分析案例翻譯

1、Designing High Performance Web-Based ComputingServitoPromoteemedicineDatabaseManagement SystemAbstractMany web computing systems are running real time database serviwhere their informationchange continuously and expand incrementally.his context, web data servihave a majorrole and draw significant im

2、provements in monitoring and controlling the informationtruthfulness and data propagation. Currently, webemedicine database serviare of centralgrowth ofimportance to distributed systems. However, the increasing complexity and the rathe real world healthcare challenging applications make it hard to i

3、nduce the databaseadministrative staff. In this presponse time for large scale database management wir, we build anegrated web data servit satisfy faste-health database management systems. Our focus will be onpplication scenarios in dynamicemedicine systems to increasecare admiss and decrease care d

4、ifficultiech as distance, travel, and time limiions. Weproe three-fold approach based on data fragmenion, database web sites clustering andelligent data distribution. This approach reduthe amount of data migrated betn websites during applications execution; achieves cost-effectivecommunications duri

5、ng applications prosing and improves applications response time andthroughput. The proed approach is validatedernally by measuring the impact of using ourcomputing servi techniques on various performance features like communications cost,response time, and throughput. The external validation is achi

6、eved by comparing theperformance of our approach tot of other techniqueshe literature. The results showtouregrated approach significantly improves the performance of web database systems andoutperforms its countarts.Keywords: Webemedicine Database Systems (WTDS); database fragmenion; datadistributio

7、n; sites clustering.1.roductionThe ragrownd continuous change of the real world software applications have provokedresearchers to proe several computing servi techniques to achieve more efficient andeffective management of webemedicine database systems (WTDS). Significant researchprogress has been m

8、adehe past few years to improve WTDS performance. In particular,databases as a critical component of these systems have attracted many researchers. The webplays an important role in enabling healthcare servilikeemedicine to serve inacsibleareas where there are few medical resour. It offers an easy a

9、nd global acs to patients datawithouving toeract with them inand it provides fast channels to consult spelistsin emergency situations. Different kinds of patients information such as ECG, temperature, andheart rate need to be acsed by means of various cnt deviin heterogeneouscommunications environme

10、nts. WTDS enable high quality continuous delivery of patientsinformation wherever and whenever needed. Several benefits can be achieved by using webemedicirviincluding: medical consulion delivery, transporion cost savings, datastorage savings, and mobile applicationpportte obstacles related to thepe

11、rformance (e.g. bandwidth, battery life, and storage), security (e.g. privacy, and reliability), andenvironment (e.g. scalability, heterogeneity, and availability). The objectives of such serviareto: (i) develop large applicationst scale as the scope and workload increases, (ii) achieveprecise contr

12、ol and monitoring on medical dao generate highemedicine database systemperformance, (iii) provide large data archive of medical data records, accurate decisystems, and trusted event-based notificationsypical clinical centers.supportRecently, many researchers have focused on designing web medical dat

13、abase managementsystemst satisfy certain performance levels. Such performance is evaluated by measuring theamount of relevant and irrelevant data acsed and the amount of transferred medical dataduring tranions prosing time. Several techniques have been proed in order to improveemedicine database per

14、formance, optimize medical data distribution, and control medicaldata proliferation. These techniques be vedt high performance for such systems can beachieved by improvingeast one of the database web management servi, namelydatabase fragmenion, data distribution, web sites clustering, distributed ca

15、ching, and databasescalability. However, theractable time complexity of prosing large number of medicaltranions and managing huge number of communications make the design of suethods anon-trivial task. Moreover, none of the existing methods consider the three-fold servitogether whiakes them impracti

16、cablehe field of web database systems. Additionally,using multiple medical servifrom different web database providers may not fit the needs forimproving theemedicine database system performance. Furthermore, the servifromdifferent web database providers may not be compatible or in some cases it may

17、increase theprosing time because of the constras on the network 1. Finally, there has been lackhetoolst support the design,ysis and cost-effective deployments of webemedicinedatabase systems 1.Designing and develofast, efficient, and reliable incorporated techniquest can handlehuge number of medical

18、 tranions on large number of web healthcare sites in near optimalpolynomial time are key challengeshe area of WTDS. Data fragmenion, web sites clustering,and data allocation are the main components of the WTDSt continue to create great researchchallenges as their current best near optimal solutions

19、are allplete.To improve the performance of medical distributed database systems, we incorporate datafragmenion, web sites clustering, and data distribution computing servitogether in a newwebemedicine database system approach. This new approachends to decrease datacommunication, increase system thro

20、ughput, reliability, and data availability.Thedatabase tranition of webemedicine database relationso disjofragments allowsions to be executed concurrently and hence minimizes the total response time.Fragmenion typically increases the level of concurrency and, therefore, the system throughput.The ben

21、efits of generatingemedicine disjofragments cannot be deemed unless distributingthese fragments over the web sites, sot they reduce communication cost of databasetranions. Database disjofragments are initially distributed over logical clusters (a group ofweb sitest satisfy a certain physical propert

22、y, e.g. communications cost). Distributingdatabase disjofragments to clusters where a benefit allocation is achieved, rathernallocating the fragments to all web sites, have an important impact on database systemthroughput. This type of distribution reduthe number of communications required for query

23、prosingerms of retrieval and update tranions; is always a significant impact on thewebemedicine database system performance. Moreover, distributing disjofragmentsamong the web sites where it is needed most, improves database system performance byminimizing the data transferred and acsed during the e

24、xecution time, reducing the storageoverheads, and increasing availability and reliability as multiple copies of the same data areallocated.Database partitioning techniques aim at improving database systems throughput by reducing theamount of irrelevant data packets (fragments) to be acsed and transf

25、erred among different web sites. However, data fragmenion raises some difficulties; particularly when webemedicine database applications have contradictory requirementst avert breakdown of therelationo mutually exclusive fragments. Those applications whose views are defined on moren one fragment may

26、 sufferformance ruin.his case, it might be nesary to retrievedata from two or more fragments and take their join, which is costly 31. Data fragmen iontechnique describes how each fragment is derived from the database global relations. Three main classes of data fragmen ion have been discussed in the

27、 literature; horizontal 23, vertical 45, and hybrid 67. Although there are various schemes describing data partitioning, few are known for the efficiency of their algorithms and the validity of their results 33.The Clustering technique identifies groups of network sites in large web database systems

28、 anddiscovers better data distributions among them. This technique is considered to be an efficientmethods a major role in reducing the amount of transferred and acsed data duringprosing database tranions. Accordingly, clustering techniques help in eliminating the extracommunications costs betn web

29、sites and thus enhandistributed database systemsperformance 32. However, the amptions on the web communications and the restrictions onthe number of network sites, make clustering solutions impractical 1631. Moreover, someconstras about network connectivity and tranions prosing time bound the applic

30、abilityof the proed solutions to small number of clusters 910.Data distribution describes the way of allocating the disjoand their respective sites of the database system. This profragments among the web clusterss addresses the assignment of eachdata fragment to the distributed database web sites 81

31、71821. Data distribution relatedtechniques aim at improving distributed database systems performance. This can beplished by reducing the number of database fragmentst are transferred and acsedduring the execution time. Additionally, Data distribution techniques attempt to increase dataavailability,

32、elevate database reliability, and redutorage overhead 1127. However, therestrictions on database retrieval and update frequencies in some data allocation methods maynegatively affect the fragments distribution over the web sites 20.his work, we address the previous drawbacks and proe a three-fold ap

33、proachtmanages the computing web servit are required to promoteemedicine databasesystem performance. The main contributions are:1 Develop a fragmenion computing service technique by splittingemedicine databaserelationso small disjofragments. This technique generates the minimum number of disjofragme

34、ntst would be allocated to the web servershe data distribution phase. Thisurnreduthe daransferred and acsed through different web sites and accordingly reduthe communications cost.2roduce a high speed clustering service techniquet groups the webemedicine databasesitessiteso sets of clusters accordin

35、g to their communications cost. This helps in grouthe webt are more suitable to be in one cluster to minimize data allocation operations, which inturn helps to avoid allocating redundant data.Proe a new computing service technique foremedicine data allocation and redistributionservibased on tranions

36、 prosing cost functions. These functions guarantee theminimum communications cost among web sites and henceplish better data distributioncompared to allocating dao all web sites evenly.Develop a user-friendly experimental tool to perform serviofemedicine datafragmenion, web sites clustering, and fra

37、gments allocation, as well as assist databaseadministratorseasuring WTDS performance.egrateallocationemedicine database fragmenion, web sites clustering, and data fragmentso one scenario toplish ultimate webemedicine system throughput intermsofconcurrency,reliability,anddataavailability.Wecallthissc

38、enariocts theegrated-Fragmenarchitecture of the proion-Clustering- Allocation (IFCA) approach. Figure 1 dedemedicine IFCA approach.the data request is initiated from theemedicine database system sites. The requested data isdefined as SQL queriest are executed on the database relations to generate da

39、ta set records.Some of these data records may be overlapped or even redundant, which increase the I/Otranions prosing time and so the system communications overhead. To solve this problem,we execute the proed fragmenion technique which generatesemedicine disjoemedicine databasefragmentst represent t

40、he minimum number of data records. The websites are groupedo clusters by using our clustering service technique in a phase prior to dataallocation. The pure of this clustering is to reduce the communications cost needed for dataallocation. Accordingly, the proed allocation service technique is appd

41、to allocate thegenerated disjofragments at the clusterst showitive benefit allocation. Then thefragments are allocated to the sites within the selected clusters. Database administrator isresponsible for recovering any site failurehe WTDS.The remainder of the pBasic concepts of the webr isanized as f

42、ollows. Section 22 summarizes the related work.emedicine database settings and amptions are discussed inSection 3.emedicine compuion serviand estimation mare discussed in Section 4.Experimental results and performance evaluation are presented in Section 35. Finally, in Section 6,we draws and outline

43、 the future work.2. Related WorkMany research works have attempted to improve the performance of distributed databasesystems. These works have mostly investigated fragmenion, allocation and sometimesclustering problems.his section,resent the main contributions related to these problems,discuss and c

44、ompare their contributions with our proed solutions.2.1. Data FragmenionWith respect to fragmenion, the unit of data distribution is a vital ie. A relation is notappropriate for distribution as application views are usually subsets of relations *31+. Therefore,the locality of applications acses is d

45、efined on the derivative relations subsets. Hence it isimportant to divide the relationo smaller data fragments and consider it for distribution overthe network sites. The authors in 8 considered each record in each database relation as adisjofragmentt is subject for allocation in a distributed data

46、base sites. However, largenumber of database fragments is generatedhis method, causing a high communication costfor transmitting and prosing the fragments. In contrast to this approach, the authors in 11considered the whole relation as a fragment, not all the records of the fragment have to beretrie

47、ved or updated, and a selectivity matrixt indicates the percentage of acsing afragment by a tranfragments overlapion is pro.ed. However, this research suffers from data redundancy and2.2. Clustering Web SitesClustering service technique identifies groups of networking sites and discoverserestingdist

48、ributions among large web database systems. This technique is considered as an efficientmethodt has a major role in reducing transferred and acsed data during tranionsprosing 9. Moreover, groudistributed network sitesn the sites and then enhano clusters helps to eliminate thethe distributed database

49、 systemextra communication costs betperformance by minimizing the communication costs required for prosing the tranions atrun time.In a web database system environment where the number of sites has expanded tremendously and amount of data has increased enormously, the sites are required to manage th

50、ese data and should allow data transparency to the users of the database. Moreover, to have a reliabledatabase system, the tranions should be executed very fast in a flexible load balancingdatabase environment. When the number of sites in a web database system increases to a largescale, the problem

51、of supporting high system performance with consistency and availabilityconstrases crul. Different techniques could be developed for this pure; one ofthem is web sites clustering.Grouweb siteso clusters reducommunications cost and then enhantheperformance of the web database system. However, clusteri

52、ng network sites is still an openproblem and the optimal solution to this problem isplete 12. Moreover, in case of acomplex network where large numbers of sites are connected to each other, a huge number ofcommunications are required, which increases the system load and degrades its performance.The

53、authors in 13 have proed a hierarchical clustering algorithmt uses similarity upperapproximation derived from a toleranimilarity) relation and based on rough set theorytdoes not require any prior information about the data. The presented approach results in roughclusters in which an object is a memb

54、er of moren one cluster. Rough clustering can helpresearchers to discover multiple needs anderests in a sesby looking at the multipleclusterst a sesbelongs to. However, in order to carry out rough clustering, two additionalrequirements, namely, an ordered value set of each attribute and a distance m

55、easure forclustering need to be specified 14. Clustering coefficients are needed in many approaches inorder tofy the structural network properties. In 15, the authors proed higher orderclustering coefficients defined as probabilitiest determine the shortest distance betn anytwo nearest neighbors of

56、a certain node when neglecting all paths crossing this node. Thees of this method declaret the average shortest distancehe nodes neighborhoodis smallern all network distan. However, independent constant values and naturallogarithm function are used in the shortest distance approximation function to

57、determine theclustering mechanism, which results in generating small number of clusters.2.3. Data Allocation (Distribution)Data allocation describes the way of distributing the database fragments among the clusters and their respective sites in distributed database systems. This pro s addresses the

58、assignment of network node(s) to each fragment 8. However, finding an optimal data allocation isplete problem 12. Distributing data fragments among database web sites improvesdatabase system performance by minimizing the daransferred and acsed during execution,reducing the storage overhead, and incr

59、easing availability and reliability where multiple copies ofthe same data are allocated.Many data allocation algorithms are describedhe literature. The efficiency of these algorithmsis measurederm of response time. Authors in 19 proed an approachndles the fullreplication of data allocation in databa

60、se systems.his approach, a database file is fully copiedto all participating nodes through the master node. This approach distributes the sequenthrough fragments wiround-robin strategy for sequence input set already ordered by size,where the number of sequenis about the same and number of characters