Pervasive and Mobile Computing: A 3tier Architecture

1
Pervasive and Mobile Computing A 3-tier
Architecture

• Yanyun lu
• ylu_at_site.uottawa.ca

2
What is Pervasive computing ?

• Pervasive computing is computing power that
  enables software applications available anytime
  and anywhere.
• Personal Digital Assistant (PDA) and cell phone
  are the first widely available and used pervasive
  computing devices.

3
traditional applications VS mobile applications

• The behavior and use of traditional applications
  (desktop) is greatly influenced by the input
  explicitly provided by its users
• dealing with traditional desktop interfaces,
  developers could give users applications that
  took advantage of high bandwidth, large screens
  with extended graphics capabilities, and rich
  input methods.

4
traditional applications VS mobile applications

• mobile users have to interact with applications
  on small screens with limited graphics
  capabilities and minimal bandwidth.
• pervasive applications must provide the user with
  relevant and timely information that enables them
  to act on that information while minimizing the
  navigation and text entry needed.
• pervasive applications must become adaptive to
  the intent of the users and their environment in
  order to make mobile workforce more effective.

5
What is the key to make mobile workforce more
effective?

• context is a key to achieve this goal
• context is the aggregate knowledge of mobile user
  or devices state.
• including location, situation,
  surroundings, and users preferences and activity

6
existing research and industry approaches
addressing the context adaptation problems

• As mobile devices or users move from one location
  to another, pervasive applications must adapt
  themselves to new environments
• Sentient Computing framework , Hewlett-Packards
  Cooltown project , PIMA and so on

7
Disadvantage of those approaches

• These approaches mainly support a specific task
• None of them provide a suitable architectural
  framework that includes the following
  capabilities
• (i) gathering context information from a
  variety of sources, such as corporate system,
  wireless service providers, indoor and outdoor
  location information
• (ii) easily adapting themselves to existing
  IT infrastructure and enterprise applications
• (iii) adapting current software development
  process, therefore providing a generic approach
  to minimize the developers learning curve.

8
A three-tier architecture

• The 3-tier model is designed to better resolve
  the issues that exist in the framework solutions
  described before.
• 3-tier architecture takes a variety of context
  sources (not just location-aware)
• This architecture is made up of three separate
  layers (namely, Collection, Analysis and Action)

9
Some useful background knowledge

• Context-aware Pervasive Computing
• A pervasive computing system that strives to
  be minimally intrusive has to be context-aware.
  It must be cognizant of its users state and
  surroundings, and must modify its behavior based
  on this information.
• A key challenge is obtaining the information
  needed to function in a context-aware manner .
• Examples of such information including
  position, orientation, the identities of people
  nearby, locally observable objects and actions,
  and emotional and physiological state.

10
Some useful background knowledge

• Context Model
• Context model is designed based on the
  activities identified by the business analyst.
• Context model designed and created in
  conjunction with the data model and uses similar
  terminology and modeling techniques to minimize
  the learning curve.

11
Some useful background knowledge

• Three main components of the context model
• Entities
• (1) similar to database entities they could
  be a person, place, or thing.
• (2) the difference is they are considered
  relevant to the behavior of the application.
• (3) have unique keys and can also have
  properties associated with them, which store more
  static information which is needed to make a
  state determination.
• (4) Entities also have a lease or lifespan
  within the context server. If the entity is no
  longer relevant it should no longer exist in the
  context server.

12
Some useful background knowledge

• 2. Relationships
• (1) Relationships link entities together in
  the context server.
• (2)one principle link between two entities and
  it is a directed link (e.g., Employee ATTENDS
  Meeting).

13
Some useful background knowledge

• 3. States
• the most critical part of the context model and
  serve as the differentiator between a data model
  and a context model
• Their role is to capture an entity or
  relationships current circumstance or situation.
  

14
Some useful background knowledge

• The context designer must make some decisions and
  tradeoffs in determining what information to
  model.
• Everything can be modeled as an entity in the
  context model, but there must be a state change
  affecting that entity, otherwise it has no value
  in the context model.
• A good rule for modeling is to design the model
  you believe will represent the context of the
  scenario.

15
THREE-TIER ARCHITECTURE MODEL



16
THREE-TIER ARCHITECTURE MODEL

• As shown in Figure 1 , three layers include
• (i) Collection layer
• identifies and collects the users context
  and environment data by deploying sensors to
  connect to the data from enterprise applications,
  environmental sources, or other content
  providers.
• (ii) Analysis layer
• manages and interprets context data and
  references it at any time and updates it
  accordingly with minimal involvement from the
  user. Interpreters can be deployed within the
  context server to mediate conflicting context
  information, abstract higherlevel context states
  or predict context state information based on
  history, preferences, and intent.
• (iii) Action layer
• takes the information managed by the
  context server and uses it to change the behavior
  of an application, such as presentation,
  navigation, and application logic. The platform
  provides the capabilities to send actionable
  alerts to multiple devices, alter the user
  interface, or make requests and queries based on
  context state information and/or events.

17
Collection Tier - How to Get Context Data?

• The collection tier is the most critical tier in
  the three-tier approach. Its role is to gather
  context data from a variety of sources and
  implement the Context Model components based on
  that data.
• In building this tier, the developer must
  consider that context data comes from static,
  dynamic, environment, and predictive sources
  (show in Table 1 next page)

18
Collection Tier - How to Get Context Data?
19
Gathering context related information

• The collection layer focuses on two main
  components to gather context related information

20
Sensors and Adapters in collection tier

• Adapters are not unique to context servers they
  are conduits to existing enterprise and
  thirdparty data sources that gather context
  information.
• Ideally, they are the same adapters used in
  existing web and client applications.
• Rather than build our own adapter layer, we will
  use existing adapter frameworks that
  organizations have already spent a considerable
  amount of time and effort to implement.
• The sensor is the critical component in the
  collection tier. The sensor gathers data from the
  adapter and creates, updates, or deletes
  information in the context server.

21
Analysis Tier

• The analysis tier provides the framework to
  manage context entities, relationships, and
  states within the context server.
• Interpreters are components within the analysis
  tier that focus on reading context state
  information, applying application or business
  logic, and setting a higher level or derived
  context state.

22
Action Tier

• Based on these context states, specific actions
  are triggered in response.
• The action tier affects the behavior of an
  application based on the previously preformed
  analysis
• Our goal in achieving this is to focus on three
  action areas presentation, navigation, and
  application logic.

23
Action Tier

• Once the context server has determined that state
  change has occurred it can act on that state in
  different ways
• (i) Presentation Actions that cause the
  content or format of the application to change as
  the person is using it.
• (ii) Navigation Actions that minimize the
  needed navigation path a user has to perform
  based on a proactive context event.
• (iii) Application Logic Enables the developer
  to alter the behavior and rules for accessing and
  using specified applications and services.

24
Action Tier

• Presentation and Navigation affects the
  applications user interface based on context
  data. Sample actions include
• When the user is driving, change the presentation
  mode from text to voice.
• Display menu options that are only relevant to
  the users current situation and intent.
• The system indicates the best way for salesperson
  A to contact salesperson B (cellphone,
  email, landline phone of meeting place, etc.).

25
CONCLUSION

• A three-tier architecture divides the process of
  building pervasive applications into three
  stages Collection, Analysis, and Action.
• In this building process, collecting contextual
  information is the most critical part because of
  the variety of the context sources as well as the
  complexity of the human behavior.
• Despite the context collection challenges,
  three-tier approach enables developers to address
  a number of mobility scenarios, reduces
  developers learning curve by adapting normal
  software development process to build
  context-aware applications that focus on the
  users activity
• Using three-tier approach enables pervasive
  applications to easily adapt to existing IT
  infrastructure and enterprise applications

26
Questions?
27
Reference

• Group 1
• 1 Michel Barbeau, Mobile, Distributed, and
  Pervasive Computing, Handbook of
• Wireless Networks and mobile computing
  (Ivan Stojmenovic, edit), Wiley, Jan. 2002
• 2 G.D. Abowd, Software Engineering Issues for
  Ubiquitous Computing, in
• proceedings of the 1999 International
  Conference on Software Engineering,
• Apr. 1999
• 3 Yuval Boger, Context Awareness and Mobile
  Computing, Pocket PC Summit,
• May 2002
• 4 M. Satyanarayanan, Pervasive Computing
  Vision and Challenges, IEEE Personal
  Communications, August 2001
• 5 Karen Henricksen, Jadwiga Indulska and Andry
  Rakotonirainy, Infrastructure for
• Pervasive Computing Challenges, 2001

28
Reference

• Group 2
• 6 Jerey Hightower and Gaetano Borriello,
  Location Systems for Ubiquitous
• Computing, Aug. 2001
• 7 Guanling Chen and David Kotz, A Survey of
  Context-Aware Mobile Computing
• Research, Dartmouth Computer Science
  Technical Report TR2000-381, 2000
• 8 M. Weiser, Some Computer Science Issues in
  Ubiquitous Computing, July 1993
• 9 Guruduth Banavar, James Beck, Eugene
  Gluzberg, Jonathan Munson, Jeremy
• Sussman, Deborra Zukowski, Challenges an
  application model for pervasive
• computing, 6th Proc annual Intl. Conference
  on Mobile Computing and Networking
• MOBICOM 2000, August 2000
• 10 Want, Z. and Garlan, D., Task-Driven
  Computing, Technical Report, CMU-CS-
• 00-154, School of Computer Science,
  Carnegie Mellon University, May 2000.
• 11 Harter, A. et al., The Anatomy of a
  Context-Aware Application, Proceedings 5th
• Annual Intl. Conference on Mobile
  Computing and Networking (MobiCom99),
• August 1999.

29
Reference

• Group 3
• 12 Kindberg, T. et al., People, Places,
  Things Web Presence for the Real World,
• http//www.cooltown.hpl.hp.com, Hewlett-
  Packard Labs Technical Report HPL-
• 2000-16, 2000.
• 13 Blair, G,. Blair, L., Issarny, V,. Tuma, P,.
  Zarras, A,. The Role of Software
• Architecture in Constraining Adaptation in
  Component-Based Middleware
• Platforms, Middleware 2000 Proc LNCS 1795
  IFIP/ACM NY, USA, April 2000
• 14 Esler, M. et al., Next Century Challenges
  Data-Centric Networking for Invisible
• Computing, Proceedings 5th Annual Intl.
  Conference on Mobile Computing and
• Networking (MobiCom99), August 1999.
• 15 N.M. Belaramani, Y. Chow, V.W. Kwan, A
  Component-based Software
• Architecture - For Pervasive Computing,
  2001
• 16 Thomas Phan, Lloyd Huang, Chris Dulan,
  Challenge Integrating Mobile Wireless
• Devices Into the Computational Grid,
  http//pcl.cs.ucla.edu/pub/papers/mobicom02,
• 2002

30
Proposed questions

• What is Pervasive computing ?

31
Proposed solution

• Pervasive computing is computing power that
  enables software applications available anytime
  and anywhere.

32
Proposed questions

• What are the three layers of the THREE-TIER
  ARCHITECTURE MODEL ?

33
Proposed solution

• Collection, Analysis and Action

34
Proposed questions

• What is the key to make mobile workforce more
  effective?

35
Proposed solution

• context is the key
• Context is the aggregate knowledge of mobile
  user or devices state including location,
  situation, surroundings, and users preferences
  and activity.

36
Thank you