Download presentation
Presentation is loading. Please wait.
Published byMaria Lambert Modified over 9 years ago
1
Spring 20021 Efficient Dissemination of Enterprise Summary Data to Mobile Clients Mohamed A. Sharaf University of Pittsburgh
2
2 Motivation “…Currently handheld and palmtop computers are widely used for personal information management. In the near future they will also be used to access enterprise data…”, IBM Corp., 2000 “…For organizations to be successful in today's fast-paced digital economy, decision makers require access to all business-critical information on any platform. Wireless devices are quickly becoming alternative platforms for e-enabling the enterprise, as they provide instant access to relevant enterprise data for Mobile Decision Making…”, Hummingbird Communications Ltd., 2000
3
3 Outline OLAP (On-Line Analytical Processing) Data Model Wireless OLAP Model Scheduling Algorithms Simulation Results Conclusion
4
4 Multi-Dimensional Model [Codd93] Product TV VCR PC Date 1Qtr 2Qtr 3Qtr 4Qtr Country U.S.A Canada Mexico Group-By (P,C,D), Sum(Sales) Dimensions Measures
5
5 A Sample Data Cube Product TV VCR PC Date 1Qtr 2Qtr 3Qtr 4Qtr Country U.S.A Canada Mexico G(P,C) G(P) Derivation Dependency
6
6 Traditional OLAP Server Point to Point Access
7
7 Wireless OLAP Server Broadcast Uplink Channel Power Consumption
8
8 Wireless Environments Asymmetry in the communication Broadcast for data dissemination Periodic (push-based) On-Demand (pull-based) Hybrid A broadcast schedule determines what and when to broadcast Metrics Access Time = Wait + Tune Power Consumption = Active + Doze
9
9 On-Demand Scheduling Algorithms First-Come First-Serve (FCFS) Shortest Service Time First (SSTF) RxW: broadcast a page either because it is popular or because it has at least one long-outstanding request [Franklin 99] Most Request First (MRF) [Ammar 86] Summary Tables : 1) Heterogeneous 2) Skewed Access 3) Derivation Dependency
10
10 Broadcast Organization Header Packet = Identifier + Pointer A table T X is characterized by set of dimensional attributes X. T X subsumes T Y, iff Y X, similarly, T Y is dependent on T X X is the dimensionality degree 100G(Supp)111G(Supp, Prod, Cust)… TuneWaitTune Target Table Header Table
11
11 RxW Variants Strict RxW/S: For each request Q X for a summary table T X, the server maintains the following values: R: The number of requests for T X. W: The age of the first request has for table T X. S: The size of table T X. Table with highest RxW/S is the one to broadcast. Flexible RxW/S: Decision is same as RxW/S, but using “Derivation Dependency” allows: Server to remove dependent tables from queue Client to tune to the first subsuming table
12
12 Controlling the Flexibility Why ? Compromise between access time and power consumption How ? Integrate derivation dependency with scheduling decision Classify dependents into beneficial and impairing according to dimensionality
13
13 d3d3 d1d1 d2d2 d4d4 d 1,d 2 d 1,d 3 d 1,d 2,d 3 d 1,d 2,d 4 Benefit(B) Impairment(I) d3d3 d1d1 d2d2 d4d4 d 1,d 2 d 1,d 3 d 1,d 2,d 3 d 1,d 2,d 4 Benefit(B) Scheduling Intuition d 1,d 2,d 3,d 4,d 5 d5d5 d 4,d 5 d3d3 d1d1 d2d2 d4d4 d 1,d 2 d 1,d 3 d 1,d 2,d 3 d 1,d 2,d 4 d 3,d 4,d 5 d 1,d 2,d 3,d 4 d 2,d 3,d 4,d 5 d 1,d 2,d 3,d 5 QXQX distance = X /2
14
14 Benefit/Impairment Scheduling (BI) The BI for Q i is computed as: The highest BI value request is broadcast next and dependents B are removed A priority queue is used to store requests
15
15 Experiments A synthesized six-dimension lattice. Packet capacity = 10 attribute values Each Mobile host poses 100 queries according to a Zipf distribution Each experiment was run 5 times Metrics: Average Access Time in simulation ticks Average Power consumption in doze units Active power = 20 times doze power Fairness: Standard Deviation of requests’ stretch [Acharya 98] stretch = access time/service time
16
16 Average Access Time
17
17 Power Consumption
18
18 Fairness
19
19 Varying Skewness
20
20 Conclusion We introduced the new problem of scheduling objects with a derivation dependency property We proposed a variety of scheduling algorithms that minimize access time and preserve power consumption LoadAATPC LowBIRxW/S MedBIRxW/S HighFlex. RxW/SBI 65% less than RxW & 55% less than RxW/S 70% less than RxW & 55% less than RxW/S 77% less than RxW 15% less than RxW 20% less than RxW 24% less than RxW
21
21 Future Work We are planning to extend the research to include: Subscribe push environment Caching mechanisms More detailed cost model
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.