Using Semantic Caching to Manage Location Dependent Data

Using Semantic Caching to Manage Location Dependent Data in Mobile Computing Qun Ren, Margaret H. Dunham Presented by Jekkin Shah

Objective • Application of semantic caching to location dependent applications • Eg. Mobile user, navigation system.

Contributions • Mobility model to represent moving objects • Formal definition of location dependent queries • Strategies for query processing • Cache management strategies

Modeling mobility • L = (LX, , LY) location of object in 2 D ( lat-long ) • V = < VX, VY> velocity at any instant of time • Ldt = ( (VX * dt +LX, ) , (VY *dt ) + LY )

Query model • (QR, QA, QP, QL, QC ) QL is the location QC is the result of the query

Query predicate example (price < 100) ^ (LX – 20 < xposition < LX + 20 ) ^ (LY – 20 < xposition < LY + 20 ) • a operator ( LX + c ) • a = attribute , c = constant, LX = location variable

LDD Cache model • Cache consists of LDD semantic segments • Each segment starts with a new page • Additional parameters like Timestamp is also stored

LDD query processing • Query trimming – Probe query – Remainder query Coalesing partial results

LDD Cache Management • FAR ( Farthest Away Replacement ) • Improvisation on existing technique: • Calculating Manhattan distance taking location and direction into account • ( in-direction , out-direction )

Performance study • Performance of semantic caching scheme • Cache replacement strategy • • Workload Design Database design with various parameters Query design ( select only queries ) Moving path design – One-way, return trip , random