data miniaturization- implications for mobile devices - john archbold, windspring inc., and robert...
TRANSCRIPT
© copyright 2004 by OSGi Alliance All rights reserved.
Data MiniaturizationFor Telematics and Mobile Devices
Store More …. Send Faster
© copyright 2004 by OSGi Alliance. All rights reserved.
Agenda
• Telematics delivery stages and performance challenges
• Introduction to Data Miniaturization (DM)
• Mobile data challenges & device solutions
• DM/OSGi integration opportunity
© copyright 2004 by OSGi Alliance. All rights reserved.
Telematics Information Flow
1) Data collection
2) Message generation
3) Validation
4) Communication
5) In-car processing
Road Authority
Traffic Police
Auto Club informants
Private data collection
Private floating car
data
General TTI Report
Individual Service TTI
Broadcast Two-way link
Mobile telephone service, cell broadcast
Private TTI data computer
Traffic Information Center
Broadcast Editorial Unit
Broadcasting TTI
RADIO
Private TTI service
Route guidance
NAVIGATION
Dynamic Route Guidance
Road Authority
Road Authority
Traffic PoliceTraffic Police
Auto Club informantsAuto Club
informantsPrivate data
collectionPrivate data
collection
Private floating car
data
Private floating car
data
General TTI Report
General TTI Report
Individual Service TTIIndividual Service TTI
BroadcastBroadcast Two-way link
Two-way link
Mobile telephone service, cell broadcast
Private TTI data computer
Traffic Information CenterTraffic Information Center
Broadcast Editorial UnitBroadcast Editorial Unit
Broadcasting TTI
RADIORADIO
Private TTI service
Route guidance
NAVIGATIONNAVIGATION
Dynamic Route GuidanceInformation Flow
© copyright 2004 by OSGi Alliance. All rights reserved.
Telematics Systems Challenges
1) Data collection: Various source formatsMultiple data sourcesUnknown future sources
2) Message generation: Need to translate source formats to standards and tokens
3) Validation: Complexity of validating multiple formats
4) Communication: Speed and cost of transmissionError rates
5) In-car processing: Real Time processing and responseMultiple communication sources
© copyright 2004 by OSGi Alliance. All rights reserved.
Telematics Data Repetition
• Telematics data contain sequences that are repeated over time– Within individual streams/files
– Across streams/files from common data sources
– Across different data sources
• Current approaches for addressing data repetition– Compression: discovery of common sequences within
individual streams/files only
– Tokenization: replacement of longer (standard) sequences with shorter sequences
© copyright 2004 by OSGi Alliance. All rights reserved.
• Data Miniaturization– Original data transformed into Miniaturized Index File
coupled with a Sequence Dictionary
– Automated discovery, capture and use of repeated data sequences
– Loss-less compression process
• Full manipulation of data in the miniaturized form– Data may be permanently Miniaturized
– High-speed seek, search, edit, and display of any data element
• Significant reduction in data footprint– 30-90% reduction in data size
• Improves performance across the system– Reduced error susceptibility
– Reduced transmission and storage costs
– Faster response time
The DM AlternativeThe DM Alternative
© copyright 2004 by OSGi Alliance. All rights reserved.
DM Process Flow
• DM codec analyzes the original data
• Sequence Dictionary is created– Dictionary organized according to the
frequency of recurring data sequencesin the sample file or stream
– Sequence Dictionary able to be pre-loaded into vehicle
• Encodes original data into a Miniaturized Index File (MIF)or Stream (MIS)
• MIF/MIS and Sequence Dictionary operate together to perform high-speed search, edit and display of the encoded data
Original Data
Sequence Dictionary
DMT Codec MIF
© copyright 2004 by OSGi Alliance. All rights reserved.
DM Technology DemoMapping & Text
Technology Demonstration
© copyright 2004 by OSGi Alliance. All rights reserved.
DM Technology DemoSequence Dictionary
Technology Demonstration
© copyright 2004 by OSGi Alliance. All rights reserved.
Mobile Data Challenges
• Data latency is impacting service adoption, usage and subscription prices
• Application and storage space are constrained
• Network costs are accelerating
© copyright 2004 by OSGi Alliance. All rights reserved.
DM Application DemoMobile Devices
Technology Demonstration
© copyright 2004 by OSGi Alliance. All rights reserved.
DM Application DemoMultiple Data Formats
Technology Demonstration
© copyright 2004 by OSGi Alliance. All rights reserved.
OSGi Telematics/Mobile BenefitsStore More …. Send Faster
• GPS Navigation– Increases device map storage capacity
– Increases map delivery speed
– Reduces communication error rates
• Server/device management– Increases XML/HTML storage capacity on space-constrained devices
– Reduces XML/HTML synchronization time
• Remote Computing– Increases mobile database storage capacity
– Reduces synchronisation time
– Increases back-office data storage capacity
– Increases data search and query speed
• Search– Increases XML/HTML/text storage capacity
– Increases search speed
© copyright 2004 by OSGi Alliance. All rights reserved.
Potential DM/OSGi Integration
• Multiple applications– Communications (e.g. XML, BMP)
– Database (e.g. SQL)
– Search
• Multiple platforms– J2ME (javax.microedition.io)
– Windows, Windows CE
– Linux
– BREW
© copyright 2004 by OSGi Alliance. All rights reserved.
DM Performance Perspective
• Miniaturization performance–Driven by data repetition, type and size
• DM excels with repetitive data types–XML / HTML, BMP, Text Databases, Text, Multi-GIF images
• Performance scales with size–Multiple files within the same data type
–Large files (single and multiple large files)
• DM has limitations with:–Executables, binary files, small single file, single GIF images
–Lossy applications (e.g. JPEG, MPEG, MP3)
© copyright 2004 by OSGi Alliance. All rights reserved.
DM/OSGi Issues Summary
• Telematics–Multiple data formats and sources
–Near real-time requirements for data validation, transmission and processing
• Mobile Devices–Data latency
–Network transmission costs
–Storage constraints on mobile devices
© copyright 2004 by OSGi Alliance. All rights reserved.
DM/OSGi Benefits
• Data Miniaturization addresses both Telematic and Mobile Device storage and performance issues–Data may be permanently Miniaturized–Full manipulation of data in the Miniaturized form–Patent filed in 1999
• Key benefits– Increases storage on new and legacy devices– Increases effective transmission speed–Reduces error exposure–Enables standardized output formats, leading to reduced
complexity
© copyright 2004 by OSGi Alliance. All rights reserved.
Back Up Slides
© copyright 2004 by OSGi Alliance. All rights reserved.
Internet Gateways
HomeIndividual
Government Service ProvidersOffice
CellPhoneWiFi DSRC
Radio
NAVIGATIONGATEWAY AUDIO
AIRCON
INSTRUMENTS
INTERIOR LOCKING
ENGINE ABS POWERTRAIN
ONBOARD SERVER CARD READERH.I.D.
--Information LAN --
--Body LAN --
--PowerTrain LAN --
C
CAMERA
Basic FunctionsHands FreeEmergencyCar eCommerceCar Probe Info
Telematics Data User
© copyright 2004 by OSGi Alliance. All rights reserved.
Data Transform Process
• Maintains an exact one-to-one relationship between data in the original file to data in the MIF file
• Original database is no longer required after the miniaturization process
4MBOriginal
File
1MBDMTFile
OriginalData
MIFData
Data Location: 0.4 MB
Data Location: 1.6 MB
Data Miniaturization
Dictionary options:
• Internal
• External
• Both
SD
SD
SD
SD
© copyright 2004 by OSGi Alliance. All rights reserved.
WindSpring, Inc.
Leadership Team
• Robert F. Mitro, Chairman• Mark Arman, President and CEO• John Archbold, CTO and Inventor• Par Sheth, COO and VP Eng• Alan Knitowski, Director• Steve Liebeskind, Director• John Thomas, Director
• Founded in Queensland, Australia in 1996• Migrated to California in December 2003• Invented Data Miniaturization Technology – DMT• Filed Patent in 1999, 10 Countries/Regions
Background
Data Miniaturization Technology
• DMT miniaturizes large databases for high speed search, retrieval and data transfer
• Enables data search, edit and display in theminiaturized state
• Works without changing existing data formats• Prototype in 2003, FCS in March 2004