RAID-RMS: A fault tolerant stripped mirroring RAID architecture for distributed systems

Size: px
Start display at page:

Download "RAID-RMS: A fault tolerant stripped mirroring RAID architecture for distributed systems"

Transcription

1 computers & security 28 (2009) available at joural homepage: -RMS: A fault tolerat stripped mirrorig for distributed systems Javad Akbari Torkestai a, *, Mohammad Reza Meybodi b,c a Computer Egieerig Departmet, Islamic Azad Uiversity, Arak, Ira b Computer Egieerig Departmet, Amirkabir Uiversity of Techology, Tehra, Ira c Istitute for Studies a Theoretical Physics ad Mathematics (IPM), School of Computer Sciece, Tehra, Ira article ifo Article history: Received 5 September 2006 Received i revised form 14 Jue 2008 Accepted 2 September 2008 Keywords: Disk mirrorig Reliability Data stripig abstract Disk arrays, or s, have become the solutio to icrease the capacity, badwidth ad reliability of most storage systems. I spite of its high redudacy level, disk mirrorig is a popular paradigm, because replicatig data also doubles the badwidth available for processig requests, improves the reliability ad achieves fault tolerace. I this paper, we preset a ew called -RMS i which a special hybrid mechaism is used to map the data blocks to the cluster. The mai idea behid the proposed algorithm is to combie the data block stripig ad disk mirrorig techique with a data block rotatio. The resultig improves the parallelism reliability ad efficiecy of the array. We show that the proposed is able to serve may more disk requests compared to the other mirrorig-based s. We also argue that a more balaced disk load is attaied by the give, especially whe there are some disk failures. ª 2008 Elsevier Ltd. All rights reserved. 1. Itroductio A Redudat Array of Idepedet Disks- system is composed of a disk array, ad a disk array cotroller. The cotroller s fuctio is to receive a operatio from the user of the disk array, ad to carry it out by performig a set of lowlevel operatios o specific disks. There are several s that are classified ito five levels (Patterso et al., 1988; Gibso et al., 1998). Differet s ca be distiguished based o their type of ecodig, mappig ad algorithms used to access data. I other words, each system ca be defied by these three compoets. The proposed i this paper carefully cosiders all the subjects above. The ecodig idicates the type of redudacy iformatio used to ecode the data. The mappig itroduces the patter used to place data ad redudat iformatio o the disk array. Algorithms used to access data ca be classified as ormal mode ad failed mode oes. I ormal mode, either there is o failure i the disk array or the cotroller kows about the failed disks, if ay. I failed mode, a disk failure has occurred i the middle of cotroller operatio. The cotroller the eeds to recover the error ad complete the operatio. This process is called error recovery. There have bee umerous reliability ad performace evaluatio studies of, sice the publicatio of Patterso et al. (1988). This paper focuses o the reliability ad badwidth, which are the mai reasos of formatio systems. Gibso et al. (1998), Thomasia ad Blaum (2006) ad Thomasia ad Blaum (2006), Thomasia (2006) preseted a reliability-based compariso amog some existig s, especially for those are based o mirrorig techique. There are several studies cosiderig ew disk mirrorig orgaizatios i which the icreasig disk loads due to the failures are balaced. For example, iterleaved * Correspodig author. Computer Egieerig Departmet, Islamic Azad Uiversity, Arak, Ira Tel.: þ ; fax: þ addresses: j-akbari@iau-arak.ac.ir (J. Akbari Torkestai), mmeybodi@aut.ac.ir (M. Reza Meybodi) /$ see frot matter ª 2008 Elsevier Ltd. All rights reserved. doi: /j.cose

2 computers & security 28 (2009) declusterig method-id (Tera data Corporatio, 1985), implemeted i the origial teradata database machie, partitios N disks ito c clusters ad the primary data o each disk is distributed evely amog the other disks i the cluster. Hsiao ad DeWitt (1993) itroduced a called chaied declusterig which is a improvemet over ID ad tolerates a larger umber of disk failures tha ID. I this, each disk is partitioed ito a primary ad secodary area, ad the primary data o each disk is replicated o the secodary area. Che ad Towsley (1996) proposed the group rotate declusterig techique i which similar to the basic mirrorig, disks are divided ito primary ad secodary disks. Data is striped ad the stripe uits of each primary disk are located i a rotatio maer o secodary disks. The s, based o their redudat data, provide differet reliability levels ad effective badwidth to access data. I spite of the high redudacy level i -1, due to the basic mirrorig techique, the caot balace the load icreased by disk failures ad so it reduces the reliability ad access badwidth for processig requests. To overcome this shortcomig, we propose a ew i which the data block stripig ad disk mirrorig techiques are combied. The former techique is used to balace disk loads ad improves the badwidth to access data ad the latter techique replicates the data blocks ad icreases the reliability of the array. The theoretical aalysis shows that the proposed outperforms the existig mirrorig-based s i ivestigated metrics ad the experimetal results completely cofirm it. It is show that the proposed is able to serve may more disk requests compared to the other systems whose s are based o disk mirrorig scheme. We also argue that a more balaced disk loads is attaied by the give model, especially whe there are some disk failures. The rest of the paper is orgaized as follows. I the ext sectio, we propose a ew, which is explaied i full detail i Sectio 3. The proposed is evaluated through theoretical aalysis i terms of the maximum badwidth, disk operatio cost ad reliability i Sectio 4, ad the compared to other similar s through simulatio i Sectio 5. The cocludig remarks are preseted i Sectio The proposed I this paper, we adopt the followig otatioal covetios to itroduce the proposed. Let N ad deote the block sequece umber i the memory address space ad the umber of disks i the cluster, respectively. M deotes the umber of blocks per disk, each of size S ad i ad j are idices to specify the horizotal ad vertical positios of a data or parity blocks, respectively. Formally, a redudat array of idepedet disks- ca be described by a quituple < ~ A; B ; R ; f B ; g R >, where ~ A ¼fa 0 ; a 1 ;.; a N g deotes the sequece of the data blocks, B¼fb ði;jþ j0 i m 1; 0 j 1g is the fiite set of data blocks i the disk array, R ¼ FWfr ði;jþ j0 i m 1; 0 j 1sg is the fiite set of redudat blocks which may be a ull set. It implies that a may provide o redudacy, like 0. Let x B ad x R be the cardiality of B ad R, respectively. The redudacy rate for a give ca be obtaied by x R /(x B þ x R ). f B :A / B is the block fuctio, which maps to each block sequece umber a N a two-dimessioal address b (i,j ) i the cluster. Let b deotes a give subset of B. g R :B / R is the redudat fuctio, which assigs a redudat block r (i,j ) to a give subset b. It gives rise to the s ca be easily studied. The basic idea behid the ew hybrid called -RMS 1 is to combie the data block stripig ad disk mirrorig techiques with the block rotatio to improve the access badwidth ad reliability of the array. To do so, the data blocks are divided ito two distict set of blocks: The primary data blocks ad the Mirrored (or Backup) blocks. Each disk is also partitioed ito a primary ad secodary area (Hsiao ad DeWitt, 1993) ad the former data blocks are distributed i the primary disk area ad the latter oes i the secodary. I this, Data is striped (Che et al., 1994) ad the primary data is distributed evely amog all the disks i the cluster, the the resultig primary data blocks o each disk are replicated amog the other ( 1) disks i a rotated maer fully described i the ext sectio. Now, if a sigle disk fails, a effectively balaced load ca be attaied at all survivig disks. So, this ca icrease the parallelism for processig requests, the reliability ad the effective badwidth to access data. It is show that the proposed is able to serve may more disk requests compared to the other systems whose s are based o disk mirrorig. Besides, such ca tolerate more disk failures ad decrease the rate of missig data due to the preseted data distributio. The disk failure may cause a egligible reductio to access data ad system performace compared to the other existig s (Akbari Torkestai ad Meybodi, 2005a; Akbari Torkestai ad Meybodi, 2005b; Akbari Torkestai et al., 2006; Akbari Torkestai ad Haghighat, 2003; Cortes, 1999; Stoebraker et al., 2000). The method proposed to distribute the data, forms the cluster so that each disk icludes both primary ad backup data copies. That is, the primary data blocks are located i the upper halves of the disks (i.e. primary area) ad their mirrored copies are located i the lower halves (i.e. primary area). The most importat issue which should be cosidered is desigig a efficiet data distributio algorithm for distributig the blocks i the maer metioed earlier. Such method descriptio is comprehesively preseted i the ext sectio. I this data distributio method the data blocks ad their backups are distributed by a right rotatio i the separated disks. Thus, less data will be lost after disk failures ad it ca be used as a data storig model for those systems i which data correctio, protectio ad security are really importat. Furthermore, i this model, while the system is i the recovery state to repair some failed disks or blocks, other blocks ad disks remai available ad this improves the system performace dramatically. 1 Rotatig-Mirrorig-Stripig (RMS).

3 42 computers & security 28 (2009) Descriptio of -RMS As metioed earlier, the proposed suggests a combiatio of the data block stripig ad disk mirrorig techiques with a data block rotatio to distribute the data i the cluster. I the remaiig of this sectio, the data distributio method proposed i this paper will be formulated ad described i detail. The formulas as show by (1) (3) provide two mappigs from the block sequece umbers to the primary data blocks i the cluster ad vice versa, as well as a mappig from the data blocks to their backup copies. I other words, the block fuctio f B :A / B ad redudat fuctio g R :B / R ca be realized by the metioed formulas. Sice the sequece block umbers i the memory space are addressed i a sigle-dimesio fashio, they must be trasformed to the two-dimesio addresses to be stored i the cluster. I our proposed array, each disk is partitioed ito a primary ad secodary area (Hsiao ad DeWitt, 1993). The data blocks are also divided ito two distict set of blocks: The primary data blocks ad the mirrored blocks. Let B (i,j ) s deote the primary data blocks which are located withi the upper halves of the disks, where 0 i 1ad 0 j m/2 1, adm (i 0,j 0 ) s deote the mirrored copies located i the lower halves, where m/2 i 0 m 1ad0j 0 1. Now, the data distributio method ca be described as follows: whe the cotroller receives a request to store a sigle or a set of data blocks, the followig steps should be take by the cotroller: Step 1. Data is striped (Che et al., 1994) ad the primary data is distributed evely amog all the disks i the cluster. The addresses i which the primary data blocks should be located are obtaied by (1). This formula defies the block fuctio f B :A / B for our method, where B deotes the primary data block. It s a mappig from the block sequece umber A N to the physical address (primary data block) of the cluster B (i,j ). i ¼ P N B ði;jþ ¼ R j ¼ P NR þðn mod Þ (1) mod Where 0 i m/2 1 ad 0 j 1. Let A N ¼ N for otatioal coveiece. Step 2. I this step, the primary data blocks o each disk are replicated i a rotated maer amog the other ( 1) disks modula- as stated by (2-a). It maps a mirrored block M (i 0,j 0 ) to each block sequece umber A N ad defies the redudat fuctio g R :B / R. ( i 0 ¼ P NR þ m=2 M ði0;j0þ ¼ h j 0 ¼ P ðnþ1þ i R þððnþ1þ mod Þ mod (2-a) Where m/2 i 0 m 1 ad 0 j 0 1. I the steps above, a right side rotatio techique is implicitly ivolved to fid the positio of data ad mirrored blocks withi disk array. For retrievig the missig blocks, the data blocks ad their backups eed to be immediately coverted, so (2-a) should be rewritte based o i ad j. Substitutig PN=R ¼ i ad j 0 ¼ j þ 1i (2-a) fially yields (2-b). i M ði0;j0þ ¼ 0 ¼ i þ m=2 j 0 ¼ðjþ1Þmod (2-b) Whe disk i fails, the correspodig backup copies of the missig primary data blocks, B (i,j ) s(for all j ¼ 0,1,.,m/2 1), ca be directly foud by (2-b) ad easily retrieved. The lost mirrored blocks M (i 0,j 0 ) (for all j ¼ m/2,.,m 1) ca be recostructed by iversely applyig (2-b). Besides, we represet a iverse-block fuctio i which the Cartesia coordiates B (i,j ) of the block-iterleaved data are mapped to the sequece block umber A N. ði Þþðj mod Þ i; A N ¼ ði þ 1Þþj i; i_j i j Where 0 i m/2 1ad 0 j Evaluatio ad compariso As metioed earlier, our proposed exploits the data stripig techique i order to balace the disk loads ad the disk mirrorig techique to retrieve the missig blocks. Therefore, it seems appropriate to implemet i those storage systems which hold huge amouts of critical data. I the remaiig part of this sectio, the preseted is theoretically aalyzed i terms of the maximum badwidth, disk operatio cost ad reliability ad its results are compared to those of the existig mirrorig-based s The maximum badwidth The s, based o their redudat data, provide differet reliability levels ad effective badwidth to access data. These are some criteria for evaluatig the effectiveess of the disk storage s. The access badwidth eeds to be cosidered whe the system is either i the ormal or failed modes to process the requests. Although the proposed, due to the block stripig techique, is able to access the total badwidth provided by all disks for processig both small ad large requests, the mirrorig-based s ad, cosequetly, the proposed replicate the data blocks o the secodary disks so the data blocks ad their backups would be ivolved for processig both small ad large requests. Therefore, all the s whose codig methods are based o the data mirrorig techique, ca efficietly exploit oly half of the total badwidth available to process large writigs. At first, we itroduce some cocepts upo which the ext discussios are based i the rest of the paper. Read/Write requests are those kids of disk operatios i which the whole or part of a data block is processed ad their scopes do ot exceed the give block at all. Therefore, such I/O operatios may ifluece the differet blocks of the same or differet stripes simultaeously (Corbett et al., 1993). Read/Write requests are those kids of disk operatios i which the several blocks of the same stripe or eve the differet stripes may be ivolved to process the requests. Note that, i most cases, the empirical results give i the literatures are ot as good as the theoretical results are. This is because the badwidth provided by the is cosidered i a maximum case ad so it is ot always reachable (Corme ad Kotz, 1993; Foster et al., 1997; Ho et al., 2000). (3)

4 computers & security 28 (2009) Let ad B be the total umber of disks i the arrays ad the maximum disk badwidth, respectively. As show i Table 1, the maximum badwidth provided by all give disk arrays to process the small requests is B. I spite of the high redudacy level i -1, due to the disk mirrorig techique, the effective badwidth provided by the model is limited to oly half of the badwidth available i the cluster. So whe a large is requested to process, the cotroller is able to process oly two cosecutive data blocks i parallel. The maximum badwidth provided by the other existig blockiterleaved s is B. I -5, to solve the bottleeck problem which is associated with modifyig the parity blocks i 4, the idividual parity disk is removed ad the parity blocks are distributed across the all disks. Sice the give parity blocks should be modified after processig the small or the large requests, the maximum badwidth available for processig small operatios is reduced to ( 1)/2 B ad for large operatios to ( 1) B Disk operatio cost Let N Totaldisk ad N Checkdisk be the total umber of disks ad the umber of check disks i the cluster, respectively. Let T R be the mea time to a give block ad L Data be the legth of a give request (i.e. the umber of blocks per file). As metioed before, i all s discussed i the previous sectios, idividual small requests ca be processed i parallel, where ¼ N Totaldisk. As show i Table 1, the average cost, due to processig the small operatios, usig such s is T R. We itroduce the followig formula to calculate the disk operatio cost eeded to process a request as log as L Data, where L Data is the umber of blocks eeded to be processed. This formula ca be applied for all s except -1. The resultig costs have bee show i Table 1. ðl Data T R Þ=ðN Totaldisk N Checkdisk Þ (4) The proposed ot oly does ot cause ay side effects o seek time ad disk cache hit, but it achieves better results tha the other disk arrays. I -5, the parity blocks, associated to the modified strips, should be updated after each small or large operatio. This process is described as follows: the modified data blocks ad the old parity block should be to obtai the ew parity block. The the ew oe is rewritte. Thus, as show i the last two colums of Table 1, each request has a additioal cost for ig the modified data ad the old parity blocks Reliability aalysis The more parallelism betwee requests ad the higher badwidth to access data, due to the striped data blocks, ad the higher reliability agaist disk failures due to the data mirrorig techique, are some of the advatages of the proposed. I this subsectio, the reliability of the proposed is evaluated i terms of the maximum disk failure coverage, which meas the maximum umber of disk failures that ca be tolerated by the disk array (Cortes, 1999; Aderso et al., 1996), ad the compared to those of the arrays metioed earlier. As metioed before, due to usig the data mirrorig techique i -1 ad -10 to backup the disk failures, the levels of the failure recovery provided by both give s are the same. Assume that two idepedet disks fail successively so that the secod failure occurs before the first oe ca be recovered completely. It implies that the rebuild process which is retrievig the first failed disk caot be completed before the secod disk failure. Sice the MTTR 2 is assumed to be much smaller tha the MTTF 3, the probability that the failed disk ca be retrieved ad the array retured to the origial state is close to uity. Therefore, the probability that each disk failure or missig data block ca be recovered by the give s is 1 (1/( 1)), where is the total umber of disks i the cluster. Despite their high redudacy level, because of the weak data distributio patter, these arrays caot tolerate the disks failures effectively. But, there is a special case, whe these s tolerate /2 successive idepedet disk failures (Corme ad Kotz, 1993; Foster et al., 1997; Harry et al., 1995; Howard et al., 1988; Hsiao ad DeWitt, 1990; Hu et al., 1999). -5 uses the distributed parity blocks to detect ad correct errors. I this, the cotroller checks the strips by xorig the data blocks ad parity block, ad recogizes the happeed failure, if ay. -5 ca oly tolerate the sigle disk failures ad detect the strips i which some failures happeed ad it is ot capable of exactly detectig the damaged data blocks. I the case that, the umber of failures happeed i a give strip is eve, the cotroller of such a array caot detect ay failures (Asami et al., 1999; Cabrera ad Log, 1991; Cao et al., 1994). As metioed earlier, -RMS uses a special hybrid mechaism to map the data blocks to the cluster. The mai idea behid the proposed algorithm is to combie the data block stripig ad disk mirrorig techiques with a data block rotatio to improve the parallelism, reliability ad efficiecy of the array. I this, the data is blockiterleaved ad the distributed i the other ( 1) disks modula-. The mirrored blocks are also replicated diagoally throughout the secodary area i a similar maer. Such a, whose data ad mirrored blocks are distributed diagoally, alleviates the effects of the both disk ad block failures. The results of the compariso betwee these s show that the proposed leads to cosiderably better results of both data block ad disk failure recoveries. Assume that two successive idepedet disk failures be happeed i the differet disks of the cluster so that the secod failure occurs before recoverig the first oe completely. I this case, the probability that the missig data blocks ca be recovered is 1 (1/( 1)( 2)). The system we cosidered i Sectio 4.1 to evaluate ad compare the give s, cocered the maximum badwidth provided by the models whe the system is i the ormal state. I this state, the cotroller does ot eed to perform ay special fuctios to support the system. I the remaiig part of this sectio, we cosider the maximum available badwidth of the s whe the system is i failed mode. I this 2 Mea time to repair. 3 Mea time to failure.

5 44 computers & security 28 (2009) Table 1 The theoretical results of the four s i terms of the maximum I/O badwidth, I/O rate ad recovery rate level Maximum badwidth -1 N B B/2 (/2) B (/2) B -10 B B (/2) B (/2) B -5 B ( 1) B ( 1)/2 B ( 1) B -RMS B B (/2) B (/2) B Read/ cost -1 R L/2 R W L W -1 0 R L/ R W L W/(/2) -5 R L R /( 1) R þ W L (R þ W )/( 1) -RMS R L/ R W L W/(/2) Recovery rate -1 ( 1) B B -10 ( 1) B (/2) B -5 ( 1) B ( 2) B -RMS ( 1) B ( 1) B state, the cotroller is recoverig the disk failures. First of all, the followig questios eed to be cosidered. 1 How much is the able to utilize the available badwidth after some disk failures? 2 How much is the system able to utilize the available badwidth of the cluster for retrievig the missig data? To aswer these questios, we describe the recovery process i detail. Whe a disk fails, rebuild processig is used to retrieve the cotets of the failed disk o a spare disk. Durig the recostructio process, the ormal fuctios of the system should be temporarily stopped. Therefore, to achieve higher performace, the recostructio iterval should be shorteed as much as possible. It improves the reliability of the array as the MTTR/MTTF ratio is reduced. Sice all the missig data blocks should be rewritte sequetially, the expected time to re a failed disk is the same for all s. So we cosider the expected time to do the ig process. I -1 ad -10, this process sequetially s the cotets of the mirrored disk ad the available badwidth to recostruct the failed disk is limited to oly the mirrored disk. I -5, this process ivolves the ig of successive blocks from the ( 1) survivig disks. Although these ig operatios ca be doe i parallel, but the metioed process should be repeated for all the missig blocks ad so here the expected time is also similar to those of the earlier s. The proposed ca ivolve the ( 1) survivig disks to the block-iterleaved backup copies simultaeously ad so provides the highest effective badwidth compared to the other s. As show i Table 1, after a sigle disk failure, the proposed model provides the most effective badwidth compared to the other s as we argued i Sectio 5.1. Furthermore, the proposed array ca tolerate all double disk failures whereas the other arrays will lose some data. I Table 1, let N total ¼, T R ¼ R, T W ¼ W ad L Data ¼ L for otatioal coveiece. 5. Experimetal results We coducted some simulatio experimets to study the performace of the proposed, ad compared the obtaied results to the other studied s i terms of the access badwidth, I/O rate ad reliability. We first experimeted the effective badwidth provided by the studied s to access data. The parallel s ad s are cosidered separately ad the large request size, for either or operatios, is set to 20 MB. The size of each block strip is set to 32 KB. This cofiguratio implies a high degree of stripig ad parallelism. The requests to access disks ad data are geerated with a uiform distributio. The measured access badwidth has bee show i Table 2 as a fuctio of the umber of cliet requests for parallel data accesses. The results summarized i all tables are averages over oe hudred rus. We cosidered both light ad heavy I/O traffic rates, ad varied the umber of cliet requests from 1 to 16 parallel cliet requests, while the umber of disks is fixed at 16. To reveal the parallel I/O capability of the disk arrays, the large ad the large are studied ad to test the idividual disk performace, the small or are cosidered. For a small, the access badwidth provided by -5 is very close to that of our proposed disk array ad -RMS performs oly slightly better tha -5, while it shows much better results tha -10. As the umber of cliet requests icreases, both -RMS ad -5 show higher scalability ad -10 lags behid. The simulatio results for large requests are very close to that for small requests. For parallel s of either a large file or a small block, the -RMS achieves the best scalability amog the other Table 2 Maximum badwidth compariso of -RMS, -10 ad -5 ad the effects of I/O rate (MB/S) of cliets RMS

6 computers & security 28 (2009) s with a highest 16 MB/s for 16 cliets. -5 scales slowly due to the bottleeck problem metioed earlier for modifyig the parity blocks. -10 scales slower tha our proposed, but very faster tha -5. To sum up, the -RMS, due to the full parallelism i accessig all data blocks i the upper half of the disk array, outperforms the others. Sice more disks i the cluster may satisfy more cliet requests, scalig of a disk array size may icrease the effective access badwidth. O the other had, it may also itroduce more cotetios. The effects of disk array size are very similar to the effects of the cliet request rate. I both cases, we cosider the saturated cases of accessig a large umber of cliet requests simultaeously. These experimets take ito accout the array size i the simulatios ad measure the access badwidth provided by the s metioed above, as the umber of disks icreases from 2 to 16 i array. The total umber of cliet requests is fixed at 16. The workload is fixed at 320 MB for large s or for large s, regardless of the array size. For small s or s, the fixed workload is set to 512 KB for 16 cliets. As show i Table 3, these experimetal results are slightly differet from those of the earlier oes, especially for operatios. For large or small s, the rakig give for previous experimets remais uchaged ad -RMS is raked above the other s ad -5 below. The differeces i access badwidth are attributed maily to the architectural characteristics of each disk array. The stregth of the -RMS lies maily i its superior performace i performig parallel operatios, regardless of the graularity. The I/O rates have bee show i Table 4 agaist the umber of disks i the cluster. I these experimets, the umber of cliet requests is fixed at 16, the umber of disks icreases from 2 to 16 ad each cliet requests to or 500 MB data o the array. The block results show that the -5 ad -RMS perform better tha -10. Especially, as the array size icreases to 16 disks, the proposed array ad -5 coverge to a output rate of aroud 4 MB/s. But the -10 speed lags far behid because it experieced more overheads. This overhead comes from a less aggressive mirrorig scheme o the -10. Table 4 also shows the results of the block experimets, where a 500 MB data is Table 3 Scalability of -RMS, -10 ad -5 for ad operatios uder a heavy I/O traffic coditio (MB/S) of cliets RMS Table 4 The I/O rates of the proposed, -10 ad -5 for ad operatios (MB/S) of cliets RMS writte to the disk array. The proposed ad -10 have the same output rate with a peak speed of 3.75 MB/s. This rate is almost idepedet of the disk array size. I -5, the overhead of excessive parity updates result i a lower output rate. I systems, the reliability of the s ca be also evaluated i terms of the maximum disk failure coverage, which meas the maximum umber of disk failures that ca be tolerated by the disk arrays. Amog the four s, -5 ca oly tolerate the sigle disk failures ad so it provides the lowest reliability. -1 ad -10, due to the duplicatio of all primary disks, ca tolerate up to /2 disk failures. -RMS, because of usig a diagoal data distributio algorithm, provides higher reliability agaist both disk ad data block failures. Therefore, we have coducted some simulatio experimets to compare the reliability of -RMS agaist the other mirrorig-based s, amely -1, -10 i terms of data recovery rate. But, sice -1 ad -10 have the same results, Table 5 oly shows a compariso betwee the proposed ad -10. These results express the probability with which the blocks of a failed disk ca be retrieved. I these experimets, the failures are geerated with a uiform distributio withi the disk arrays ad the data recovery rate is measured as the Table 5 The disk failures recovery rate of -RMS agaist -10 of disks of failures RMS

7 46 computers & security 28 (2009) umber of disks icreases from 3 to 16 ad the umber of failures icreases from 1 to /2 i each experimet. 6. Coclusio I this paper, we preseted a ew i which a special hybrid data distributio techique is used to locate the data blocks i the cluster. The mai idea behid the proposed algorithm is to combie the data block stripig ad disk mirrorig techiques with a data block rotatio to improve the parallelism, reliability ad efficiecy of the array. I such a system, if a sigle disk fails, a effectively balaced load ca be attaied at all survivig disks. It is show that the proposed is able to serve may more disk requests compared to the other existig mirrorig-based s. Besides, such ca tolerate more disk failures ad decrease the rate of missig data due to the preseted data distributio. So, this ca icrease the parallelism for processig requests, the reliability ad the effective badwidth to access data, especially after disk failures. The proposed i terms of the maximum badwidth, disk operatio cost ad reliability is theoretically aalyzed ad the obtaied results show that this outperforms the studied s i most of the ivestigated metrics ad the experimetal results completely cofirm the aalytic results. Fially, the give is appropriate to implemet i the distributed storage systems which hold huge amouts of critical data. Ackowledgemet The authors would like to thak the reviewers for their valuable commets ad costructive criticism of the origial mauscript. refereces Akbari Torkestai J, Haghighat AT. A ew redudacy algorithm for distributed eviromets. Operatig system & Security Coferece (OSSC), 2003, pp Akbari Torkestai J, Meybodi MR. A ew redudacy structure for usig i high reliable systems. Proceedigs of the Secod Iteratioal Coferece o Iformatio ad Kowledge Techology (IKT2005), 2005a. Akbari Torkestai J, Meybodi MR. A ew data mirrorig algorithm to ehace reliability ad fault tolerace. Proceedigs of the secod iteratioal coferece o iformatio ad kowledge techology (IKT2005), 2005b. Akbari Torkestai J, Izadi T, Meybodi MR., Implemetatio of a ew data security method for. Proceedigs of 11th aual CSI computer coferece of Ira, Fudametal Sciece Research Ceter (IPM), Computer Sciece Research Lab., 2006, pp Aderso T, Dahli M, Patterso D, Wag R. Serverless etwork filesystems. ACM Trasactios o Computer Systems 1996: Asami S, Talagala N, Patterso DA. Desigig a self-maitaiig storage system. Proceedigs of 16th IEEE Symposium o Mass Storage Systems, 1999, pp Cabrera LF, Log DE. Usig distributed disk stripig to provide high I/O data rates. Proceedigs of USENIX Computig Systems, 1991, pp Cao P, Lim SB, Vekatarama S, Wilkes J. The TickerTAIP Parallel Architecture. ACM Trasactios o Computer System 1994;12(3): Che SZ, Towsley D. A performace evaluatio of s. IEEE Trasactio o Computers 1996;45(10): Che PM, Lee EK, Gibso GA, Katz RH, Patterso DA. : highperformace, reliable secodary storage. ACM Computig Surveys 1994;26(2): Corbett P, Feitelso DG, Prost JP, Baylor SJ. Parallel access to files i the Vesta file system. Proceedigs of Supercomputig 93, Corme TH, Kotz D. Itegratig theory ad practice i parallel file systems. Proceedigs of DAGS 93 Symposium, 1993, pp Cortes T. Software ad parallel filesystems. I: Buyya Rajkumar, editor. High performace cluster computig s ad systems. Pretice Hall PTR; p Foster A, Kohr D, Krishaiyer JR, Mogill J. Remote I/O: Fast Access to Distat Storage, Proceedigs of the Fifth Workshop o I/O i Parallel ad Distributed Systems, 1997, pp Gibso G, Nagle D, Amiri K, Chag F, Gobioff H, Riedel E. A cost-effective, high-badwidth storage. Proceedigs of the 8th Coferece o Architectural Support for Programmig Laguages ad Operatig Systems, 1998, pp Harry M, Rosario JMD, Choudhary A. VIP-FS: a virtual, parallel file system for high performace parallel ad distributed computig. Proceedigs of the 9th Iteratioal Parallel Processig Symposium (IPPS 95), 1995, pp Ho R, Hwag K, Ji H. Desig ad aalysis of clusters with sigle I/O space,. Proceedigs of 20 Iteratioal Coferece o Distributed Computig Systems (ICDCS 2000), 2000, Taiwa, pp Howard JH, Kazar ML, Meees SG, Nichols DA, Satyaarayaa M, Sidebotham RN, West MJ. Scale ad Performace i a Distributed File System. ACM Trasactios o Computer System 1988;6(1): Hsiao HI, DeWitt D, Chaied declusterig: a ew availability strategy for multiprocessor database machies. Proceedigs of 6th Iteratioal Data Egieerig Coferece, 1990, pp Hsiao HI, DeWitt DJ. A performace study of three high availability data replicatio strategies. Joural of Distributed Parallel Databases 1993;1(1): Hu Y, Yag Q, Nightigale T. RAPID-Cache a reliable ad iexpesive cache for Disk I/O Systems. Proceedigs of the 5th Iteratioal Symposium o High Performace Computer Architecture (HPCA-5), 1999, pp Patterso DA, Gibso GA, Katz RH. A case for redudat arrays of iexpesive disks (). ACM SIGMOD, It. Cof. o Maagemet of Data, Jue 1988, pp Stoebraker M, Gerhard A, Schloss A. Distributed Raid a ew multiple copy algorithm. Berkeley, CA: Uiversity of Califoria; Tera data Corporatio. DBC/1012: database computer system maual, Release 2, Thomasia A, Blaum M. Mirrored disk orgaizatio reliability aalysis. IEEE Trasactios o Computers 2006;55(12): Thomasia. Shortcut method for reliability comparisos i. Joural of Systems ad Software 2006;79:

Improvement of the Orthogonal Code Convolution Capabilities Using FPGA Implementation

Improvement of the Orthogonal Code Convolution Capabilities Using FPGA Implementation Improvemet of the Orthogoal Code Covolutio Capabilities Usig FPGA Implemetatio Naima Kaabouch, Member, IEEE, Apara Dhirde, Member, IEEE, Saleh Faruque, Member, IEEE Departmet of Electrical Egieerig, Uiversity

More information

Operating System Concepts. Operating System Concepts

Operating System Concepts. Operating System Concepts Chapter 4: Mass-Storage Systems Logical Disk Structure Logical Disk Structure Disk Schedulig Disk Maagemet RAID Structure Disk drives are addressed as large -dimesioal arrays of logical blocks, where the

More information

3D Model Retrieval Method Based on Sample Prediction

3D Model Retrieval Method Based on Sample Prediction 20 Iteratioal Coferece o Computer Commuicatio ad Maagemet Proc.of CSIT vol.5 (20) (20) IACSIT Press, Sigapore 3D Model Retrieval Method Based o Sample Predictio Qigche Zhag, Ya Tag* School of Computer

More information

Ones Assignment Method for Solving Traveling Salesman Problem

Ones Assignment Method for Solving Traveling Salesman Problem Joural of mathematics ad computer sciece 0 (0), 58-65 Oes Assigmet Method for Solvig Travelig Salesma Problem Hadi Basirzadeh Departmet of Mathematics, Shahid Chamra Uiversity, Ahvaz, Ira Article history:

More information

New HSL Distance Based Colour Clustering Algorithm

New HSL Distance Based Colour Clustering Algorithm The 4th Midwest Artificial Itelligece ad Cogitive Scieces Coferece (MAICS 03 pp 85-9 New Albay Idiaa USA April 3-4 03 New HSL Distace Based Colour Clusterig Algorithm Vasile Patrascu Departemet of Iformatics

More information

CSC 220: Computer Organization Unit 11 Basic Computer Organization and Design

CSC 220: Computer Organization Unit 11 Basic Computer Organization and Design College of Computer ad Iformatio Scieces Departmet of Computer Sciece CSC 220: Computer Orgaizatio Uit 11 Basic Computer Orgaizatio ad Desig 1 For the rest of the semester, we ll focus o computer architecture:

More information

A Study on the Performance of Cholesky-Factorization using MPI

A Study on the Performance of Cholesky-Factorization using MPI A Study o the Performace of Cholesky-Factorizatio usig MPI Ha S. Kim Scott B. Bade Departmet of Computer Sciece ad Egieerig Uiversity of Califoria Sa Diego {hskim, bade}@cs.ucsd.edu Abstract Cholesky-factorizatio

More information

Appendix D. Controller Implementation

Appendix D. Controller Implementation COMPUTER ORGANIZATION AND DESIGN The Hardware/Software Iterface 5 th Editio Appedix D Cotroller Implemetatio Cotroller Implemetatios Combiatioal logic (sigle-cycle); Fiite state machie (multi-cycle, pipelied);

More information

Adaptive Resource Allocation for Electric Environmental Pollution through the Control Network

Adaptive Resource Allocation for Electric Environmental Pollution through the Control Network Available olie at www.sciecedirect.com Eergy Procedia 6 (202) 60 64 202 Iteratioal Coferece o Future Eergy, Eviromet, ad Materials Adaptive Resource Allocatio for Electric Evirometal Pollutio through the

More information

Algorithms for Disk Covering Problems with the Most Points

Algorithms for Disk Covering Problems with the Most Points Algorithms for Disk Coverig Problems with the Most Poits Bi Xiao Departmet of Computig Hog Kog Polytechic Uiversity Hug Hom, Kowloo, Hog Kog csbxiao@comp.polyu.edu.hk Qigfeg Zhuge, Yi He, Zili Shao, Edwi

More information

Lecture 28: Data Link Layer

Lecture 28: Data Link Layer Automatic Repeat Request (ARQ) 2. Go ack N ARQ Although the Stop ad Wait ARQ is very simple, you ca easily show that it has very the low efficiecy. The low efficiecy comes from the fact that the trasmittig

More information

Load balanced Parallel Prime Number Generator with Sieve of Eratosthenes on Cluster Computers *

Load balanced Parallel Prime Number Generator with Sieve of Eratosthenes on Cluster Computers * Load balaced Parallel Prime umber Geerator with Sieve of Eratosthees o luster omputers * Soowook Hwag*, Kyusik hug**, ad Dogseug Kim* *Departmet of Electrical Egieerig Korea Uiversity Seoul, -, Rep. of

More information

On-line cursive letter recognition using sequences of local minima/maxima. Robert Powalka

On-line cursive letter recognition using sequences of local minima/maxima. Robert Powalka O-lie cursive letter recogitio usig sequeces of local miima/maxima Summary Robert Powalka 19 th August 1993 This report presets the desig ad implemetatio of a o-lie cursive letter recogizer usig sequeces

More information

Reversible Realization of Quaternary Decoder, Multiplexer, and Demultiplexer Circuits

Reversible Realization of Quaternary Decoder, Multiplexer, and Demultiplexer Circuits Egieerig Letters, :, EL Reversible Realizatio of Quaterary Decoder, Multiplexer, ad Demultiplexer Circuits Mozammel H.. Kha, Member, ENG bstract quaterary reversible circuit is more compact tha the correspodig

More information

Evaluation scheme for Tracking in AMI

Evaluation scheme for Tracking in AMI A M I C o m m u i c a t i o A U G M E N T E D M U L T I - P A R T Y I N T E R A C T I O N http://www.amiproject.org/ Evaluatio scheme for Trackig i AMI S. Schreiber a D. Gatica-Perez b AMI WP4 Trackig:

More information

Pruning and Summarizing the Discovered Time Series Association Rules from Mechanical Sensor Data Qing YANG1,a,*, Shao-Yu WANG1,b, Ting-Ting ZHANG2,c

Pruning and Summarizing the Discovered Time Series Association Rules from Mechanical Sensor Data Qing YANG1,a,*, Shao-Yu WANG1,b, Ting-Ting ZHANG2,c Advaces i Egieerig Research (AER), volume 131 3rd Aual Iteratioal Coferece o Electroics, Electrical Egieerig ad Iformatio Sciece (EEEIS 2017) Pruig ad Summarizig the Discovered Time Series Associatio Rules

More information

9.1. Sequences and Series. Sequences. What you should learn. Why you should learn it. Definition of Sequence

9.1. Sequences and Series. Sequences. What you should learn. Why you should learn it. Definition of Sequence _9.qxd // : AM Page Chapter 9 Sequeces, Series, ad Probability 9. Sequeces ad Series What you should lear Use sequece otatio to write the terms of sequeces. Use factorial otatio. Use summatio otatio to

More information

An Efficient Algorithm for Graph Bisection of Triangularizations

An Efficient Algorithm for Graph Bisection of Triangularizations A Efficiet Algorithm for Graph Bisectio of Triagularizatios Gerold Jäger Departmet of Computer Sciece Washigto Uiversity Campus Box 1045 Oe Brookigs Drive St. Louis, Missouri 63130-4899, USA jaegerg@cse.wustl.edu

More information

Fast Fourier Transform (FFT) Algorithms

Fast Fourier Transform (FFT) Algorithms Fast Fourier Trasform FFT Algorithms Relatio to the z-trasform elsewhere, ozero, z x z X x [ ] 2 ~ elsewhere,, ~ e j x X x x π j e z z X X π 2 ~ The DFS X represets evely spaced samples of the z- trasform

More information

A New Morphological 3D Shape Decomposition: Grayscale Interframe Interpolation Method

A New Morphological 3D Shape Decomposition: Grayscale Interframe Interpolation Method A ew Morphological 3D Shape Decompositio: Grayscale Iterframe Iterpolatio Method D.. Vizireau Politehica Uiversity Bucharest, Romaia ae@comm.pub.ro R. M. Udrea Politehica Uiversity Bucharest, Romaia mihea@comm.pub.ro

More information

arxiv: v2 [cs.ds] 24 Mar 2018

arxiv: v2 [cs.ds] 24 Mar 2018 Similar Elemets ad Metric Labelig o Complete Graphs arxiv:1803.08037v [cs.ds] 4 Mar 018 Pedro F. Felzeszwalb Brow Uiversity Providece, RI, USA pff@brow.edu March 8, 018 We cosider a problem that ivolves

More information

Elementary Educational Computer

Elementary Educational Computer Chapter 5 Elemetary Educatioal Computer. Geeral structure of the Elemetary Educatioal Computer (EEC) The EEC coforms to the 5 uits structure defied by vo Neuma's model (.) All uits are preseted i a simplified

More information

BOOLEAN MATHEMATICS: GENERAL THEORY

BOOLEAN MATHEMATICS: GENERAL THEORY CHAPTER 3 BOOLEAN MATHEMATICS: GENERAL THEORY 3.1 ISOMORPHIC PROPERTIES The ame Boolea Arithmetic was chose because it was discovered that literal Boolea Algebra could have a isomorphic umerical aspect.

More information

Chapter 1. Introduction to Computers and C++ Programming. Copyright 2015 Pearson Education, Ltd.. All rights reserved.

Chapter 1. Introduction to Computers and C++ Programming. Copyright 2015 Pearson Education, Ltd.. All rights reserved. Chapter 1 Itroductio to Computers ad C++ Programmig Copyright 2015 Pearso Educatio, Ltd.. All rights reserved. Overview 1.1 Computer Systems 1.2 Programmig ad Problem Solvig 1.3 Itroductio to C++ 1.4 Testig

More information

Optimization for framework design of new product introduction management system Ma Ying, Wu Hongcui

Optimization for framework design of new product introduction management system Ma Ying, Wu Hongcui 2d Iteratioal Coferece o Electrical, Computer Egieerig ad Electroics (ICECEE 2015) Optimizatio for framework desig of ew product itroductio maagemet system Ma Yig, Wu Hogcui Tiaji Electroic Iformatio Vocatioal

More information

Improving Information Retrieval System Security via an Optimal Maximal Coding Scheme

Improving Information Retrieval System Security via an Optimal Maximal Coding Scheme Improvig Iformatio Retrieval System Security via a Optimal Maximal Codig Scheme Dogyag Log Departmet of Computer Sciece, City Uiversity of Hog Kog, 8 Tat Chee Aveue Kowloo, Hog Kog SAR, PRC dylog@cs.cityu.edu.hk

More information

New Fuzzy Color Clustering Algorithm Based on hsl Similarity

New Fuzzy Color Clustering Algorithm Based on hsl Similarity IFSA-EUSFLAT 009 New Fuzzy Color Clusterig Algorithm Based o hsl Similarity Vasile Ptracu Departmet of Iformatics Techology Tarom Compay Bucharest Romaia Email: patrascu.v@gmail.com Abstract I this paper

More information

Mobile terminal 3D image reconstruction program development based on Android Lin Qinhua

Mobile terminal 3D image reconstruction program development based on Android Lin Qinhua Iteratioal Coferece o Automatio, Mechaical Cotrol ad Computatioal Egieerig (AMCCE 05) Mobile termial 3D image recostructio program developmet based o Adroid Li Qihua Sichua Iformatio Techology College

More information

HADOOP: A NEW APPROACH FOR DOCUMENT CLUSTERING

HADOOP: A NEW APPROACH FOR DOCUMENT CLUSTERING Y.K. Patil* Iteratioal Joural of Advaced Research i ISSN: 2278-6244 IT ad Egieerig Impact Factor: 4.54 HADOOP: A NEW APPROACH FOR DOCUMENT CLUSTERING Prof. V.S. Nadedkar** Abstract: Documet clusterig is

More information

Quorum Based Data Replication in Grid Environment

Quorum Based Data Replication in Grid Environment Quorum Based Data Replicatio i Grid Eviromet Rohaya Latip, Hamidah Ibrahim, Mohamed Othma, Md Nasir Sulaima, ad Azizol Abdullah Faculty of Computer Sciece ad Iformatio Techology, Uiversiti Putra Malaysia

More information

Chapter 3 Classification of FFT Processor Algorithms

Chapter 3 Classification of FFT Processor Algorithms Chapter Classificatio of FFT Processor Algorithms The computatioal complexity of the Discrete Fourier trasform (DFT) is very high. It requires () 2 complex multiplicatios ad () complex additios [5]. As

More information

Redundancy Allocation for Series Parallel Systems with Multiple Constraints and Sensitivity Analysis

Redundancy Allocation for Series Parallel Systems with Multiple Constraints and Sensitivity Analysis IOSR Joural of Egieerig Redudacy Allocatio for Series Parallel Systems with Multiple Costraits ad Sesitivity Aalysis S. V. Suresh Babu, D.Maheswar 2, G. Ragaath 3 Y.Viaya Kumar d G.Sakaraiah e (Mechaical

More information

GPUMP: a Multiple-Precision Integer Library for GPUs

GPUMP: a Multiple-Precision Integer Library for GPUs GPUMP: a Multiple-Precisio Iteger Library for GPUs Kaiyog Zhao ad Xiaowe Chu Departmet of Computer Sciece, Hog Kog Baptist Uiversity Hog Kog, P. R. Chia Email: {kyzhao, chxw}@comp.hkbu.edu.hk Abstract

More information

n Learn how resiliency strategies reduce risk n Discover automation strategies to reduce risk

n Learn how resiliency strategies reduce risk n Discover automation strategies to reduce risk Chapter Objectives Lear how resiliecy strategies reduce risk Discover automatio strategies to reduce risk Chapter #16: Architecture ad Desig Resiliecy ad Automatio Strategies 2 Automatio/Scriptig Resiliet

More information

An Improved Shuffled Frog-Leaping Algorithm for Knapsack Problem

An Improved Shuffled Frog-Leaping Algorithm for Knapsack Problem A Improved Shuffled Frog-Leapig Algorithm for Kapsack Problem Zhoufag Li, Ya Zhou, ad Peg Cheg School of Iformatio Sciece ad Egieerig Hea Uiversity of Techology ZhegZhou, Chia lzhf1978@126.com Abstract.

More information

Accuracy Improvement in Camera Calibration

Accuracy Improvement in Camera Calibration Accuracy Improvemet i Camera Calibratio FaJie L Qi Zag ad Reihard Klette CITR, Computer Sciece Departmet The Uiversity of Aucklad Tamaki Campus, Aucklad, New Zealad fli006, qza001@ec.aucklad.ac.z r.klette@aucklad.ac.z

More information

Morgan Kaufmann Publishers 26 February, COMPUTER ORGANIZATION AND DESIGN The Hardware/Software Interface. Chapter 5.

Morgan Kaufmann Publishers 26 February, COMPUTER ORGANIZATION AND DESIGN The Hardware/Software Interface. Chapter 5. Morga Kaufma Publishers 26 February, 208 COMPUTER ORGANIZATION AND DESIGN The Hardware/Software Iterface 5 th Editio Chapter 5 Virtual Memory Review: The Memory Hierarchy Take advatage of the priciple

More information

Avid Interplay Bundle

Avid Interplay Bundle Avid Iterplay Budle Versio 2.5 Cofigurator ReadMe Overview This documet provides a overview of Iterplay Budle v2.5 ad describes how to ru the Iterplay Budle cofiguratio tool. Iterplay Budle v2.5 refers

More information

Solving Fuzzy Assignment Problem Using Fourier Elimination Method

Solving Fuzzy Assignment Problem Using Fourier Elimination Method Global Joural of Pure ad Applied Mathematics. ISSN 0973-768 Volume 3, Number 2 (207), pp. 453-462 Research Idia Publicatios http://www.ripublicatio.com Solvig Fuzzy Assigmet Problem Usig Fourier Elimiatio

More information

. Written in factored form it is easy to see that the roots are 2, 2, i,

. Written in factored form it is easy to see that the roots are 2, 2, i, CMPS A Itroductio to Programmig Programmig Assigmet 4 I this assigmet you will write a java program that determies the real roots of a polyomial that lie withi a specified rage. Recall that the roots (or

More information

Probability of collisions in Soft Input Decryption

Probability of collisions in Soft Input Decryption Issue 1, Volume 1, 007 1 Probability of collisios i Soft Iput Decryptio Nataša Živić, Christoph Rulad Abstract I this work, probability of collisio i Soft Iput Decryptio has bee aalyzed ad calculated.

More information

LU Decomposition Method

LU Decomposition Method SOLUTION OF SIMULTANEOUS LINEAR EQUATIONS LU Decompositio Method Jamie Traha, Autar Kaw, Kevi Marti Uiversity of South Florida Uited States of America kaw@eg.usf.edu http://umericalmethods.eg.usf.edu Itroductio

More information

On (K t e)-saturated Graphs

On (K t e)-saturated Graphs Noame mauscript No. (will be iserted by the editor O (K t e-saturated Graphs Jessica Fuller Roald J. Gould the date of receipt ad acceptace should be iserted later Abstract Give a graph H, we say a graph

More information

Stone Images Retrieval Based on Color Histogram

Stone Images Retrieval Based on Color Histogram Stoe Images Retrieval Based o Color Histogram Qiag Zhao, Jie Yag, Jigyi Yag, Hogxig Liu School of Iformatio Egieerig, Wuha Uiversity of Techology Wuha, Chia Abstract Stoe images color features are chose

More information

A SOFTWARE MODEL FOR THE MULTILAYER PERCEPTRON

A SOFTWARE MODEL FOR THE MULTILAYER PERCEPTRON A SOFTWARE MODEL FOR THE MULTILAYER PERCEPTRON Roberto Lopez ad Eugeio Oñate Iteratioal Ceter for Numerical Methods i Egieerig (CIMNE) Edificio C1, Gra Capitá s/, 08034 Barceloa, Spai ABSTRACT I this work

More information

IMP: Superposer Integrated Morphometrics Package Superposition Tool

IMP: Superposer Integrated Morphometrics Package Superposition Tool IMP: Superposer Itegrated Morphometrics Package Superpositio Tool Programmig by: David Lieber ( 03) Caisius College 200 Mai St. Buffalo, NY 4208 Cocept by: H. David Sheets, Dept. of Physics, Caisius College

More information

Searching a Russian Document Collection Using English, Chinese and Japanese Queries

Searching a Russian Document Collection Using English, Chinese and Japanese Queries Searchig a Russia Documet Collectio Usig Eglish, Chiese ad Japaese Queries Fredric C. Gey (gey@ucdata.berkeley.edu) UC Data Archive & Techical Assistace Uiversity of Califoria, Berkeley, CA 94720 USA ABSTRACT.

More information

The Counterchanged Crossed Cube Interconnection Network and Its Topology Properties

The Counterchanged Crossed Cube Interconnection Network and Its Topology Properties WSEAS TRANSACTIONS o COMMUNICATIONS Wag Xiyag The Couterchaged Crossed Cube Itercoectio Network ad Its Topology Properties WANG XINYANG School of Computer Sciece ad Egieerig South Chia Uiversity of Techology

More information

Low Complexity H.265/HEVC Coding Unit Size Decision for a Videoconferencing System

Low Complexity H.265/HEVC Coding Unit Size Decision for a Videoconferencing System BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 15, No 6 Special Issue o Logistics, Iformatics ad Service Sciece Sofia 2015 Prit ISSN: 1311-9702; Olie ISSN: 1314-4081 DOI:

More information

Analysis Metrics. Intro to Algorithm Analysis. Slides. 12. Alg Analysis. 12. Alg Analysis

Analysis Metrics. Intro to Algorithm Analysis. Slides. 12. Alg Analysis. 12. Alg Analysis Itro to Algorithm Aalysis Aalysis Metrics Slides. Table of Cotets. Aalysis Metrics 3. Exact Aalysis Rules 4. Simple Summatio 5. Summatio Formulas 6. Order of Magitude 7. Big-O otatio 8. Big-O Theorems

More information

An Efficient Algorithm for Graph Bisection of Triangularizations

An Efficient Algorithm for Graph Bisection of Triangularizations Applied Mathematical Scieces, Vol. 1, 2007, o. 25, 1203-1215 A Efficiet Algorithm for Graph Bisectio of Triagularizatios Gerold Jäger Departmet of Computer Sciece Washigto Uiversity Campus Box 1045, Oe

More information

performance to the performance they can experience when they use the services from a xed location.

performance to the performance they can experience when they use the services from a xed location. I the Proceedigs of The First Aual Iteratioal Coferece o Mobile Computig ad Networkig (MobiCom 9) November -, 99, Berkeley, Califoria USA Performace Compariso of Mobile Support Strategies Rieko Kadobayashi

More information

Enhancing Efficiency of Software Fault Tolerance Techniques in Satellite Motion System

Enhancing Efficiency of Software Fault Tolerance Techniques in Satellite Motion System Joural of Iformatio Systems ad Telecommuicatio, Vol. 2, No. 3, July-September 2014 173 Ehacig Efficiecy of Software Fault Tolerace Techiques i Satellite Motio System Hoda Baki Departmet of Electrical ad

More information

An Algorithm to Solve Multi-Objective Assignment. Problem Using Interactive Fuzzy. Goal Programming Approach

An Algorithm to Solve Multi-Objective Assignment. Problem Using Interactive Fuzzy. Goal Programming Approach It. J. Cotemp. Math. Scieces, Vol. 6, 0, o. 34, 65-66 A Algorm to Solve Multi-Objective Assigmet Problem Usig Iteractive Fuzzy Goal Programmig Approach P. K. De ad Bharti Yadav Departmet of Mathematics

More information

CIS 121 Data Structures and Algorithms with Java Fall Big-Oh Notation Tuesday, September 5 (Make-up Friday, September 8)

CIS 121 Data Structures and Algorithms with Java Fall Big-Oh Notation Tuesday, September 5 (Make-up Friday, September 8) CIS 11 Data Structures ad Algorithms with Java Fall 017 Big-Oh Notatio Tuesday, September 5 (Make-up Friday, September 8) Learig Goals Review Big-Oh ad lear big/small omega/theta otatios Practice solvig

More information

Using a Dynamic Interval Type-2 Fuzzy Interpolation Method to Improve Modeless Robots Calibrations

Using a Dynamic Interval Type-2 Fuzzy Interpolation Method to Improve Modeless Robots Calibrations Joural of Cotrol Sciece ad Egieerig 3 (25) 9-7 doi:.7265/2328-223/25.3. D DAVID PUBLISHING Usig a Dyamic Iterval Type-2 Fuzzy Iterpolatio Method to Improve Modeless Robots Calibratios Yig Bai ad Dali Wag

More information

Python Programming: An Introduction to Computer Science

Python Programming: An Introduction to Computer Science Pytho Programmig: A Itroductio to Computer Sciece Chapter 1 Computers ad Programs 1 Objectives To uderstad the respective roles of hardware ad software i a computig system. To lear what computer scietists

More information

Alpha Individual Solutions MAΘ National Convention 2013

Alpha Individual Solutions MAΘ National Convention 2013 Alpha Idividual Solutios MAΘ Natioal Covetio 0 Aswers:. D. A. C 4. D 5. C 6. B 7. A 8. C 9. D 0. B. B. A. D 4. C 5. A 6. C 7. B 8. A 9. A 0. C. E. B. D 4. C 5. A 6. D 7. B 8. C 9. D 0. B TB. 570 TB. 5

More information

CS2410 Computer Architecture. Flynn s Taxonomy

CS2410 Computer Architecture. Flynn s Taxonomy CS2410 Computer Architecture Dept. of Computer Sciece Uiversity of Pittsburgh http://www.cs.pitt.edu/~melhem/courses/2410p/idex.html 1 Fly s Taxoomy SISD Sigle istructio stream Sigle data stream (SIMD)

More information

Lecture 1: Introduction and Strassen s Algorithm

Lecture 1: Introduction and Strassen s Algorithm 5-750: Graduate Algorithms Jauary 7, 08 Lecture : Itroductio ad Strasse s Algorithm Lecturer: Gary Miller Scribe: Robert Parker Itroductio Machie models I this class, we will primarily use the Radom Access

More information

AN EFFICIENT GROUP KEY MANAGEMENT USING CODE FOR KEY CALCULATION FOR SIMULTANEOUS JOIN/LEAVE: CKCS

AN EFFICIENT GROUP KEY MANAGEMENT USING CODE FOR KEY CALCULATION FOR SIMULTANEOUS JOIN/LEAVE: CKCS Iteratioal Joural of Computer Networks & Commuicatios (IJCNC) Vol.4, No.4, July 01 AN EFFICIENT GROUP KEY MANAGEMENT USING CODE FOR KEY CALCULATION FOR SIMULTANEOUS JOIN/LEAVE: CKCS Melisa Hajyvahabzadeh

More information

Bayesian approach to reliability modelling for a probability of failure on demand parameter

Bayesian approach to reliability modelling for a probability of failure on demand parameter Bayesia approach to reliability modellig for a probability of failure o demad parameter BÖRCSÖK J., SCHAEFER S. Departmet of Computer Architecture ad System Programmig Uiversity Kassel, Wilhelmshöher Allee

More information

Data diverse software fault tolerance techniques

Data diverse software fault tolerance techniques Data diverse software fault tolerace techiques Complemets desig diversity by compesatig for desig diversity s s limitatios Ivolves obtaiig a related set of poits i the program data space, executig the

More information

Extending The Sleuth Kit and its Underlying Model for Pooled Storage File System Forensic Analysis

Extending The Sleuth Kit and its Underlying Model for Pooled Storage File System Forensic Analysis Extedig The Sleuth Kit ad its Uderlyig Model for Pooled File System Foresic Aalysis Frauhofer Istitute for Commuicatio, Iformatio Processig ad Ergoomics Ja-Niclas Hilgert* Marti Lambertz Daiel Plohma ja-iclas.hilgert@fkie.frauhofer.de

More information

n n B. How many subsets of C are there of cardinality n. We are selecting elements for such a

n n B. How many subsets of C are there of cardinality n. We are selecting elements for such a 4. [10] Usig a combiatorial argumet, prove that for 1: = 0 = Let A ad B be disjoit sets of cardiality each ad C = A B. How may subsets of C are there of cardiality. We are selectig elemets for such a subset

More information

Effect of control points distribution on the orthorectification accuracy of an Ikonos II image through rational polynomial functions

Effect of control points distribution on the orthorectification accuracy of an Ikonos II image through rational polynomial functions Effect of cotrol poits distributio o the orthorectificatio accuracy of a Ikoos II image through ratioal polyomial fuctios Marcela do Valle Machado 1, Mauro Homem Atues 1 ad Paula Debiasi 1 1 Federal Rural

More information

How do we evaluate algorithms?

How do we evaluate algorithms? F2 Readig referece: chapter 2 + slides Algorithm complexity Big O ad big Ω To calculate ruig time Aalysis of recursive Algorithms Next time: Litterature: slides mostly The first Algorithm desig methods:

More information

Lecture Notes 6 Introduction to algorithm analysis CSS 501 Data Structures and Object-Oriented Programming

Lecture Notes 6 Introduction to algorithm analysis CSS 501 Data Structures and Object-Oriented Programming Lecture Notes 6 Itroductio to algorithm aalysis CSS 501 Data Structures ad Object-Orieted Programmig Readig for this lecture: Carrao, Chapter 10 To be covered i this lecture: Itroductio to algorithm aalysis

More information

High Availability with Diagonal Replication in 2D Mesh (DR2M) Protocol for Grid Environment

High Availability with Diagonal Replication in 2D Mesh (DR2M) Protocol for Grid Environment High Availability with Diagoal Replicatio i 2D Mesh (DR2M) Protocol for Grid Eviromet Rohaya Latip (Correspodig author) Faculty of Computer Sciece ad Iformatio Techology Uiversiti Putra Malaysia Tel: 60-3-8946-6536

More information

BASED ON ITERATIVE ERROR-CORRECTION

BASED ON ITERATIVE ERROR-CORRECTION A COHPARISO OF CRYPTAALYTIC PRICIPLES BASED O ITERATIVE ERROR-CORRECTIO Miodrag J. MihaljeviC ad Jova Dj. GoliC Istitute of Applied Mathematics ad Electroics. Belgrade School of Electrical Egieerig. Uiversity

More information

Economical Structure for Multi-feature Music Indexing 1

Economical Structure for Multi-feature Music Indexing 1 Proceedigs of the Iteratioal MultiCoferece of Egieers ad Computer Scietists 2008 Vol I IMECS 2008, 19-21 March, 2008, Hog Kog Ecoomical Structure for Multi-feature Music Idexig 1 Yu-Lug Lo, ad Chu-Hsiug

More information

Dynamic Programming and Curve Fitting Based Road Boundary Detection

Dynamic Programming and Curve Fitting Based Road Boundary Detection Dyamic Programmig ad Curve Fittig Based Road Boudary Detectio SHYAM PRASAD ADHIKARI, HYONGSUK KIM, Divisio of Electroics ad Iformatio Egieerig Chobuk Natioal Uiversity 664-4 Ga Deokji-Dog Jeoju-City Jeobuk

More information

Running Time. Analysis of Algorithms. Experimental Studies. Limitations of Experiments

Running Time. Analysis of Algorithms. Experimental Studies. Limitations of Experiments Ruig Time Aalysis of Algorithms Iput Algorithm Output A algorithm is a step-by-step procedure for solvig a problem i a fiite amout of time. Most algorithms trasform iput objects ito output objects. The

More information

EE260: Digital Design, Spring /16/18. n Example: m 0 (=x 1 x 2 ) is adjacent to m 1 (=x 1 x 2 ) and m 2 (=x 1 x 2 ) but NOT m 3 (=x 1 x 2 )

EE260: Digital Design, Spring /16/18. n Example: m 0 (=x 1 x 2 ) is adjacent to m 1 (=x 1 x 2 ) and m 2 (=x 1 x 2 ) but NOT m 3 (=x 1 x 2 ) EE26: Digital Desig, Sprig 28 3/6/8 EE 26: Itroductio to Digital Desig Combiatioal Datapath Yao Zheg Departmet of Electrical Egieerig Uiversity of Hawaiʻi at Māoa Combiatioal Logic Blocks Multiplexer Ecoders/Decoders

More information

On Nonblocking Folded-Clos Networks in Computer Communication Environments

On Nonblocking Folded-Clos Networks in Computer Communication Environments O Noblockig Folded-Clos Networks i Computer Commuicatio Eviromets Xi Yua Departmet of Computer Sciece, Florida State Uiversity, Tallahassee, FL 3306 xyua@cs.fsu.edu Abstract Folded-Clos etworks, also referred

More information

1.2 Binomial Coefficients and Subsets

1.2 Binomial Coefficients and Subsets 1.2. BINOMIAL COEFFICIENTS AND SUBSETS 13 1.2 Biomial Coefficiets ad Subsets 1.2-1 The loop below is part of a program to determie the umber of triagles formed by poits i the plae. for i =1 to for j =

More information

Euclidean Distance Based Feature Selection for Fault Detection Prediction Model in Semiconductor Manufacturing Process

Euclidean Distance Based Feature Selection for Fault Detection Prediction Model in Semiconductor Manufacturing Process Vol.133 (Iformatio Techology ad Computer Sciece 016), pp.85-89 http://dx.doi.org/10.1457/astl.016. Euclidea Distace Based Feature Selectio for Fault Detectio Predictio Model i Semicoductor Maufacturig

More information

Improving Template Based Spike Detection

Improving Template Based Spike Detection Improvig Template Based Spike Detectio Kirk Smith, Member - IEEE Portlad State Uiversity petra@ee.pdx.edu Abstract Template matchig algorithms like SSE, Covolutio ad Maximum Likelihood are well kow for

More information

The isoperimetric problem on the hypercube

The isoperimetric problem on the hypercube The isoperimetric problem o the hypercube Prepared by: Steve Butler November 2, 2005 1 The isoperimetric problem We will cosider the -dimesioal hypercube Q Recall that the hypercube Q is a graph whose

More information

Data Structures and Algorithms. Analysis of Algorithms

Data Structures and Algorithms. Analysis of Algorithms Data Structures ad Algorithms Aalysis of Algorithms Outlie Ruig time Pseudo-code Big-oh otatio Big-theta otatio Big-omega otatio Asymptotic algorithm aalysis Aalysis of Algorithms Iput Algorithm Output

More information

Fuzzy Minimal Solution of Dual Fully Fuzzy Matrix Equations

Fuzzy Minimal Solution of Dual Fully Fuzzy Matrix Equations Iteratioal Coferece o Applied Mathematics, Simulatio ad Modellig (AMSM 2016) Fuzzy Miimal Solutio of Dual Fully Fuzzy Matrix Equatios Dequa Shag1 ad Xiaobi Guo2,* 1 Sciece Courses eachig Departmet, Gasu

More information

CMSC Computer Architecture Lecture 12: Virtual Memory. Prof. Yanjing Li University of Chicago

CMSC Computer Architecture Lecture 12: Virtual Memory. Prof. Yanjing Li University of Chicago CMSC 22200 Computer Architecture Lecture 12: Virtual Memory Prof. Yajig Li Uiversity of Chicago A System with Physical Memory Oly Examples: most Cray machies early PCs Memory early all embedded systems

More information

Parallel Polygon Approximation Algorithm Targeted at Reconfigurable Multi-Ring Hardware

Parallel Polygon Approximation Algorithm Targeted at Reconfigurable Multi-Ring Hardware Parallel Polygo Approximatio Algorithm Targeted at Recofigurable Multi-Rig Hardware M. Arif Wai* ad Hamid R. Arabia** *Califoria State Uiversity Bakersfield, Califoria, USA **Uiversity of Georgia, Georgia,

More information

BOOLEAN DIFFERENTIATION EQUATIONS APPLICABLE IN RECONFIGURABLE COMPUTATIONAL MEDIUM

BOOLEAN DIFFERENTIATION EQUATIONS APPLICABLE IN RECONFIGURABLE COMPUTATIONAL MEDIUM MATEC Web of Cofereces 79, 01014 (016) DOI: 10.1051/ mateccof/0167901014 T 016 BOOLEAN DIFFERENTIATION EQUATIONS APPLICABLE IN RECONFIGURABLE COMPUTATIONAL MEDIUM Staislav Shidlovskiy 1, 1 Natioal Research

More information

Chapter 10. Defining Classes. Copyright 2015 Pearson Education, Ltd.. All rights reserved.

Chapter 10. Defining Classes. Copyright 2015 Pearson Education, Ltd.. All rights reserved. Chapter 10 Defiig Classes Copyright 2015 Pearso Educatio, Ltd.. All rights reserved. Overview 10.1 Structures 10.2 Classes 10.3 Abstract Data Types 10.4 Itroductio to Iheritace Copyright 2015 Pearso Educatio,

More information

Primitive polynomials selection method for pseudo-random number generator

Primitive polynomials selection method for pseudo-random number generator Joural of hysics: Coferece Series AER OEN ACCESS rimitive polyomials selectio method for pseudo-radom umber geerator To cite this article: I V Aiki ad Kh Alajjar 08 J. hys.: Cof. Ser. 944 0003 View the

More information

ISSN (Print) Research Article. *Corresponding author Nengfa Hu

ISSN (Print) Research Article. *Corresponding author Nengfa Hu Scholars Joural of Egieerig ad Techology (SJET) Sch. J. Eg. Tech., 2016; 4(5):249-253 Scholars Academic ad Scietific Publisher (A Iteratioal Publisher for Academic ad Scietific Resources) www.saspublisher.com

More information

APPLICATION NOTE PACE1750AE BUILT-IN FUNCTIONS

APPLICATION NOTE PACE1750AE BUILT-IN FUNCTIONS APPLICATION NOTE PACE175AE BUILT-IN UNCTIONS About This Note This applicatio brief is iteded to explai ad demostrate the use of the special fuctios that are built ito the PACE175AE processor. These powerful

More information

Harris Corner Detection Algorithm at Sub-pixel Level and Its Application Yuanfeng Han a, Peijiang Chen b * and Tian Meng c

Harris Corner Detection Algorithm at Sub-pixel Level and Its Application Yuanfeng Han a, Peijiang Chen b * and Tian Meng c Iteratioal Coferece o Computatioal Sciece ad Egieerig (ICCSE 015) Harris Corer Detectio Algorithm at Sub-pixel Level ad Its Applicatio Yuafeg Ha a, Peijiag Che b * ad Tia Meg c School of Automobile, Liyi

More information

FACIAL MOTION TRACKING AND ANIMATION: AN ICA-BASED APPROACH

FACIAL MOTION TRACKING AND ANIMATION: AN ICA-BASED APPROACH FACIAL MOTION TRACKING AND ANIMATION: AN ICA-BASED APPROACH Lucas D Terissi ad Jua C Gómez Laboratory for System Dyamics ad Sigal Processig FCEIA, Uiversidad Nacioal de Rosario Riobamba 245 Bis, 2000 Rosario

More information

GE FUNDAMENTALS OF COMPUTING AND PROGRAMMING UNIT III

GE FUNDAMENTALS OF COMPUTING AND PROGRAMMING UNIT III GE2112 - FUNDAMENTALS OF COMPUTING AND PROGRAMMING UNIT III PROBLEM SOLVING AND OFFICE APPLICATION SOFTWARE Plaig the Computer Program Purpose Algorithm Flow Charts Pseudocode -Applicatio Software Packages-

More information

A Comparative Study on the Algorithms for a Generalized Josephus Problem

A Comparative Study on the Algorithms for a Generalized Josephus Problem Appl. Math. If. Sci. 7, No. 4, 1451-1457 (2013) 1451 Applied Mathematics & Iformatio Scieces A Iteratioal Joural http://dx.i.org/10.12785/amis/070425 A Comparative Study o the Algorithms for a Geeralized

More information

Cache-Optimal Methods for Bit-Reversals

Cache-Optimal Methods for Bit-Reversals Proceedigs of the ACM/IEEE Supercomputig Coferece, November 1999, Portlad, Orego, U.S.A. Cache-Optimal Methods for Bit-Reversals Zhao Zhag ad Xiaodog Zhag Departmet of Computer Sciece College of William

More information

Hash Tables. Presentation for use with the textbook Algorithm Design and Applications, by M. T. Goodrich and R. Tamassia, Wiley, 2015.

Hash Tables. Presentation for use with the textbook Algorithm Design and Applications, by M. T. Goodrich and R. Tamassia, Wiley, 2015. Presetatio for use with the textbook Algorithm Desig ad Applicatios, by M. T. Goodrich ad R. Tamassia, Wiley, 2015 Hash Tables xkcd. http://xkcd.com/221/. Radom Number. Used with permissio uder Creative

More information

Image Segmentation EEE 508

Image Segmentation EEE 508 Image Segmetatio Objective: to determie (etract) object boudaries. It is a process of partitioig a image ito distict regios by groupig together eighborig piels based o some predefied similarity criterio.

More information

Copyright 2016 Ramez Elmasri and Shamkant B. Navathe

Copyright 2016 Ramez Elmasri and Shamkant B. Navathe Copyright 2016 Ramez Elmasri ad Shamkat B. Navathe CHAPTER 22 Database Recovery Techiques Copyright 2016 Ramez Elmasri ad Shamkat B. Navathe Itroductio Recovery algorithms Recovery cocepts Write-ahead

More information

EE 459/500 HDL Based Digital Design with Programmable Logic. Lecture 13 Control and Sequencing: Hardwired and Microprogrammed Control

EE 459/500 HDL Based Digital Design with Programmable Logic. Lecture 13 Control and Sequencing: Hardwired and Microprogrammed Control EE 459/500 HDL Based Digital Desig with Programmable Logic Lecture 13 Cotrol ad Sequecig: Hardwired ad Microprogrammed Cotrol Refereces: Chapter s 4,5 from textbook Chapter 7 of M.M. Mao ad C.R. Kime,

More information

Derivation of perspective stereo projection matrices with depth, shape and magnification consideration

Derivation of perspective stereo projection matrices with depth, shape and magnification consideration Derivatio of perspective stereo projectio matrices with depth, shape ad magificatio cosideratio Patrick Oberthür Jauary 2014 This essay will show how to costruct a pair of stereoscopic perspective projectio

More information

UH-MEM: Utility-Based Hybrid Memory Management. Yang Li, Saugata Ghose, Jongmoo Choi, Jin Sun, Hui Wang, Onur Mutlu

UH-MEM: Utility-Based Hybrid Memory Management. Yang Li, Saugata Ghose, Jongmoo Choi, Jin Sun, Hui Wang, Onur Mutlu UH-MEM: Utility-Based Hybrid Memory Maagemet Yag Li, Saugata Ghose, Jogmoo Choi, Ji Su, Hui Wag, Our Mutlu 1 Executive Summary DRAM faces sigificat techology scalig difficulties Emergig memory techologies

More information

On Infinite Groups that are Isomorphic to its Proper Infinite Subgroup. Jaymar Talledo Balihon. Abstract

On Infinite Groups that are Isomorphic to its Proper Infinite Subgroup. Jaymar Talledo Balihon. Abstract O Ifiite Groups that are Isomorphic to its Proper Ifiite Subgroup Jaymar Talledo Baliho Abstract Two groups are isomorphic if there exists a isomorphism betwee them Lagrage Theorem states that the order

More information