Enabling Location Specific Real-time Mobile Applications Ravi - - PowerPoint PPT Presentation

Enabling Location Specific Real-time Mobile Applications

Ravi Kokku, Karthik Sundaresan, Geoff Jiang NEC Laboratories America, Princeton

1

Trends in the Mobile Networking Domain

Our objective: enable rapid deployment of mobile applications

• Ubiquity of Mobile Devices
• Feature-rich Mobile Devices
• Emerging openness

– Open Services, Open Networks, Open Devices,…

8/26/2008

2

Maps Weather Social n/w

Verizon approved devices Verizon approved services Verizon network

Maps Weather Social n/w

Open devices Open Services and applications Open networks

Verizon ATT Sprint

• F

Cognitive spectrum access

Download assistance

Focus: Location Specific Real-time Applications

8/26/2008

3

Service

C B A

Info A Info A Info B Info B

Mobiles themselves can act as sensors to provide updates to applications in real-time.

Proposed Approach and Challenges

1.

Design and implement the framework

2.

Build a few sample applications

8/26/2008

4

Network provider

Info. delivery

Mobile user devices

Repository

Data mining

Other Apps.

Data svcs

Service composition Existing data

Framework for Location-specific real-time services

Data collection, validation, representation

Incentives, privacy, trust

8/26/2008

5

Related Work

• Vs. P2P file sharing

– Amount of information a user has to upload is inversely proportional to the number of users present at the location

Incentive design more complex

• Vs. Pub-sub systems

– Subscribers are also participating in publishing

incentive design, privacy, trust very important

– Publishers and subscribers per location change dynamically

need highly adaptive info. gathering and delivery mechanisms

Some other related work in the proposal…

6

PARK-ASSI ST: Application Case Study

Internet

Help each other find parking

Credit Management

• Pricing function is the key

– Relative value of an update p(t), – Relative charge per info. delivered q(t), – Max. absolute value of an update: Vm, – Max. absolute charge for info: C

m

• Updates: Credit per update = Vm * p(t)

– t is time since last (set of) useful updates

• Information requests: charge = Cm * q(t)

– t captures info. recentness/confidence

• Limitations (Work in Progress):

– Update credit does not capture # of other users – Users oblivious to variable credit

8/26/2008

7

1

Time

1

Time

1

Time

1

Time

p(t) q(t)

Information Generator Key Manager (Periodic key generate) User credit manager

uid Credits Periodic update

• f Keys

Keys Periodic key request (uid) Broadcast Information (encrypted with keys)

Application Service

(1) (2) (3)

Decrypt information with keys

(4) (6) (5) (1)

I ncentive compatible I nformation Delivery

Summary

Enabling Location-Specific Real-time Mobile Applications

– Ubiquity, resourcefulness, openness

Several interesting challenges to be addressed

– Incentives, trust, information delivery,…

Approach: Develop core set of mechanisms

– enable rapid deployment of such applications

8/26/2008

9