Laurent Ciarletta

1
Emulating the Future with/of Pervasive Computing
RD

• Laurent Ciarletta
• LORIA ENSMN
• Pervasive 2005, Workshop
• What make for good application-led research
• in ubiquitous computing?

2
Visibility of Ubicomp News and Internet
Browsing on Pervasive Computing Applications

• PDAs, smart phones, WiFi, 3G, RFID
• where/what are pervasive applications?
• Smart fridge, smart cars, robotic hovers/mowers
  
• where are the mainstream offers, where is the
  interoperability?
• Weather /mood connected lamp, (AR) PacMan in the
  streets of Taiwan
• where is the public?
• Etc.

Are we there yet?
Why?
3
Personal Background and Experience Pervasive
Computing Work at NIST

• Application-led RD
• Building test-bed applications
• Collaborating with industry and academics
  (networking, systems, software, embedded systems,
  AI)
• Standardization
• Evaluation metrics, tools
• Reference implementation
• Modeling and simulation

1998-2002
4
Pervasive Computing is User-centric

• User is part of the LPC model
• New metrics
• Usability
• Reliability
• Service vs. Cost
• Various users
• Consumer/client/patient
• Disabled/Senior/Child
• John Doe (is not a designer, nor a researcher!)

5
From Application-led RD to EmulationAroma
Smart Conference Room, EXiST

• Resource consuming
• Money-wise
• Time and energy
• Moderately rewarding
• Demo effect
• Emerging technologies
• Quick prototyping techniques
• Maintenance issue
• Quickly obsolete (although)
• Non usable, lost when gone
• Isolated, cant keep up with the big players
• -gt Emulation/Simulation
• EXiST

1999-2000
6
The 2 Faces of Emulation

• To emulate
• (Computer Science) To imitate the function of
  another system (not necessary at the same speed)
• Effort or ambition to equal or surpass another
• To strive to equal/match or excel, especially
  through imitation

• Existing
• Bricks
• Architectures
• In computer science
• and other scientific fields

Hardware emulators
Virtual Reality Toolkit
Network emulators
Tangible User Interface
Intelligent agents
DIS
Missing parts ?
HLA
7
Real Test-beds vs. Emulation
Real Test-bed Emulation/simulation
Expensive Difficult to set up/maintain Limited scale Quick obsolescence Fixed location Replay difficult Class of applications - oriented Safety and privacy issues for specific applications Reduced cost Quick set-up, prototyping, development Scalability Consistent/coherent replay Safe
Reality check! Its not reality Reality is more complex Modeling is difficult Real-time is even more difficult Validity?
8
Tools for Application-led DevelopmentTools for
the Community
Developing applications shouldnt/doesnt imply
lots of money and hardware and rebuilding almost
everything from scratch

• Development tool (toolkit/commonly agreed on
  architecture)
• Large-scale / distributed
• Mixed-mode
• Integration of real applications, devices,
  architecture with emulation/simulation
• Scenario-based
• Methodology
• Allows
• Test, evaluation, incremental development
• Emulation of creativity, showroom
• Model/classification, for cooperation between
• Research teams
• Communities
• Large scale  open  community test-bed
  Pervasive Computing Bone
• Taskforce / workgroup
• Common tools
• Standards
• Mailing list/ Forum
• Publicize/advertise

What features would you like to see?
9
Thank you