1

Last Time

Context Aware Applications Defining context

Active vs. passive

Survey of some applications No killer application yet; but maybe in the future? Location (and time) most heavily used contexts

Sensing context Location:

• Outside: GPS

• Inside: Need a location traction mechanism

2

Sensing other low level contexts

Time? Physical sensors possible

Light, acceleration, tilt, sound, temperature, pressure, proximity of humans

Nearby objects? If system keeps track of location of objects, query to a

database Bandwidth? Odyssey Orientation?

Orientation sensor based on two mercury switches (Newton MessagePad)

Directional antennas/directional transmitters?

3

High Level contexts

This is by far the more difficult problem Not just an engineering challenge

How does a node know whether this is a party or a wake? Check your calendar or detect the cake/coffin?

4

Sensing Context changes

Publish subscribe model? Monitor polls the current context Notifies subscribers to the context Centralized or on a per-node?

Polling rate is a function of rate of change of context

5

Modeling Context Information

6

Location Model

How do we answer questions such as: Given an object what is its location Given a location, return the set of objects there Determining paths between locations How does a vehicle navigation system do it?

Geographical Information Systems/Spatial Databases Already pretty mature (part of Oracle, mysql, etc..) Model geometry and topology Typically locations are organized hierarchically Sonal Dedhia’s project

7

Security and Privacy Issues

I personally hate the idea that someone can know my location all the time I still refuse to get a cell phone, but it looks like I will finally

cave in! Par for the course, many of these systems ignore privacy

and security issues

8

Second Paper: Anatomy of a Context Aware Application

9

Outline

• The Anatomy of a Context-Aware Application – Andy Harter, Andy Hopper, Pete Steggles, Andy Ward and Paul Webster. AT&T Labs, Cambridge, UK (cont)

10

Overview Paper describes the system components necessary to support

a context-aware application1. Location system2. Data model3. Distributed object model4. Resource Monitors5. Spatial monitors

Working system, real details

So what is the application? Teleporting/follow-me Application follows you as you move around a building Hmm!! Virtual dog?

• Definitely some gaming applications

• But what else? Virtual nanny?

• Framework not specific to this application

11

Location System

Would like: Fine grained (accurate in space) High update rate (accurate in time) Inexpensive Works indoors

Argue for ultrasonic Optical: expensive detectors; line of sight limitations GPS: dies indoors RF: multipath

12

Bat Unit

Radio transceiver, ultrasonic transducer and control logic

Each bat has a GUID Base station transmits a periodic message with the GUID;

corresponding bat responds with ultrasound

Use speed of sound in air (estimated from ambient temperature) to estimate location Ultrasonic receivers detect the delay and map it to distance

Use multiple receivers to get 3D location using multilateration

Reflections of ultrasonic waves – statistical outlier elimination (can same techniques be applied to RF multipath?)

13

Bat (cont’d)

14

Bat Unit (cont.)

It takes 20 ms between bat readings = 50 timeslots per base station per second Allow echos to die

Location can be used to measure orientation Attach many bats to the same object. Use the measurements to

infer the orientation If too cumbersome, can check “shadow” of a single bat

Base station can provide Location Quality-of-service to allocate time slots to bats based on the expected update frequency Bats carried by people – few times a second Bats attached to workstation – once every few minutes

15

Bat Unit (cont.)

Bats perform handover when moving from one base station to another (similar to the cellular networks) Hand off decisions can also be made based on the Bat location

Battery consumption is low, power consumed depends on the update frequency and power state Several updates a second = several month lifetime

95% of samples within 9cm’s of actual location 95% of samples within 25 degrees with multiple bats; within 70

degrees for a single bat (orientation)

Bat is good enough to be used as a 3D mouse; it has buttons and can communicate with basestation

16

How well does it work?

17

Modeling the environment

Detailed model describing entities in the real world and their possible interactions Environment consists of real objects: should use OO

modeling (agree?)

Modeling language based on entity relation diagrams and multiple inheritance Modeled people, computers, keyboards, monitors, networks,

telephones and furniture

Use CORBA and databases to implement persistent distributed objects

18

3 Tier Architecture

19

Populating and Updating the Model

Some elements are static (e.g., furniture) …but some are dynamic (e.g., is a keyboard in use)

• need to be updated automatically

Centralized data repository (to optimize information access)

Three classes of resource monitors: Machine activity e.g. keyboard activity Machine resource e.g. CPU usage, memory usage Network point-to-point bandwidth and latency

20

Client level event filters

Update Frequency The frequency at which items are monitored is based on

how quickly the item tends to change

Relevancy If a value has not changed significantly, it is not sent. This

value depends on the data being monitored

Caching Caching improves performance at the cost of consistency

21

Location Update

Each Bat location is tracked Translated into object location in a type specific way (e.g.,

person different than workstation)

22

API Absolute and relative spatial facts

“Person is at (x,y,z) facing in direction ” .vs. “Person is standing in front of the monitor”

Geometric containment is used for relative spatial facts

monitor person

Contained(person, screenspace)

23

Reasoning about Space

24

Scalability

With many devices, containment is complex

They use containment tree indexing system (a quad-tree based approach)

25

Implementation

BAT teleportation system With their earlier active badge based teleportation system,

they only knew that a user was in the room and so they had to cycle between multiple displays in a room

If a particular display was being use, they would still cycle that display because they did not monitor machines

If a machine is dead, their system would still wait because they did not monitor machines

With the BAT system, they have more accurate location information

Event driven programming style

26

Zones and Buttons

Action zone triggers teleportation Within maintain zone, teleported desktops are

maintained

Action

monitor

Maintain

27

Active Names: Flexible Location and Transport of Wide-Area Resources

Michael Dahlin, Thomas Anderson, and Amit Aggarwal. In Proceedings of the Second

USENIX Symposium on Internet Technologies and Systems, October 1999

28

Active Names

Idea: Intelligence in the network User specifies a name carrying some service intent Active name is mapped to a chain of mobile programs that can

customize how the service is located and presented Each service owning a portion of the namespace has control over

which protocols are used to access the service and where they run in the network

Encapsulates Service location, customization and transport Can be used as an extensible general framework for deploying

wide-area distributed services

How does relate to mobility/context-aware applications?

29

What exactly are they proposing?

“when a mobile user in Europe refers to cnn.com, the user probably wants to go to a different replica, fetch different data, and transform that data differently than a user in the US with a large screen and T3 connection”

30

Properties of the System

1. Completeness: it can support a wide variety of extensions

2. Simple extensions can yield big performance wins

3. End to end performance information can be gathered

4. Programming model can support location independent program execution

5. Programmable model supports extensibility

31

Naming – brief intro.

DNS translates from machine names to IP addresses opal.cs.binghamton.edu 128.226.123.101 static translation (usually valid for a few days)

DNS round-robin used to dynamically match name to host (based on machine load, for example) www.cnn.com 207.25.71.23 or 207.25.71.24 or …

Network level switch (for e.g. CISCO director) Automatically forward network packets to some server

32

Naming Intent

A URL is a hint to get some value from some server that is “appropriate” (in terms of locality and access costs) and that fits in my display

The exact host names and paths can be remapped to the appropriate forms.

My view of cnn.com depends on my device, my advertisement preferences etc. Your view of USA Today may be different from mine

33

Active Names Goals

Server selection – select appropriate server from replicas spread across the Internet

Client customization – Customize the page for the present client (e.g. Transcoding, customizable portals [e.g. my.yahoo.com])

Server customization – Advertisements customized for the user, collecting statistics (e.g. hit counts, ad rotation etc.)

34

Naming today

ClientProxy

HTTPServer2 Database

Server

HTTPServer1

DNSServer

1. Name2. host

3. URL - redirect

4. URL

5. Name

RESULTS

www.cnn.com

35

Active name system

Clients generate active names (domain:name) and name of a namespace program to resolve it.

Clients hand them to a resolver.

Name space program locates next program to run and then transports data to that program

Each program acts as a filter that transports and transforms its input to its output.

36

Active Name Architecture

37

Active name system

Active name resolver determines domain-specific program These programs are location independent and can run

anywhere Application specific, the name is resolved in domain-specific

manner

Domain specific code (e.g. ad rotation) After methods are associated with each active name

After methods are a list of programs guaranteed to be called They can perform client-specific transformation of data

38

Multi-way RPC for efficiency

ClientProxy Proxy

Proxy

Server

Traditional RPC•Results passed down the nodes.•Adds latency

39

Multi-way RPC for efficiency

ClientProxy Proxy

Proxy

Server

Multi-way RPC•Results sent directly•Minimal latency

40

Multi-way RPC

Security implications – someone else responds to your queries

Use capability certificates to authenticate response

Resource consumption limit should be managed. Resolvers use compute resources on foreign hosts.

Hierarchical namespaces.

41

Applications

Extensibility Compared DNS round robin (next server), distributed

director (closest server as measured by hop count) and Active names (number of hops; biased by a decaying histogram of previous performance)

Average latency follows Distributed Director at low load and DNS round robin at high load

42

Using Transcoding

Low JPEGLow JPEGQualityQuality10 KB10 KB

Thumbnail2KB

Crop40KB

Grayscale85KB

Original116 KB

Transliterate few bytes

Foggy road with fall Foggy road with fall foliage canopyfoliage canopy

43

Decision on where to transcode

Can transcode either on the server or proxy

Proxy is closer to client

Wide area network from server to proxy could be congested. Tradeoff between sending a smaller image across the country vs performing transcoding at the proxy

Active name migrates based on the current load

44

Composability

Use server-side include to update page based on the current request

Banner ad rotation Logs cookies

Implemented using server side and active namespaces.

Active namespaces are shown to be composable