1. Fushen Wang, XinZhou, Carlo Zaniolo Using XML to Build Efficient Transaction- Time Temporal Database Systems on Relational Databases In Time Center, 2005 資工所．莊政道． D95922014

2. Contents Introduction Viewing Relation History in XML Temporal Queries using XQuery The ArchIS System Temporal Clustering and Indexing Performance Study Database History Compression Conclusion 1

3. Introduction The additional complexity of going from standard queries into temporal ones: XML/XQuery < relation tables and SQL The evolution history of a relational database: be viewed naturally using XML be queried effectively using Xquery the temporal data and temporal queries can be supported efficiently: data-compression, clustering, indexing and query- mapping techniques data model ： temporally ungrouped vs. temporally grouped 2

4. Contents Introduction Viewing Relation History in XML Temporal Queries using XQuery The ArchIS System Temporal Clustering and Indexing Performance Study Database History Compression Conclusion 3

5. Viewing Relation History in XML Traditional transaction-time databases (temporally ungrouped) 4

6. Viewing Relation History in XML (cont.) Traditional transaction-time databases (temporally ungrouped) Change an attribute value  add a new history tuple Redundancy information Temporal queries  frequently coalesce tuples Complex and hard to scale in RDBMS Overcome using: Time-stamped history of each attribute is grouped under the attribute 5

7. 6

8. 7

9. Viewing Relation History in XML (cont.) H-documents (or H-views when these are virtual representations) the nested representations hard hard to be represented in flat tables, they can be naturally represented by XML- based hierarchical views. greatly reduces the need for coalescing the effectiveness of expressing complex temporal queries with XQuery 8

10. 9

11. 10

12. Contents Introduction Viewing Relation History in XML Temporal Queries using XQuery The ArchIS System Temporal Clustering and Indexing Performance Study Database History Compression Conclusion 11

13. Temporal Queries using XQuery temporal projection temporal snapshot temporal slicing temporal join temporal aggregate restructuring 12

14. Temporal Queries using Xquery (cont.) Grouped data model: result is already coalesced Ungrouped data model: coalesced on the results 13

15. Temporal Queries using Xquery (cont.) xs: namespace of XML Schema tstart($e): start date tend($d): end date 14

16. Temporal Queries using Xquery (cont.) toverlaps($a, $b):returns true if one node overlaps with another one, and false otherwise telement($a, $b) constructs an element with a and b as its attributes 15

17. Temporal Queries using Xquery (cont.) 16 overlapinterval($a,$b) as Ch.3

18. Temporal Queries using Xquery (cont.) 17

19. Temporal Queries using Xquery (cont.) function restructure takes two lists and returns all the overlapped intervals. 18

20. Temporal Queries using Xquery (cont.) tcontains($a,$b) : to check if one interval covers another. 19

21. Temporal Queries using Xquery (cont.) tequals($d1,$d2) :to check if two nodes have equal intervals 20

22. Temporal Queries using Xquery (cont.) Temporal Functions Restructuring functions coalesce($l): coalesce a list of nodes restructure($a,$b): return all the overlapped intervals on two set of nodes. Interval functions (such as Ch.3) toverlaps($a,$b), tprecedes($a,$b), tcontains($a,$b), tequals($a,$b), tmeets($a,$b) overlapinterval($a,$b) 21

23. Temporal Queries using Xquery (cont.) Temporal Functions Duration and date/time functions timespan($e): the time span of a node tstart($e): the start time of a node tend($e): the end time of a node tinterval($e): the interval of a node telement($Ts, $Te): constructs an empty element telement with attributes tstart and tend rtend($e): recursively replaces all the occurrence of “9999-12-31” with the value of current date externalnow($e): recursively replaces all the occurrence of “9999-12-31” with the string “now”. 22

24. Temporal Queries using Xquery (cont.) Support for ‘now’ Now  current_timestamp, current_date Value(now)  end-of-time (as 9999-12-31) Access with tstart($e), tend($d) Search based on indexes Temporal ordering can be used without any change query-1  output(9999-12-31 for ‘now’)  input of query-2 23

25. Contents Introduction Viewing Relation History in XML Temporal Queries using XQuery The ArchIS System Temporal Clustering and Indexing Performance Study Database History Compression Conclusion 24

26. The ArchIS System underlying database systems: RDBMSs design issues: how to map (shred) the XML views representing the H- documents into tables (which we call Htables) how to translate queries from the XML views to the H- tables, and which indexing, clustering and query mapping techniques should be used for high performance. 25

27. The ArchIS System (cont.) 26

28. The ArchIS System (cont.) 27 H-Tables Table in Relational DB Key table attribute history table Global relation table +tstart, tend

29. The ArchIS System (cont.) Relation in the current database: employee(id, name, salary, title, deptno) id  key attribute The Key Table: employee_id(id, tstart, tend) id  not change Composite keys ex. (supplierno, itemno)  lineitem_id(id,supplierno, itemno, tstart, tend) id : a unique value generated from (supplierno,itemno) 28

30. The ArchIS System (cont.) Attribute History Tables: employee_name(id,name,tstart,tend) employee_deptno(id, deptno, tstart,tend) employee_salary(id, salary, tstart,tend) employee_title(id, title, tstart,tend) Index on ‘id’  efficiently join 29

31. The ArchIS System (cont.) Insert a new tuple: TSTART =current timestamp TEND = now Delete a current tuple: TEND =current timestamp Update a current tuple: Delete + Insert 30

32. The ArchIS System (cont.) Global Relation Table: relations(relationname, tstart, tend) record all the relations history in the database schema, i.e., the time spans covered by the various tables in the database. the root elements of H-documents. 31

33. The ArchIS System (cont.) Query Mapping XQuery on H-views  equivalent SQL/XML expressions on H-tables on H-tables, SQL/XML constructs: XMLElement: return elements XMLAttributes: return attributes XMLAgg: aggregate function, constructs an XML value from a collection of XML value expressions. 32

34. The ArchIS System (cont.) Query: to return an new employees element containing all the employees hired after 02/04/2003 33 SQL/XML query: Output:

35. The ArchIS System (cont.) 34 SQL/XML query:

36. The ArchIS System (cont.) 35 SQL/XML query:

37. The ArchIS System (cont.) mapping(H-views  H-tables): 5 steps Identification of variable range tuple vs. attribute key relation vs. attribute relation Variables(XQuery:for,let)  Varibales(SQL/XML:from) Generation of join conditions Generation of the where conditions Translation of built-in functions Output generation SQL/XML queries often contan many natural joins (N.id=T.id), Joins execute very fast (id attribute is sotred) 36

38. Contents Introduction Viewing Relation History in XML Temporal Queries using XQuery The ArchIS System Temporal Clustering and Indexing Performance Study Database History Compression Conclusion 37

39. Temporal Clustering and Indexing (cont.) Usefulness-Based Clustering Improve snapshot queries Store attribute history in a segment U = U live /U all, where U live is the count of live(or current) tuples and u all is the count of all tuples As U <U min : A new segment S i is allocated; The interval of this segment is recorded in the table segment (segno, segstart, segend) All tuples in SEG live are copied into a new segment S i, sorted by ID; All live tuples in SEG live are copied into a new live segment SEG live’, and the old live segment is dropped. 38

40. Temporal Clustering and Indexing(cont.) 39

41. Temporal Clustering and Indexing (cont.) 40

42. Temporal Clustering and Indexing (cont.) Advantages for segment-based clustering: the current live segment always has a high usefulness, which assures effective updates. records are globally temporally clustered on segments. for snapshot queries, only one segment is used, and for temporal slicing queries, only segments involved are used, thus such queries can be more efficient. 41

43. Temporal Clustering and Indexing (cont.) Storage usage 42

44. Temporal Clustering and Indexing (cont.) Query Mapping with Clustering First find the segment number satisfying the tstart and tend conditions and then search those segments 43 SQL/XML query:

45. Temporal Clustering and Indexing (cont.) unless the number of segments becomes very large and exceeds the number of main memory blocks available for sort-merge joins, joining H-tables remains a very efficient one-pass operation. 44

46. Contents Introduction Viewing Relation History in XML Temporal Queries using XQuery The ArchIS System Temporal Clustering and Indexing Performance Study Database History Compression Conclusion 45

47. Performance Study Query: 46

48. Performance Study (cont.) Query Perofrmance: Segment-based archiving > good no segment-based arching 47

49. Performance Study (cont.) 48

50. Performance Study (cont.) Storate Utiltization 49

51. Performance Study (cont.) 50

52. Contents Introduction Viewing Relation History in XML Temporal Queries using XQuery The ArchIS System Temporal Clustering and Indexing Performance Study Database History Compression Conclusion 51

53. Database History Compression Block-based Compression: BlockZIP compresses the data as block-sized blocks after compression with BlockZIP, the output consists of a set of block-sized compressed blocks concatenated together. uncompressing of the whole file is not needed. BlockZIP facilitates uncompression at the granularity of a block, thus snapshot and temporal slicing queries can be efficient, since only a small number of blocks need to be uncompressed. 52

54. Database History Compression (cont.) 53

55. Database History Compression (cont.) 54

56. Contents Introduction Viewing Relation History in XML Temporal Queries using XQuery The ArchIS System Temporal Clustering and Indexing Performance Study Database History Compression Conclusion 55

57. Conclusion the transaction time histories of relational databases can be stored and queried efficiently by using XML and SQL/XML. the query mapping, indexing, clustering, and compression techniques used to achieve performance levels well above those of a native XML DBMS. The approach of ArchIS can be used to add a transaction-time capability to any existing RDBMS. realization does not require the invention of new techniques, nor costly extensions of existing standards 56

58. THE END !