Presentation is loading. Please wait.

Presentation is loading. Please wait.

June 10, 2007, IDAR’07, Beijing, China 1 Ensuring Consistent Termination of Composite Web Services An Liu 1,2,3 and Qing Li 2,3 1 Department of Computer.

Similar presentations


Presentation on theme: "June 10, 2007, IDAR’07, Beijing, China 1 Ensuring Consistent Termination of Composite Web Services An Liu 1,2,3 and Qing Li 2,3 1 Department of Computer."— Presentation transcript:

1 June 10, 2007, IDAR’07, Beijing, China 1 Ensuring Consistent Termination of Composite Web Services An Liu 1,2,3 and Qing Li 2,3 1 Department of Computer Science and Technology University of Science & Technology of China (USTC), Hefei, China 2 Joint Research Lab of Excellence CityU-USTC Advanced Research Institute, Suzhou, China 3 Department of Computer Science City University of Hong Kong (CityU), Hong Kong, China liuan@ustc.edu itqli@cityu.edu.hk

2 June 10, 2007, IDAR’07, Beijing, China 2 Outline Problem Statement and Background Problem Modeling and Methodology Framework Conclusion and Future Work

3 June 10, 2007, IDAR’07, Beijing, China 3 Web Services Composition Client App Travel Agent Booking Request Hotel Res. OK Timeline Web Services FlightTaxiHotel Flight Request Hotel Request Taxi Request Flight Res. OK Taxi Res. OK Booking Response Hotel Cancellation Taxi Cancellation Reliable Composition Needs Transactional Support !

4 June 10, 2007, IDAR’07, Beijing, China 4 ACID Transaction Insufficient Composite web services  Last for hours or days  Work in an open environment  Cross administrative boundaries ACID transaction is insufficient  Trust and timeliness no longer apply Compensation-based transaction No common transaction semantics

5 June 10, 2007, IDAR’07, Beijing, China 5 Some Representative Related Work Transaction semantics declaration  Mikalsen et al. (2002): declare transaction semantics at operation level and transactional property at service level Transactional property (TP) deduction  Vidyasankar and Vossen (2004): deduct TP of composite services from TP of component services Different Atomicity Requirements  Fauvet et al. (2005): support flexible atomicity through TP-based service selection  Bhiri et al. (2005): verify Accepted-Termination-States (ATS) based atomicity by some transactional rules  Montagut and Molva (2006): support ATS-based atomicity through TP-based service selection and automatic coordination

6 June 10, 2007, IDAR’07, Beijing, China 6 State of the Art Current research has  Defined TP of web services  Guaranteed Reliable Service Composition in the context of heterogeneous TP Autonomy of services was overlooked Compensation has temporal and cost constraints (Benatallah et al. 2004)  e.g., deadline of hotel reservation cancellation

7 June 10, 2007, IDAR’07, Beijing, China 7 Problem Statement How to get reliable composition when both autonomy and heterogeneity are considered? Given  The control flow of a composite service  Component services with heterogeneous TP  Temporal constraints of compensation How to get reliable composition where the composite service will terminate in a consistent state?

8 June 10, 2007, IDAR’07, Beijing, China 8 Outline Problem Statement and Background Problem Modeling and Methodology Framework Conclusion and Future Work

9 June 10, 2007, IDAR’07, Beijing, China 9 Definitions Transactional Property (TP) of web services  Compensatable (c): normal operation (o n ) and compensation operation (o c )  Retriable (r): Only normal operation  Pivot (p): Only normal operation Temporal property of operations  Execution time (t x ): e.g., 5 seconds  Available time (t a ): e.g., [9am, 5pm] Deadline: Maximal Compensatable Time (MCT) For compensation op, t a =[t, t+MCT]

10 June 10, 2007, IDAR’07, Beijing, China 10 Definitions (cont.) Service s: 5-tuple(tp, o n, o c, t, sta)  tp  {r, c, rc, p}  o n & o c : normal and compensation operation  t = (t s, t e ): starting and ending time  sta  {initial, active, completed, aborted, failed, cancelled, compensated}: current state Process: G(V t, V o, E)  V o : set of control flow op. (cfo), where cfo  {AND-split, AND-join, XOR-split, XOR-join, OR-split, OR-join, start-loop, end-loop} G(V s, V o, E): an instance of process G(V t, V o, E)

11 June 10, 2007, IDAR’07, Beijing, China 11 Consistent Termination Condition A composite service s c will always terminate in a consistent state if the consistent termination condition (CTC) holds: v i.tp  {c, rc}, s c.t.t e  v i.o c.t a for v i  V s in G Example:  s c, s 1  s 2  if s 1.o n.t x =2 & s 1.o c.MCT=5 & s 2.o n.t x =4  then s 1.o c.t a =[2,7], s c.t.t e =6

12 June 10, 2007, IDAR’07, Beijing, China 12 CTC Verification Precondition: global view of a composite service: control flow, transactional and temporal property Method: simple calculation Problem: NOT easy to get a global view  Dynamic selection  Multi-level composition

13 June 10, 2007, IDAR’07, Beijing, China 13 Methodology Assumptions  A1: only elementary component services  A2: static service selection  A3: only compensatable component services  A4: only AND-split and AND-join operators

14 June 10, 2007, IDAR’07, Beijing, China 14 Outline Problem Statement and Background Problem Modeling and Methodology Framework Conclusion and Future Work

15 June 10, 2007, IDAR’07, Beijing, China 15 Framework Overview

16 June 10, 2007, IDAR’07, Beijing, China 16 CTC Verification Examples Table1ABCD txtx 2846 MCT1520812 A B C D T2ABCD tsts 02210 tete 2 616 tata [2,17][10,30][6,14][16,28] T3ABCD tsts 024 tete 28 tata [2,17] T4ABCD tsts 026 tete 210 tata [10,30] T5ABCD tsts 025 tete 2 tata [2,17] [10,30] 10 [11,31] 11 10 [16,28] 10 16 10 [8,16] [11,31] 17 11 [8,16] [10,18] 12 [12,20] 12 16 [16,28] 18 [18,30] [2,17] 16 [16,28] 9 [9,17][10,30] 1011 [11,31] [11,19] 11 17 [17,29]

17 June 10, 2007, IDAR’07, Beijing, China 17 Scheduling Algorithms Motivation: different scheduling algorithms lead to different results of CTC verification, which decide whether to abort the service or not Objective: design a scheduling algorithm which can decrease the possibility of service abortion Some algorithms  Immediate scheduling  Minimal delay scheduling  Maximal delay scheduling  Average delay scheduling  …, optimal scheduling algorithm?

18 June 10, 2007, IDAR’07, Beijing, China 18 Dynamic Monitoring

19 June 10, 2007, IDAR’07, Beijing, China 19 Inaccurate Global View Problems  Uncertainty of Execution Time (UET)  Information Missing (IM) IM1: execution time of normal operation is unknown IM2: available time of compensation operation is unknown Solution  UET Scheduling algorithms Dynamic monitoring mechanism  IM1: UET variation (increased execution time)  IM2: Dynamic monitoring mechanism

20 June 10, 2007, IDAR’07, Beijing, China 20 Removing Assumptions A4 (only AND-split and AND-join operators)  Loop: UET variation  XOR/OR: pseudo AND operator A3 (only compensatable component services)  Retriable: safely ignored  Pivot: process graph division A2 (static service selection)  IM1 & IM2 A1 (only elementary component services)  IM1 & IM2

21 June 10, 2007, IDAR’07, Beijing, China 21 Time Complexity CTC verification: O(n+e) Dynamic monitoring: O(n 2 +ne) Total time complexity: O(n 2 +ne)

22 June 10, 2007, IDAR’07, Beijing, China 22 Outline Problem Statement and Background Problem Modeling and Methodology Framework Conclusion and Future Work

23 June 10, 2007, IDAR’07, Beijing, China 23 Conclusion and Future Work A framework  Assists selection of services based on transactional and temporal property  Ensures consistent termination under constrained compensation Some next steps  Minimal-Abort-Possibility scheduling algorithms  Cost constraints  Integration with web services execution engine

24 June 10, 2007, IDAR’07, Beijing, China 24


Download ppt "June 10, 2007, IDAR’07, Beijing, China 1 Ensuring Consistent Termination of Composite Web Services An Liu 1,2,3 and Qing Li 2,3 1 Department of Computer."

Similar presentations


Ads by Google