ASample of an XML stream is:

Transcription

1 1 Efficiet Multichael i XML Wireless Broadcast Stream Arezoo Khatibi* 1 ad Omid Khatibi 2 1 Faculty of Computer Sciece, Uiversity of Kasha, Kasha, Ira 2 Faculty of Mathematics, Uiversity of Viea,Viea, Austria arxiv: v1 [cs.dc] 14 Ju 2018 Abstract I this paper we recommed the use of multichael for XML data i wireless broadcastig. First we divide XML data ito iformatio uits as bucket, the extract path iformatio (XPath) for ay uit ad build a idex tree from the data path. Fially, make wireless data stream with mergig parts of idex tree ad parts of XML data i multichael XML. The, create a protocol that allows mobile users access to the wireless XML stream geerated with our method. We study 11 chaels i server side ad 3 orthogoal chaels i cliet side. Idex Terms Keywords:XML idexig; Multichael; Wireless broadcastig; Mobile databases; Wireless iformatio systems. I. INTRODUCTION ASample of a XML stream is: < Root >< a1 >< b1 >< c1/ >< c2/ >< c3/ > < /b1 >< b2 >< c4/ >< c5/ >< c6/ >< /b2 > < b3 >< c7/ >< c8/ >< c9/ >< /b3 >< /a1 > < a2 >< b4 >< c10/ >< c11/ >< c12/ >< /b4 > < b5 >< c13/ >< c14/ >< c15/ >< /b5 >< b6 > < c16/ >< c17/ >< c18/ >< /b6 >< /a2 >< a3 > < b7 >< c19/ >< c20/ >< c21/ >< /b7 >< b8 > < c22/ >< c23/ >< c24/ >< /b8 >< b9 >< c25/ > < c26/ >< c27/ >< /b9 >< /a3 >< /root > Ad tree presetatio from this XML stream is i the figure 1. Yo Doh Chug ad Ji Leo Lee make a full study of a idexig for wireless broadcast XML data[1]. We develop it by usig multichael ad multiradio. Data broadcast is widely used because of badwidth limitatios of wireless systems, where the server broadcasts data through a distributio chael ad the the customers liste to the chael [3]. Customer ca retrieve relevat iformatio, furthermore, the cliet does ot eed to sed a request to the server [4]. I order to retrieve iformatio, we have a idex broadcastig direct iformatio of Published data items icluded miutes, secods ad millisecods. Idex i broadcast re-sults i eergy savig. XML data is semi-structured ad accessed with XPath (characteristic of path or XPath query). Previous methods of data broadcastig were adequate for Arezoo Khatibi*, correspod arezoo.khatibi@grad.kashau.ac.ir, Faculty of Computer Sciece, Uiversity of Kasha, BLVD Ghotb Ravadi 6 kilometers, Kasha Ira. flat data as database records accessed with attribute value of primary key ad foreig key [2]. A. (1, x) method II. RELATED WORK (1, m) method satisfied flat data; however there is (1, x) i XML stream. I (1, x) method the idex is represeted as (XPath specificatio, start address, ed address) ad replicated x times durig broadcast. But icreased access time ad wasted badwidth. Therefore, we set partial ad relevat idex iformatio i appropriate locatio i the broadcast stream. We divide data ad idex with usig the cotrol parameter replicatio ito two parts: replicated ad o-replicated. B. defie tp strategy Tp stads for the tree of idex ad path of data. This buildig strategy of the replicatio stream, replicates high level data path ad high level idex tree ad put them before low level data path ad low level idex tree o the broadcast stream. It is suitable because of access time ad tuig time for uique chael i compare of others method as pp, tt, (1, x). III. OUR PROPOSED METHOD We first implemet, the XML tree traversal that we have oted i the itroductio to the Java laguage alog with the output of it. 1. import java.io.file; 2. import java.util.likedlist; 3. import javax.xml.parsers.documetbuilder; 4. import javax.xml.parsers. DocumetBuilderFactory; 5. import org.w3c.dom.documet; 6. import org.w3c.dom.elemet; 7. import org.w3c.dom.node; 8. import org.w3c.dom.nodelist; 9. /** 10. */ 11. public class Mai { 12. public static void mai(strig[] args) { 13. try { 14. DocumetBuilderFactory dbf = DocumetBuilderFactory.ewIstace(); 15. DocumetBuilder db = dbf. ewdocumetbuilder();

2 2 Fig. 1: tree presetatio from first sample of xml stream. 16. File file = ew File("C:\\New folder\\c. xml"); 17. if (file.exists()) { Documet doc = db.parse(file); Node rootnode = doc.getfirstchild(); 18. // System.out.pritl(rootNode.getNodeName ()); 19. // System.out.pritl(rootNode. getfirstchild().getnodename()); 20. // Sys-tem.out.pritl(rootNode. getfirstchild().getnextsiblig(). getnodename()); 21. // Sys-tem.out.pritl(rootNode. getfirstchild().getnextsiblig(). getnextsiblig().getnodename()); // ode.getnodename(); // ode.getfirstchild(); // ode.getnextsiblig(); System.out.pritl(getXML(rootNode)); LikedList<Node> ll = getll(rootnode); for (it i=0;i<ll.size();i++){ Node mynode = ll.get(i); System.out.pritl(myNode.getNodeName() + " idex: " + ll.idexof(mynode)); } System.out.pritl(); // ll.get(idex) // ll.idexof(ode) 22. } 23. } catch (Exceptio e) { 24. System.out.pritl(e); 25. } 26. } 27. public static Strig getxml(node root){ 28. if (root.getchildnodes().getlegth() == 0) 29. retur "<"+root.getnodename()+"/>"; 30. Strig a = ""; 31. a += "<"+root.getnodename()+">"; 32. for (it i=0; i<root.getchildnodes(). getlegth();i++){ 33. a += getxml(root.getchildnodes().item(i)); 34. } 35. a += "</"+root.getnodename()+">"; 36. retur a; 37. } 38. public static LikedList<Node> getll(node root){ 39. LikedList<Node> ll = ew LikedList<Node >(); 40. ll.add(root); 41. Node thisnode = root.getfirstchild(); 42. while (!thisnode.equals(root)){ 43. ll.add(thisnode); 44. if (thisnode.haschildnodes()){ thisnode = thisnode.getfirstchild(); 45. } else if (thisnode.getnextsiblig()!= ull){ thisnode = thisnode.getnextsiblig(); 46. } else { while (!thisnode.equals(root) i. && thisnode.getnextsiblig() == ull) thisnode = thisnode.getparetnode(); if (thisnode!= root) thisnode = thisnode.getnextsiblig(); 47. } 48. } 49. retur ll; 50. } 51. } < Root >< a1 >< b1 >< c1/ >< c2/ >< c3/ >< /b1 >< b2 >< c4/ >< c5/ >< c6/ >< /b2 >< b3 >< c7/ >< c8/ >< c9/ >< /b3 >< /a1 >< a2 >< b4 >< c10/ >< c11/ >< c12/ >< /b4 >< b5 >< c13/ >< c14/ >< c15/ >< /b5 >< b6 >< c16/ >< c17/ >< c18/ >< /b6 >< /a2 >< a3 >< b7 >< c19/ >< c20/ >< c21/ >< /b7 >< b8 >< c22/ >< c23/ >< c24/ >< /b8 >< b9 >< c25/ >< c26/ >< c27/ >< /b9 >< /a3 >< /Root > Root idex: 0 a1 idex: 1; b1 idex: 2; c1 idex: 3; c2 idex: 4; c3 idex: 5 ;b2 idex: 6; c4 idex: 7; c5 idex: 8 ; c6 idex: 9 ; b3 idex: 10; c7 idex: 11; c8 idex: 12; c9 idex: 13 ; a2 idex: 14 ; b4 idex: 15; c10 idex: 16 ; c11 idex: 17; c12 idex: 18; b5 idex: 19 ;c13 idex: 20; c14 idex: 21; c15 idex: 22; b6 idex: 23; c16 idex: 24; c17 idex: 25 ; c18 idex: 26; a3 idex: 27; b7 idex: 28; c19 idex: 29; c20 idex: 30; c21 idex: 31 ;b8 idex: 32; c22 idex: 33; c23 idex: 34; c24 idex: 35; b9 idex: 36; c25 idex: 37; c26 idex: 38 ; c27 idex: 39. HI (high level idex fragmetatio): The sequece of idex odes i the stream that are lo-cated o the upper level of the idex tree meas. That sub trees existed with a root ad h height related to top triagle of XML idex but a XML idex tree cosists oly oe: HI. HI referred as the replicated idex, i order to our all HI have same cotets. LI (low level idex fragmet): The fractio of the wireless stream where idex odes stream exist betwee h+1 to H. We poit LI as o-replicatio because cotet of the L1 to L9 are differet. HD (high level data fragmet): The sequece of data odes i the stream located i the same path from root to the data ode i h level. Followig a strig of (HI+LI+HD+LD) called as a segmet. Wireless stream i the sigle chael is: Istead we use oe chael where we apply three chaels.put three segmets 1, 4, 7 o chael 1 ad put three segmets 3, 6, 9 o chael 2 ad we locate three

3 3 Fig. 2: Sequece d-lik, c-lik, s-lik, h-lik are data ode address, the most left child ode, the ext siblig ode ad the relevat homolog ode. Ch Ch Ch TABLE I: Show the segmets i each chael of multichael. segmets 3, 6, 9 o chael 3 ad as we kow each segmet cosisted (HI+LI+HD+LD); the, i Geeral, we ca apply at most 3 orthogoal chaels i cliet side ad at most 11 chaels i server side. IV. PROTOCOL We use protocol i cliet for receivig iformatio. Each mobile customer ca liste to oly three chaels at a momet. Mark Guide LD The total size of a LD HD The total size of a HD LI The total size of a LI HI The total size of a HI INDEX The size of total idex i a broadcast sream DATA The amout of data i a broadcast stream f(i) The size of a idex ode i a level i g(i) The size of the data ode i a level i X The umber of idex replicatio i a (1,X) method The fa out of a XML data ad idex trees H The height of a XML data ad idex trees h The replicatio level w Positive costat k Positive costat TABLE II: Show the symbols guide usig i fourmula. V. ANALYSIS At first we wat to utilize access time with our proposed method i TP strategy. I TP strategy HI replicate i form sub tree; therefore, average access time cosidered as follows: Where equals to the legth of a high level i a XML tree i oe chael.we cosider a full tree with H height ad h referred as replicatio level ad each ode is i oe bucket ad do ot cosider dowload time after fidig data. We assume value of idex ode i i level as liear fuctio of f from i ad k is a positive umber, i our proposed method f(i)=k, we assume data ode value i i level as g(i). A. For the average access time calculatig by usig TP strategy i multichael case AvgAT T P = ( 1 2 ( 3 )h + 1)k[ ( 3 )h ( 3 )h H 1 1 ] ( 3 )h + 1)w[h + ( 3 )h H 1 1 ], 3 (V.1) Fig. 3: I Shows access time of tp strategy i sigle chael.

4 4 h H K w AvgAT T P Sigle chael= Multichael=80784 TABLE III: Get value of the access time used tp strategy with specified values i multichael. Fig. 6: Shows of idex size with tp strategy i sigle chael. Fig. 4: Shows access time of tp strategy i multichael. B. For the idex size calculatig by usig TP strategy i multichael case INDEX T P = ( 3 )h ( HI T P + LI T P ) ) = k( 3 )h [ ( 3 )h ( 3 )h H 1 1 ] 3 3 (V.2) h H K INDEX T P Sigle chael= Multichael=1056 TABLE IV: get value of the idex used tp strategy with specified values i sigle chael ad multichael. Fig. 7: Shows of idex size with tp strategy i multichael. Fig. 5: Compariso access time i sigle chael ad multichael. Fig. 8: Compariso idex size of tp strategy betwee sigle chael ad multichael.

5 5 C. For the average tuig time calculatig by usig TP strategy i multichael case AvgT T T P = 1 + HI T P + LI T P = 1 + k[ ( 3 )k ( 3 )h H 1 1 ] 3 3 (V.3) h H K AvgT T T P Sigle chael= Multichael=67 TABLE V: get value of the tuig time used tp strategy with specified values i multichael. Fig. 11: compariso tuig time betwee multichael ad sigle chael. D. For the data size calculatig by usig TP strategy i multichael case DAT A T P = ( 3 )h ( HD T P + LD T P ) = ( h h H 3 )k [ g(i) + (g(i) ( 3 )i 1 ] i=1 i=1 (V.4) = w( 3 )k [h + ( 3 )h H 1 ( 3 ) 1 ] (V.5) Fig. 9: tuig time i multichael case. h H w DAT A T P Sigle chael= Multichael=5280 TABLE VI: get value of the data size used tp strategy with specified values i multichael. Fig. 10: tuig time i sigle chael case. Fig. 12: data size of tp strategy i sigle chael.

6 6 Arezoo Khatibi correspod Faculty of Computer Sciece, Uiversity of Kasha, BLVD Ghotb Ravadi 6 kilometers, Kasha, Ira. Her research iterests cover most spectral Iteret of Thig, distributed system, parallel programmig. Khatibi atteded the Nemo Iteret of Thig at Uiversity of Viea (2016). [ ]Arezoo Khatibi Fig. 13: data size of tp strategy i multichael. Fig. 14: compare data size of tp strategy betwee sigle chael ad multichael. VI. CONCLUSION With icreasig umber of chaels, data is broadcast faster i air. I this model, we must have a multiradio or have sigle device with several radio. We observed that tuig time reduced with icreasig umber of chael or reducig umber of odes. I idex size with icreasig chael or decreasig umber of odes, idex size icrease faster ad displayed more mutat. While accessig time i multichael case with icreasig replicatio level we see access time decreases faster. I future work we recommed a method to get rid of trav-erse tree structure i multichael xml wireless broadcast. Nofiacial ad o fud. REFERENCES [1] Yo Doh Chug, Ji Yeo lee, A idexig Method for Wireless broadcast xml data, Iformatio Scieces, Vol. 177, (2007), [2] Wai Ge Yee, Studet Member, Efficiet Data Allocatio Over Multiple Chaels at Broadcast Servers, IEEE Trasactios o Computers, Vol. 51,o 10, October (2002), [3] T. Imieliski, S. Viswaatha, ad B.R. Badriath, Data o Air: Orgaizatio ad Access, IEEE Trasactios o Kowledge ad Data Egieerig, Vol. 9, o 3, May/Jue (1997), [4] C C. Lee, Y. Leu, Data o Air: Orgaizatio ad Access, Iformatio Scieces, Vol. 171, (2005),