Sinivasan, O. Egecioglu, O. Ibaa, and T. Yang, Expeimental Studies on a Compact Stoage Scheme fo Wavelet-based Multiesolution Subegion Retieval, Poceedings of NASA 1996 Combined Industy, Space and Eath Science Data Compession Wokshop, Utah, Apil, 1996.

IP Network Design by Modified Branch Exchange Method

IP Network Design by Modified Branch Exchange Method Received: June 7, 207 98 IP Netwok Design by Modified Banch Method Kaiat Jaoenat Natchamol Sichumoenattana 2* Faculty of Engineeing at Kamphaeng Saen, Kasetsat Univesity, Thailand 2 Faculty of Management

More information

Module 6 STILL IMAGE COMPRESSION STANDARDS

Module 6 STILL IMAGE COMPRESSION STANDARDS Module 6 STILL IMAE COMPRESSION STANDARDS Lesson 17 JPE-2000 Achitectue and Featues Instuctional Objectives At the end of this lesson, the students should be able to: 1. State the shotcomings of JPE standad.

More information

Hierarchically Clustered P2P Streaming System

Hierarchically Clustered P2P Streaming System Hieachically Clusteed P2P Steaming System Chao Liang, Yang Guo, and Yong Liu Polytechnic Univesity Thomson Lab Booklyn, NY 11201 Pinceton, NJ 08540 Abstact Pee-to-pee video steaming has been gaining populaity.

More information

An Unsupervised Segmentation Framework For Texture Image Queries

An Unsupervised Segmentation Framework For Texture Image Queries An Unsupevised Segmentation Famewok Fo Textue Image Queies Shu-Ching Chen Distibuted Multimedia Infomation System Laboatoy School of Compute Science Floida Intenational Univesity Miami, FL 33199, USA chens@cs.fiu.edu

More information

ANALYTIC PERFORMANCE MODELS FOR SINGLE CLASS AND MULTIPLE CLASS MULTITHREADED SOFTWARE SERVERS

ANALYTIC PERFORMANCE MODELS FOR SINGLE CLASS AND MULTIPLE CLASS MULTITHREADED SOFTWARE SERVERS ANALYTIC PERFORMANCE MODELS FOR SINGLE CLASS AND MULTIPLE CLASS MULTITHREADED SOFTWARE SERVERS Daniel A Menascé Mohamed N Bennani Dept of Compute Science Oacle, Inc Geoge Mason Univesity 1211 SW Fifth

More information

Communication vs Distributed Computation: an alternative trade-off curve

Communication vs Distributed Computation: an alternative trade-off curve Communication vs Distibuted Computation: an altenative tade-off cuve Yahya H. Ezzeldin, Mohammed amoose, Chistina Fagouli Univesity of Califonia, Los Angeles, CA 90095, USA, Email: {yahya.ezzeldin, mkamoose,

More information

Illumination methods for optical wear detection

Illumination methods for optical wear detection Illumination methods fo optical wea detection 1 J. Zhang, 2 P.P.L.Regtien 1 VIMEC Applied Vision Technology, Coy 43, 5653 LC Eindhoven, The Nethelands Email: jianbo.zhang@gmail.com 2 Faculty Electical

More information

A Memory Efficient Array Architecture for Real-Time Motion Estimation

A Memory Efficient Array Architecture for Real-Time Motion Estimation A Memoy Efficient Aay Achitectue fo Real-Time Motion Estimation Vasily G. Moshnyaga and Keikichi Tamau Depatment of Electonics & Communication, Kyoto Univesity Sakyo-ku, Yoshida-Honmachi, Kyoto 66-1, JAPAN

More information

On using circuit-switched networks for file transfers

On using circuit-switched networks for file transfers On using cicuit-switched netwoks fo file tansfes Xiuduan Fang, Malathi Veeaaghavan Univesity of Viginia Email: {xf4c, mv5g}@viginia.edu Abstact High-speed optical cicuit-switched netwoks ae being deployed

More information

Dynamic Topology Control to Reduce Interference in MANETs

Dynamic Topology Control to Reduce Interference in MANETs Dynamic Topology Contol to Reduce Intefeence in MANETs Hwee Xian TAN 1,2 and Winston K. G. SEAH 2,1 {stuhxt, winston}@i2.a-sta.edu.sg 1 Depatment of Compute Science, School of Computing, National Univesity

More information

Journal of World s Electrical Engineering and Technology J. World. Elect. Eng. Tech. 1(1): 12-16, 2012

Journal of World s Electrical Engineering and Technology J. World. Elect. Eng. Tech. 1(1): 12-16, 2012 2011, Scienceline Publication www.science-line.com Jounal of Wold s Electical Engineeing and Technology J. Wold. Elect. Eng. Tech. 1(1): 12-16, 2012 JWEET An Efficient Algoithm fo Lip Segmentation in Colo

More information

Segmentation of Casting Defects in X-Ray Images Based on Fractal Dimension

Segmentation of Casting Defects in X-Ray Images Based on Fractal Dimension 17th Wold Confeence on Nondestuctive Testing, 25-28 Oct 2008, Shanghai, China Segmentation of Casting Defects in X-Ray Images Based on Factal Dimension Jue WANG 1, Xiaoqin HOU 2, Yufang CAI 3 ICT Reseach

More information

Detection and Recognition of Alert Traffic Signs

Detection and Recognition of Alert Traffic Signs Detection and Recognition of Alet Taffic Signs Chia-Hsiung Chen, Macus Chen, and Tianshi Gao 1 Stanfod Univesity Stanfod, CA 9305 {echchen, macuscc, tianshig}@stanfod.edu Abstact Taffic signs povide dives

More information

Lecture # 04. Image Enhancement in Spatial Domain

Lecture # 04. Image Enhancement in Spatial Domain Digital Image Pocessing CP-7008 Lectue # 04 Image Enhancement in Spatial Domain Fall 2011 2 domains Spatial Domain : (image plane) Techniques ae based on diect manipulation of pixels in an image Fequency

More information

The Dual Round Robin Matching Switch with Exhaustive Service

The Dual Round Robin Matching Switch with Exhaustive Service The Dual Round Robin Matching Switch with Exhaustive Sevice Yihan Li, Shivenda S. Panwa, H. Jonathan Chao Abstact Vitual Output Queuing is widely used by fixed-length highspeed switches to ovecome head-of-line

More information

Optical Flow for Large Motion Using Gradient Technique

Optical Flow for Large Motion Using Gradient Technique SERBIAN JOURNAL OF ELECTRICAL ENGINEERING Vol. 3, No. 1, June 2006, 103-113 Optical Flow fo Lage Motion Using Gadient Technique Md. Moshaof Hossain Sake 1, Kamal Bechkoum 2, K.K. Islam 1 Abstact: In this

More information

Scaling Location-based Services with Dynamically Composed Location Index

Scaling Location-based Services with Dynamically Composed Location Index Scaling Location-based Sevices with Dynamically Composed Location Index Bhuvan Bamba, Sangeetha Seshadi and Ling Liu Distibuted Data Intensive Systems Laboatoy (DiSL) College of Computing, Geogia Institute

More information

Adaptation of TDMA Parameters Based on Network Conditions

Adaptation of TDMA Parameters Based on Network Conditions Adaptation of TDMA Paametes Based on Netwok Conditions Boa Kaaoglu Dept. of Elect. and Compute Eng. Univesity of Rocheste Rocheste, NY 14627 Email: kaaoglu@ece.ocheste.edu Tolga Numanoglu Dept. of Elect.

More information

On the Conversion between Binary Code and Binary-Reflected Gray Code on Boolean Cubes

On the Conversion between Binary Code and Binary-Reflected Gray Code on Boolean Cubes On the Convesion between Binay Code and BinayReflected Gay Code on Boolean Cubes The Havad community has made this aticle openly available. Please shae how this access benefits you. You stoy mattes Citation

More information

Analysis of Wired Short Cuts in Wireless Sensor Networks

Analysis of Wired Short Cuts in Wireless Sensor Networks Analysis of Wied Shot Cuts in Wieless Senso Netwos ohan Chitaduga Depatment of Electical Engineeing, Univesity of Southen Califonia, Los Angeles 90089, USA Email: chitadu@usc.edu Ahmed Helmy Depatment

More information

Any modern computer system will incorporate (at least) two levels of storage:

Any modern computer system will incorporate (at least) two levels of storage: 1 Any moden compute system will incopoate (at least) two levels of stoage: pimay stoage: andom access memoy (RAM) typical capacity 32MB to 1GB cost pe MB $3. typical access time 5ns to 6ns bust tansfe

More information

Configuring RSVP-ATM QoS Interworking

Configuring RSVP-ATM QoS Interworking Configuing RSVP-ATM QoS Intewoking Last Updated: Januay 15, 2013 This chapte descibes the tasks fo configuing the RSVP-ATM QoS Intewoking featue, which povides suppot fo Contolled Load Sevice using RSVP

More information

A Two-stage and Parameter-free Binarization Method for Degraded Document Images

A Two-stage and Parameter-free Binarization Method for Degraded Document Images A Two-stage and Paamete-fee Binaization Method fo Degaded Document Images Yung-Hsiang Chiu 1, Kuo-Liang Chung 1, Yong-Huai Huang 2, Wei-Ning Yang 3, Chi-Huang Liao 4 1 Depatment of Compute Science and

More information

Controlled Information Maximization for SOM Knowledge Induced Learning

Controlled Information Maximization for SOM Knowledge Induced Learning 3 Int'l Conf. Atificial Intelligence ICAI'5 Contolled Infomation Maximization fo SOM Knowledge Induced Leaning Ryotao Kamimua IT Education Cente and Gaduate School of Science and Technology, Tokai Univeisity

More information

A New Finite Word-length Optimization Method Design for LDPC Decoder

A New Finite Word-length Optimization Method Design for LDPC Decoder A New Finite Wod-length Optimization Method Design fo LDPC Decode Jinlei Chen, Yan Zhang and Xu Wang Key Laboatoy of Netwok Oiented Intelligent Computation Shenzhen Gaduate School, Habin Institute of Technology

More information

Performance Optimization in Structured Wireless Sensor Networks

Performance Optimization in Structured Wireless Sensor Networks 5 The Intenational Aab Jounal of Infomation Technology, Vol. 6, o. 5, ovembe 9 Pefomance Optimization in Stuctued Wieless Senso etwoks Amine Moussa and Hoda Maalouf Compute Science Depatment, ote Dame

More information

An Improved Resource Reservation Protocol

An Improved Resource Reservation Protocol Jounal of Compute Science 3 (8: 658-665, 2007 SSN 549-3636 2007 Science Publications An mpoved Resouce Resevation Potocol Desie Oulai, Steven Chambeland and Samuel Piee Depatment of Compute Engineeing

More information

Slotted Random Access Protocol with Dynamic Transmission Probability Control in CDMA System

Slotted Random Access Protocol with Dynamic Transmission Probability Control in CDMA System Slotted Random Access Potocol with Dynamic Tansmission Pobability Contol in CDMA System Intaek Lim 1 1 Depatment of Embedded Softwae, Busan Univesity of Foeign Studies, itlim@bufs.ac.k Abstact In packet

More information

THE THETA BLOCKCHAIN

THE THETA BLOCKCHAIN THE THETA BLOCKCHAIN Theta is a decentalized video steaming netwok, poweed by a new blockchain and token. By Theta Labs, Inc. Last Updated: Nov 21, 2017 esion 1.0 1 OUTLINE Motivation Reputation Dependent

More information

a Not yet implemented in current version SPARK: Research Kit Pointer Analysis Parameters Soot Pointer analysis. Objectives

a Not yet implemented in current version SPARK: Research Kit Pointer Analysis Parameters Soot Pointer analysis. Objectives SPARK: Soot Reseach Kit Ondřej Lhoták Objectives Spak is a modula toolkit fo flow-insensitive may points-to analyses fo Java, which enables expeimentation with: vaious paametes of pointe analyses which

More information

High performance CUDA based CNN image processor

High performance CUDA based CNN image processor High pefomance UDA based NN image pocesso GEORGE VALENTIN STOIA, RADU DOGARU, ELENA RISTINA STOIA Depatment of Applied Electonics and Infomation Engineeing Univesity Politehnica of Buchaest -3, Iuliu Maniu

More information

A Recommender System for Online Personalization in the WUM Applications

A Recommender System for Online Personalization in the WUM Applications A Recommende System fo Online Pesonalization in the WUM Applications Mehdad Jalali 1, Nowati Mustapha 2, Ali Mamat 2, Md. Nasi B Sulaiman 2 Abstact foeseeing of use futue movements and intentions based

More information

Accurate Diffraction Efficiency Control for Multiplexed Volume Holographic Gratings. Xuliang Han, Gicherl Kim, and Ray T. Chen

Accurate Diffraction Efficiency Control for Multiplexed Volume Holographic Gratings. Xuliang Han, Gicherl Kim, and Ray T. Chen Accuate Diffaction Efficiency Contol fo Multiplexed Volume Hologaphic Gatings Xuliang Han, Gichel Kim, and Ray T. Chen Micoelectonic Reseach Cente Depatment of Electical and Compute Engineeing Univesity

More information

IP Multicast Simulation in OPNET

IP Multicast Simulation in OPNET IP Multicast Simulation in OPNET Xin Wang, Chien-Ming Yu, Henning Schulzinne Paul A. Stipe Columbia Univesity Reutes Depatment of Compute Science 88 Pakway Dive South New Yok, New Yok Hauppuage, New Yok

More information

Information Retrieval. CS630 Representing and Accessing Digital Information. IR Basics. User Task. Basic IR Processes

Information Retrieval. CS630 Representing and Accessing Digital Information. IR Basics. User Task. Basic IR Processes CS630 Repesenting and Accessing Digital Infomation Infomation Retieval: Basics Thosten Joachims Conell Univesity Infomation Retieval Basics Retieval Models Indexing and Pepocessing Data Stuctues ~ 4 lectues

More information

Point-Biserial Correlation Analysis of Fuzzy Attributes

Point-Biserial Correlation Analysis of Fuzzy Attributes Appl Math Inf Sci 6 No S pp 439S-444S (0 Applied Mathematics & Infomation Sciences An Intenational Jounal @ 0 NSP Natual Sciences Publishing o Point-iseial oelation Analysis of Fuzzy Attibutes Hao-En hueh

More information

(1) W tcp = (3) N. Assuming 1 P r 1. = W r (4) a 1/(k+1) W 2/(k+1)

(1) W tcp = (3) N. Assuming 1 P r 1. = W r (4) a 1/(k+1) W 2/(k+1) 1 Multi Path PERT Ankit Singh and A. L. Naasimha Reddy Electical and Compute Engineeing Depatment, Texas A&M Univesity; email: eddy@ece.tamu.edu. Abstact This pape pesents a new multipath delay based algoithm,

More information

Image Enhancement in the Spatial Domain. Spatial Domain

Image Enhancement in the Spatial Domain. Spatial Domain 8-- Spatial Domain Image Enhancement in the Spatial Domain What is spatial domain The space whee all pixels fom an image In spatial domain we can epesent an image by f( whee x and y ae coodinates along

More information

Towards Adaptive Information Merging Using Selected XML Fragments

Towards Adaptive Information Merging Using Selected XML Fragments Towads Adaptive Infomation Meging Using Selected XML Fagments Ho-Lam Lau and Wilfed Ng Depatment of Compute Science and Engineeing, The Hong Kong Univesity of Science and Technology, Hong Kong {lauhl,

More information

RANDOM IRREGULAR BLOCK-HIERARCHICAL NETWORKS: ALGORITHMS FOR COMPUTATION OF MAIN PROPERTIES

RANDOM IRREGULAR BLOCK-HIERARCHICAL NETWORKS: ALGORITHMS FOR COMPUTATION OF MAIN PROPERTIES RANDOM IRREGULAR BLOCK-HIERARCHICAL NETWORKS: ALGORITHMS FOR COMPUTATION OF MAIN PROPERTIES Svetlana Avetisyan Mikayel Samvelyan* Matun Kaapetyan Yeevan State Univesity Abstact In this pape, the class

More information

WIRELESS sensor networks (WSNs), which are capable

WIRELESS sensor networks (WSNs), which are capable IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. XX, NO. XX, XXX 214 1 Lifetime and Enegy Hole Evolution Analysis in Data-Gatheing Wieless Senso Netwoks Ju Ren, Student Membe, IEEE, Yaoxue Zhang, Kuan

More information

EE 6900: Interconnection Networks for HPC Systems Fall 2016

EE 6900: Interconnection Networks for HPC Systems Fall 2016 EE 6900: Inteconnection Netwoks fo HPC Systems Fall 2016 Avinash Kaanth Kodi School of Electical Engineeing and Compute Science Ohio Univesity Athens, OH 45701 Email: kodi@ohio.edu 1 Acknowledgement: Inteconnection

More information

Prioritized Traffic Recovery over GMPLS Networks

Prioritized Traffic Recovery over GMPLS Networks Pioitized Taffic Recovey ove GMPLS Netwoks 2005 IEEE. Pesonal use of this mateial is pemitted. Pemission fom IEEE mu be obtained fo all othe uses in any cuent o futue media including epinting/epublishing

More information

On the Forwarding Area of Contention-Based Geographic Forwarding for Ad Hoc and Sensor Networks

On the Forwarding Area of Contention-Based Geographic Forwarding for Ad Hoc and Sensor Networks On the Fowading Aea of Contention-Based Geogaphic Fowading fo Ad Hoc and Senso Netwoks Dazhi Chen Depatment of EECS Syacuse Univesity Syacuse, NY dchen@sy.edu Jing Deng Depatment of CS Univesity of New

More information

An Extension to the Local Binary Patterns for Image Retrieval

An Extension to the Local Binary Patterns for Image Retrieval , pp.81-85 http://x.oi.og/10.14257/astl.2014.45.16 An Extension to the Local Binay Pattens fo Image Retieval Zhize Wu, Yu Xia, Shouhong Wan School of Compute Science an Technology, Univesity of Science

More information

Lifetime and Energy Hole Evolution Analysis in Data-Gathering Wireless Sensor Networks

Lifetime and Energy Hole Evolution Analysis in Data-Gathering Wireless Sensor Networks 788 IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, VOL. 12, NO. 2, APRIL 2016 Lifetime and Enegy Hole Evolution Analysis in Data-Gatheing Wieless Senso Netwoks Ju Ren, Student Membe, IEEE, Yaoxue Zhang,

More information

Using SPEC SFS with the SNIA Emerald Program for EPA Energy Star Data Center Storage Program Vernon Miller IBM Nick Principe Dell EMC

Using SPEC SFS with the SNIA Emerald Program for EPA Energy Star Data Center Storage Program Vernon Miller IBM Nick Principe Dell EMC Using SPEC SFS with the SNIA Emeald Pogam fo EPA Enegy Sta Data Cente Stoage Pogam Venon Mille IBM Nick Pincipe Dell EMC v6 Agenda Backgound on SNIA Emeald/Enegy Sta fo block Intoduce NAS/File test addition;

More information

A modal estimation based multitype sensor placement method

A modal estimation based multitype sensor placement method A modal estimation based multitype senso placement method *Xue-Yang Pei 1), Ting-Hua Yi 2) and Hong-Nan Li 3) 1),)2),3) School of Civil Engineeing, Dalian Univesity of Technology, Dalian 116023, China;

More information

Topological Characteristic of Wireless Network

Topological Characteristic of Wireless Network Topological Chaacteistic of Wieless Netwok Its Application to Node Placement Algoithm Husnu Sane Naman 1 Outline Backgound Motivation Papes and Contibutions Fist Pape Second Pape Thid Pape Futue Woks Refeences

More information

Multidimensional Testing

Multidimensional Testing Multidimensional Testing QA appoach fo Stoage netwoking Yohay Lasi Visuality Systems 1 Intoduction Who I am Yohay Lasi, QA Manage at Visuality Systems Visuality Systems the leading commecial povide of

More information

Exploring non-typical memcache architectures for decreased latency and distributed network usage.

Exploring non-typical memcache architectures for decreased latency and distributed network usage. Syacuse Univesity SURFACE Electical Engineeing and Compute Science Technical Repots College of Engineeing and Compute Science 9-5-20 Exploing non-typical memcache achitectues fo deceased latency and distibuted

More information

Positioning of a robot based on binocular vision for hand / foot fusion Long Han

Positioning of a robot based on binocular vision for hand / foot fusion Long Han 2nd Intenational Confeence on Advances in Mechanical Engineeing and Industial Infomatics (AMEII 26) Positioning of a obot based on binocula vision fo hand / foot fusion Long Han Compute Science and Technology,

More information

SYSTEM LEVEL REUSE METRICS FOR OBJECT ORIENTED SOFTWARE : AN ALTERNATIVE APPROACH

SYSTEM LEVEL REUSE METRICS FOR OBJECT ORIENTED SOFTWARE : AN ALTERNATIVE APPROACH I J C A 7(), 202 pp. 49-53 SYSTEM LEVEL REUSE METRICS FOR OBJECT ORIENTED SOFTWARE : AN ALTERNATIVE APPROACH Sushil Goel and 2 Rajesh Vema Associate Pofesso, Depatment of Compute Science, Dyal Singh College,

More information

Combinatorial Mobile IP: A New Efficient Mobility Management Using Minimized Paging and Local Registration in Mobile IP Environments

Combinatorial Mobile IP: A New Efficient Mobility Management Using Minimized Paging and Local Registration in Mobile IP Environments Wieless Netwoks 0, 3 32, 200 200 Kluwe Academic Publishes. Manufactued in The Nethelands. Combinatoial Mobile IP: A New Efficient Mobility Management Using Minimized Paging and Local Registation in Mobile

More information

Color Correction Using 3D Multiview Geometry

Color Correction Using 3D Multiview Geometry Colo Coection Using 3D Multiview Geomety Dong-Won Shin and Yo-Sung Ho Gwangju Institute of Science and Technology (GIST) 13 Cheomdan-gwagio, Buk-ku, Gwangju 500-71, Republic of Koea ABSTRACT Recently,

More information

Transmission Lines Modeling Based on Vector Fitting Algorithm and RLC Active/Passive Filter Design

Transmission Lines Modeling Based on Vector Fitting Algorithm and RLC Active/Passive Filter Design Tansmission Lines Modeling Based on Vecto Fitting Algoithm and RLC Active/Passive Filte Design Ahmed Qasim Tuki a,*, Nashien Fazilah Mailah b, Mohammad Lutfi Othman c, Ahmad H. Saby d Cente fo Advanced

More information

Input Layer f = 2 f = 0 f = f = 3 1,16 1,1 1,2 1,3 2, ,2 3,3 3,16. f = 1. f = Output Layer

Input Layer f = 2 f = 0 f = f = 3 1,16 1,1 1,2 1,3 2, ,2 3,3 3,16. f = 1. f = Output Layer Using the Gow-And-Pune Netwok to Solve Poblems of Lage Dimensionality B.J. Biedis and T.D. Gedeon School of Compute Science & Engineeing The Univesity of New South Wales Sydney NSW 2052 AUSTRALIA bbiedis@cse.unsw.edu.au

More information

View Synthesis using Depth Map for 3D Video

View Synthesis using Depth Map for 3D Video View Synthesis using Depth Map fo 3D Video Cheon Lee and Yo-Sung Ho Gwangju Institute of Science and Technology (GIST) 1 Oyong-dong, Buk-gu, Gwangju, 500-712, Republic of Koea E-mail: {leecheon, hoyo}@gist.ac.k

More information

arxiv: v2 [physics.soc-ph] 30 Nov 2016

arxiv: v2 [physics.soc-ph] 30 Nov 2016 Tanspotation dynamics on coupled netwoks with limited bandwidth Ming Li 1,*, Mao-Bin Hu 1, and Bing-Hong Wang 2, axiv:1607.05382v2 [physics.soc-ph] 30 Nov 2016 1 School of Engineeing Science, Univesity

More information

(a, b) x y r. For this problem, is a point in the - coordinate plane and is a positive number.

(a, b) x y r. For this problem, is a point in the - coordinate plane and is a positive number. Illustative G-C Simila cicles Alignments to Content Standads: G-C.A. Task (a, b) x y Fo this poblem, is a point in the - coodinate plane and is a positive numbe. a. Using a tanslation and a dilation, show

More information

Modeling a shared medium access node with QoS distinction

Modeling a shared medium access node with QoS distinction Modeling a shaed medium access node with QoS distinction Matthias Gies, Jonas Geutet Compute Engineeing and Netwoks Laboatoy (TIK) Swiss Fedeal Institute of Technology Züich CH-8092 Züich, Switzeland email:

More information

Frequency Domain Approach for Face Recognition Using Optical Vanderlugt Filters

Frequency Domain Approach for Face Recognition Using Optical Vanderlugt Filters Optics and Photonics Jounal, 016, 6, 94-100 Published Online August 016 in SciRes. http://www.scip.og/jounal/opj http://dx.doi.og/10.436/opj.016.68b016 Fequency Domain Appoach fo Face Recognition Using

More information

HISTOGRAMS are an important statistic reflecting the

HISTOGRAMS are an important statistic reflecting the JOURNAL OF L A T E X CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 1 D 2 HistoSketch: Disciminative and Dynamic Similaity-Peseving Sketching of Steaming Histogams Dingqi Yang, Bin Li, Laua Rettig, and Philippe

More information

Conversion Functions for Symmetric Key Ciphers

Conversion Functions for Symmetric Key Ciphers Jounal of Infomation Assuance and Secuity 2 (2006) 41 50 Convesion Functions fo Symmetic Key Ciphes Deba L. Cook and Angelos D. Keomytis Depatment of Compute Science Columbia Univesity, mail code 0401

More information

RT-WLAN: A Soft Real-Time Extension to the ORiNOCO Linux Device Driver

RT-WLAN: A Soft Real-Time Extension to the ORiNOCO Linux Device Driver 1 RT-WLAN: A Soft Real-Time Extension to the ORiNOCO Linux Device Dive Amit Jain Daji Qiao Kang G. Shin The Univesity of Michigan Ann Abo, MI 4819, USA {amitj,dqiao,kgshin@eecs.umich.edu Abstact The cuent

More information

POMDP: Introduction to Partially Observable Markov Decision Processes Hossein Kamalzadeh, Michael Hahsler

POMDP: Introduction to Partially Observable Markov Decision Processes Hossein Kamalzadeh, Michael Hahsler POMDP: Intoduction to Patially Obsevable Makov Decision Pocesses Hossein Kamalzadeh, Michael Hahsle 2019-01-02 The R package pomdp povides an inteface to pomdp-solve, a solve (witten in C) fo Patially

More information

Cloud-Aided Wireless Networks with Edge Caching: Fundamental Latency Trade-Offs in Fog Radio Access Networks

Cloud-Aided Wireless Networks with Edge Caching: Fundamental Latency Trade-Offs in Fog Radio Access Networks 206 IEEE Intenational Symposium on Infomation Theoy Cloud-Aided Wieless Netwos with Edge Caching: Fundamental Latency Tade-Offs in Fog Radio Access Netwos Ravi Tandon Osvaldo Simeone Depatment of ECE Univesity

More information

Assessment of Track Sequence Optimization based on Recorded Field Operations

Assessment of Track Sequence Optimization based on Recorded Field Operations Assessment of Tack Sequence Optimization based on Recoded Field Opeations Matin A. F. Jensen 1,2,*, Claus G. Søensen 1, Dionysis Bochtis 1 1 Aahus Univesity, Faculty of Science and Technology, Depatment

More information

Topic -3 Image Enhancement

Topic -3 Image Enhancement Topic -3 Image Enhancement (Pat 1) DIP: Details Digital Image Pocessing Digital Image Chaacteistics Spatial Spectal Gay-level Histogam DFT DCT Pe-Pocessing Enhancement Restoation Point Pocessing Masking

More information

UCB CS61C : Machine Structures

UCB CS61C : Machine Structures inst.eecs.bekeley.edu/~cs61c UCB CS61C : Machine Stuctues Lectue SOE Dan Gacia Lectue 28 CPU Design : Pipelining to Impove Pefomance 2010-04-05 Stanfod Reseaches have invented a monitoing technique called

More information

Event-based Location Dependent Data Services in Mobile WSNs

Event-based Location Dependent Data Services in Mobile WSNs Event-based Location Dependent Data Sevices in Mobile WSNs Liang Hong 1, Yafeng Wu, Sang H. Son, Yansheng Lu 3 1 College of Compute Science and Technology, Wuhan Univesity, China Depatment of Compute Science,

More information

Spiral Recognition Methodology and Its Application for Recognition of Chinese Bank Checks

Spiral Recognition Methodology and Its Application for Recognition of Chinese Bank Checks Spial Recognition Methodology and Its Application fo Recognition of Chinese Bank Checks Hanshen Tang 1, Emmanuel Augustin 2, Ching Y. Suen 1, Olivie Baet 2, Mohamed Cheiet 3 1 Cente fo Patten Recognition

More information

An Efficient Handover Mechanism Using the General Switch Management Protocol on a Multi-Protocol Label Switching Network

An Efficient Handover Mechanism Using the General Switch Management Protocol on a Multi-Protocol Label Switching Network An Efficient andove Mechanism Using the Geneal Switch Management Potocol on a Multi-Potocol abel Switching Netwok Seong Gon hoi, yun Joo Kang, and Jun Kyun hoi Using the geneal switch management potocol

More information

INFORMATION DISSEMINATION DELAY IN VEHICLE-TO-VEHICLE COMMUNICATION NETWORKS IN A TRAFFIC STREAM

INFORMATION DISSEMINATION DELAY IN VEHICLE-TO-VEHICLE COMMUNICATION NETWORKS IN A TRAFFIC STREAM INFORMATION DISSEMINATION DELAY IN VEHICLE-TO-VEHICLE COMMUNICATION NETWORKS IN A TRAFFIC STREAM LiLi Du Depatment of Civil, Achitectual, and Envionmental Engineeing Illinois Institute of Technology 3300

More information

Quality Aware Privacy Protection for Location-based Services

Quality Aware Privacy Protection for Location-based Services In Poceedings of the th Intenational Confeence on Database Systems fo Advanced Applications (DASFAA 007), Bangkok, Thailand, Apil 9-, 007. Quality Awae Pivacy Potection fo Location-based Sevices Zhen Xiao,,

More information

Journal of Network and Computer Applications

Journal of Network and Computer Applications Jounal of Netwok and Compute Applications 34 (211) 135 142 Contents lists available at ScienceDiect Jounal of Netwok and Compute Applications jounal homepage: www.elsevie.com/locate/jnca Optimization of

More information

Dynamic Multiple Parity (DMP) Disk Array for Serial Transaction Processing

Dynamic Multiple Parity (DMP) Disk Array for Serial Transaction Processing IEEE TRANSACTIONS ON COMPUTERS, VOL. 50, NO. 9, SEPTEMBER 200 949 Dynamic Multiple Paity (DMP) Disk Aay fo Seial Tansaction Pocessing K.H. Yeung, Membe, IEEE, and T.S. Yum, Senio Membe, IEEE AbstactÐThe

More information

Modelling, simulation, and performance analysis of a CAN FD system with SAE benchmark based message set

Modelling, simulation, and performance analysis of a CAN FD system with SAE benchmark based message set Modelling, simulation, and pefomance analysis of a CAN FD system with SAE benchmak based message set Mahmut Tenuh, Panagiotis Oikonomidis, Peiklis Chachalakis, Elias Stipidis Mugla S. K. Univesity, TR;

More information

The Internet Ecosystem and Evolution

The Internet Ecosystem and Evolution The Intenet Ecosystem and Evolution Contents Netwok outing: basics distibuted/centalized, static/dynamic, linkstate/path-vecto inta-domain/inte-domain outing Mapping the sevice model to AS-AS paths valley-fee

More information

COSC 6385 Computer Architecture. - Pipelining

COSC 6385 Computer Architecture. - Pipelining COSC 6385 Compute Achitectue - Pipelining Sping 2012 Some of the slides ae based on a lectue by David Culle, Pipelining Pipelining is an implementation technique wheeby multiple instuctions ae ovelapped

More information

Assessment of query reweighing, by rocchio method in farsi information retrieval

Assessment of query reweighing, by rocchio method in farsi information retrieval Univesity of Wollongong Reseach Online Univesity of Wollongong in Dubai - Papes Univesity of Wollongong in Dubai 2008 Assessment of quey eweighing, by occhio metho in fasi infomation etieval F Sabooi Bu-Ali

More information

SCALABLE ENERGY EFFICIENT AD-HOC ON DEMAND DISTANCE VECTOR (SEE-AODV) ROUTING PROTOCOL IN WIRELESS MESH NETWORKS

SCALABLE ENERGY EFFICIENT AD-HOC ON DEMAND DISTANCE VECTOR (SEE-AODV) ROUTING PROTOCOL IN WIRELESS MESH NETWORKS SCALABL NRGY FFICINT AD-HOC ON DMAND DISTANC VCTOR (S-AODV) ROUTING PROTOCOL IN WIRLSS MSH NTWORKS Sikande Singh Reseach Schola, Depatment of Compute Science & ngineeing, Punjab ngineeing College (PC),

More information

A VECTOR PERTURBATION APPROACH TO THE GENERALIZED AIRCRAFT SPARE PARTS GROUPING PROBLEM

A VECTOR PERTURBATION APPROACH TO THE GENERALIZED AIRCRAFT SPARE PARTS GROUPING PROBLEM Accepted fo publication Intenational Jounal of Flexible Automation and Integated Manufactuing. A VECTOR PERTURBATION APPROACH TO THE GENERALIZED AIRCRAFT SPARE PARTS GROUPING PROBLEM Nagiza F. Samatova,

More information

ADDING REALISM TO SOURCE CHARACTERIZATION USING A GENETIC ALGORITHM

ADDING REALISM TO SOURCE CHARACTERIZATION USING A GENETIC ALGORITHM ADDING REALISM TO SOURCE CHARACTERIZATION USING A GENETIC ALGORITHM Luna M. Rodiguez*, Sue Ellen Haupt, and Geoge S. Young Depatment of Meteoology and Applied Reseach Laboatoy The Pennsylvania State Univesity,

More information

DEADLOCK AVOIDANCE IN BATCH PROCESSES. M. Tittus K. Åkesson

DEADLOCK AVOIDANCE IN BATCH PROCESSES. M. Tittus K. Åkesson DEADLOCK AVOIDANCE IN BATCH PROCESSES M. Tittus K. Åkesson Univesity College Boås, Sweden, e-mail: Michael.Tittus@hb.se Chalmes Univesity of Technology, Gothenbug, Sweden, e-mail: ka@s2.chalmes.se Abstact:

More information

A New and Efficient 2D Collision Detection Method Based on Contact Theory Xiaolong CHENG, Jun XIAO a, Ying WANG, Qinghai MIAO, Jian XUE

A New and Efficient 2D Collision Detection Method Based on Contact Theory Xiaolong CHENG, Jun XIAO a, Ying WANG, Qinghai MIAO, Jian XUE 5th Intenational Confeence on Advanced Mateials and Compute Science (ICAMCS 2016) A New and Efficient 2D Collision Detection Method Based on Contact Theoy Xiaolong CHENG, Jun XIAO a, Ying WANG, Qinghai

More information

Goal. Rendering Complex Scenes on Mobile Terminals or on the web. Rendering on Mobile Terminals. Rendering on Mobile Terminals. Walking through images

Goal. Rendering Complex Scenes on Mobile Terminals or on the web. Rendering on Mobile Terminals. Rendering on Mobile Terminals. Walking through images Goal Walking though s -------------------------------------------- Kadi Bouatouch IRISA Univesité de Rennes I, Fance Rendeing Comple Scenes on Mobile Teminals o on the web Rendeing on Mobile Teminals Rendeing

More information

AN ANALYSIS OF COORDINATED AND NON-COORDINATED MEDIUM ACCESS CONTROL PROTOCOLS UNDER CHANNEL NOISE

AN ANALYSIS OF COORDINATED AND NON-COORDINATED MEDIUM ACCESS CONTROL PROTOCOLS UNDER CHANNEL NOISE AN ANALYSIS OF COORDINATED AND NON-COORDINATED MEDIUM ACCESS CONTROL PROTOCOLS UNDER CHANNEL NOISE Tolga Numanoglu, Bulent Tavli, and Wendi Heinzelman Depatment of Electical and Compute Engineeing Univesity

More information

Erasure-Coding Based Routing for Opportunistic Networks

Erasure-Coding Based Routing for Opportunistic Networks Easue-Coding Based Routing fo Oppotunistic Netwoks Yong Wang, Sushant Jain, Magaet Matonosi, Kevin Fall Pinceton Univesity, Univesity of Washington, Intel Reseach Bekeley ABSTRACT Routing in Delay Toleant

More information

Lecture #22 Pipelining II, Cache I

Lecture #22 Pipelining II, Cache I inst.eecs.bekeley.edu/~cs61c CS61C : Machine Stuctues Lectue #22 Pipelining II, Cache I Wiewold cicuits 2008-7-29 http://www.maa.og/editoial/mathgames/mathgames_05_24_04.html http://www.quinapalus.com/wi-index.html

More information

Introduction To Pipelining. Chapter Pipelining1 1

Introduction To Pipelining. Chapter Pipelining1 1 Intoduction To Pipelining Chapte 6.1 - Pipelining1 1 Mooe s Law Mooe s Law says that the numbe of pocessos on a chip doubles about evey 18 months. Given the data on the following two slides, is this tue?

More information

Effective Missing Data Prediction for Collaborative Filtering

Effective Missing Data Prediction for Collaborative Filtering Effective Missing Data Pediction fo Collaboative Filteing Hao Ma, Iwin King and Michael R. Lyu Dept. of Compute Science and Engineeing The Chinese Univesity of Hong Kong Shatin, N.T., Hong Kong { hma,

More information

Modeling spatially-correlated data of sensor networks with irregular topologies

Modeling spatially-correlated data of sensor networks with irregular topologies This full text pape was pee eviewed at the diection of IEEE Communications Society subject matte expets fo publication in the IEEE SECON 25 poceedings Modeling spatially-coelated data of senso netwoks

More information

A Minutiae-based Fingerprint Matching Algorithm Using Phase Correlation

A Minutiae-based Fingerprint Matching Algorithm Using Phase Correlation A Minutiae-based Fingepint Matching Algoithm Using Phase Coelation Autho Chen, Weiping, Gao, Yongsheng Published 2007 Confeence Title Digital Image Computing: Techniques and Applications DOI https://doi.og/10.1109/dicta.2007.4426801

More information

A Novel Automatic White Balance Method For Digital Still Cameras

A Novel Automatic White Balance Method For Digital Still Cameras A Novel Automatic White Balance Method Fo Digital Still Cameas Ching-Chih Weng 1, Home Chen 1,2, and Chiou-Shann Fuh 3 Depatment of Electical Engineeing, 2 3 Gaduate Institute of Communication Engineeing

More information

Optimal Adaptive Learning for Image Retrieval

Optimal Adaptive Learning for Image Retrieval Optimal Adaptive Leaning fo Image Retieval ao Wang Dept of Compute Sci and ech singhua Univesity Beijing 00084, P. R. China Wangtao7@63.net Yong Rui Micosoft Reseach One Micosoft Way Redmond, WA 9805,

More information

A ROI Focusing Mechanism for Digital Cameras

A ROI Focusing Mechanism for Digital Cameras A ROI Focusing Mechanism fo Digital Cameas Chu-Hui Lee, Meng-Feng Lin, Chun-Ming Huang, and Chun-Wei Hsu Abstact With the development and application of digital technologies, the digital camea is moe popula

More information

A Non-blocking Directory Protocol for Large-Scale Multiprocessors. Technical Report

A Non-blocking Directory Protocol for Large-Scale Multiprocessors. Technical Report A Non-blocking Diectoy Potocol fo Lage-Scale Multipocessos Technical Repot Depatment of Compute Science and Engineeing Univesity of Minnesota 4-192 EECS Building 200 Union Steet SE Minneapolis, MN 55455-0159

More information

An effective hop-by-hop Interest shaping mechanism for CCN communications

An effective hop-by-hop Interest shaping mechanism for CCN communications An effective hop-by-hop Inteest shaping mechanism fo CCN communications Natalya Rozhnova, Sege Fdida To cite this vesion: Natalya Rozhnova, Sege Fdida. An effective hop-by-hop Inteest shaping mechanism

More information

MANET QoS support without reservations

MANET QoS support without reservations SECURITY AND COMMUNICATION NETWORKS Secuity Comm. Netwoks (2010) Published online in Wiley InteScience (www.intescience.wiley.com)..183 SPECIAL ISSUE PAPER MANET QoS suppot without esevations Soon Y. Oh,

More information