Beyond Mobile: Pervasive Computing and Clouds · 2019-03-13 · Simple Taxonomies of Clouds Deployment Model – Public, Private, Hybrid Functional – Test and Development clouds,

IBM Research

© 2010 IBM Corporation

Beyond Mobile: Pervasive Computing and Clouds

Chandra NarayanaswamiIBM TJ Watson Research CenterHawthorne, NY, [email protected]

IBM Research

© 2010 IBM Corporation2 Beyond Mobile: Pervasive Computing and the Cloud

What is Cloud Computing?

� NIST Definition

– Rapid Elasticity, Measured Service, On Demand Self Service,

Ubiquitous Network Access, Location–independent resource

pooling.

Simple Taxonomies of Clouds� Deployment Model

– Public, Private, Hybrid

� Functional

– Test and Development clouds, Storage clouds, Compute clouds,

Desktop clouds

� Industry specific clouds

– Healthcare, traffic, entertainment, automotive

IBM Research


Successful Cloud Services to date

� Collaboration

– Mail, messaging, social networks, web meetings

� Commerce

– Web commerce, product information, app-stores

� Maps

– Navigation, location-based services, street views

� Entertainment

– Music, photos, videos, social networks, hosted games

� Health

– Health care monitoring, forums

� Financial sector

– Budgeting tools, etc.

� Business Analytics - emerging

IBM Research


What is Pervasive Computing?

� Computing in places you did not traditionally expect

– Everyday objects, sensors, medical devices, household

items, shoes, clothing, sporting goods, furniture, toys, etc.

– Of course what you “expect” is changing rapidly

So, what is the connection? – Device Symbiosis

� Quite simply, clouds can expand the reach the pervasive computing by offering several key functions

– Remote computation, data aggregation, analytics, storage, test

environments, management and monitoring

IBM Research


Why Device Symbiosis and what is it?

� Mobile/Pervasive Devices are resource challenged

– Computing power, batteries, interfaces, storage, displays, bandwidth

� Mobile/Pervasive Devices have several advantages

– Personal resources, rich sensors, ubiquitous presence, always accessible, geographically dispersed

� Cloud resources are and will always be richer

– Access to larger pools of machines, steady sources of power, better bandwidth, large storage, and better connectivity

� Cloud resources/devices have limitations

– Not personal, don’t sense unless fed data, geographically limited dispersion

� Device Symbiosis

– Mitigate the disadvantages of small devices with clouds

– Provide clouds with the data it needs!

IBM Research


Some current trends

Smarter people

Smarter

Devices

Responsive and

personalized clouds

Smarter

Clouds

Smarter device

cloud

interaction

Context data,

aggregation

People

Cloud

� Clouds

– Affordable clouds, plethora of clouds

– Industry momentum, cloud services

– Deployment of network infrastructure

� Devices

– Many cheap usable devices, new applications, worldwide trend, wireless

– Phones as sensors, cheap sensors

� People

– More inter-connected to each other

– More of their data is moving to the cloud

– People-contributed data to improve processes/tasks

– Blogs, tweets, photos, Twitter

IBM Research


People-oriented clouds� People

– People will continue to interact with information and processes using variety of devices and cloud services

– However, as a result of the current trends,

• People-people interactions are going to be mediated by cloud services and devices• More personal data is moving to clouds (e.g., health vault, digital wallets) – new opportunities• Federation of clouds, data/services/authentication, lightweight runtimes, usability, etc., become

important

� Cloud Services

– Proactive, personalized and adaptive cloud services that learn and react will become ubiquitous

• Personal Concierge in the cloud - background services and agents that work even when user is not connected – e.g., teleconference “callout” instead of “callin”, automated roll-call after disaster, enterprise travel portal and schedule changes/cancellations

• Personal Data Lenses – these magnify data that is relevant to you and hide the fluff in the clouds

– “Give something get something” patterns - Data broadcast, decision building in cloud, decision broadcast

• federation of data/services/authentication among clouds and establishing user trust become important

– Programming models need to evolve to make it easier to build/deploy/orchestrate concierge micro-services on clouds that are accessed from a variety of devices

� Devices

– Seamless migration of user experience from one device to another, e.g., IM on desktop, transition active chats to mobile phone, then to tablet, etc.

– Services and data adapt to local environments and devices – wireless/cell, location

IBM Research


Some future cloud services

� “Smarter Planet” – Pervasive Monitoring, Analytics and Sensors

– Traffic, Power, Water, Healthcare, etc.

� People-oriented clouds

– Zone-specific people-clouds

• Airplane clouds - Entertainment, use of personal devices

• Hotels - Replace remotes, alarms, etc with personal phones, cleaning request services

• Retail zones - Stores, malls, augmented reality support

– “Pocket Clouds” for personal device pool

• Micro cloud services

• Nimble clouds – fired up on demand on local devices

� Emerging economy clouds

– Phones + Clouds -> No PCs required

IBM Research


With these changes – come some familiar challenges

� Security and Privacy

– More data, services, and players

• Many parties – who is reponsibl for what when services are composed?• personal devices accessing enterprise data

– Isolation between personal and enterprise data on devices

� Self describing services that can interact

– Service catalog and its maintenance

– Automated and personalized service discovery and composition

� Scaling and reliability

– Reliability of services and data – services disappear, morph

– Multi cores, redundancy, flash memories

– Not every device can directly connect to cloud

• So impedance matching - gateways, etc.

� Energy management

– Mobile devices and sensors are really challenged –

• When to rely on the cloud?• when to send data to the cloud?

IBM Research


As you put processes to sensors into clouds – some new challenges arise as well

� Business Support Services

– Service catalog mgmt, onboarding/offboarding, subscriptions, audits, reports, billing, settlement, point of service to human transition

� Pricing models

– What to charge, revenue apportionment, cheaper prices in off-hours?, mountain of charge data records – worth it?

� Operational Support Services

– software upgrades, customer support, which sub-service is broken, batteries and energy

� People contributed and sensor contributed data

– How to mix and match, reliability of data

– “Give something get something” patterns - Data broadcast, decision building in cloud, decision broadcast –

• New rewards mechanism to encourage cooperation/participation

� Lock-in, migration to other clouds

– Data, services, continuity,

IBM Research


Some specific examples of cloud services

� Enterprise mashups

– Event Maps, Organization maps

� Retail shopping

– Mobile wallets, electronic receipts,

� Sporting

– Collaborative “spectating”, in stadium viewing

IBM Research


Mashups – a’ la’Google Maps

External

Data source

Geo Map

Data

Location-based

Correlation

Geo

Map / Mashup

Visualizer

interaction

Note: Just interacting with the visualized geo map data is valuable;

Mash-up adds oomph

cool

People Centric Mashups OrgMaps

External

Data source

Org MapData

Person-based

Correlation

Org

Map / Mashup

Visualizerinteraction

cool

IBM Research


PeopleCentric Mashups – Organizational Dashboard� HR metrics and processes

– Employee training and visibility - Skills, patents, publications, education, mentors, mentees, etc.

– Employee compensation, performance history, awards, regulatory compliance, etc, and related hotspots.

� Financial performance and budgets

– what business targets have been met, how much money has been spent, etc.

� Enterprise information

– Pointers to presentations made, blogs, podcasts, media library, papers, patents, etc., created by individuals – “People RSS” – similar to geotagged RSS feeds.

� CIO and internal asset management

– Track machines, sw resources (licenses used, User ids, Notes cost, storage costs), etc.

� Project Management/Tracking and Staffing

– Data relating to project staffing could include plotting skills of employees and then zooming and panning the mashed up OrgMap to find suitable staff for a project.

– Impact of organizational distance on product quality, etc.

� Cross organizational sharing

Many other applications are possible!

IBM Research


Lotus Mashups - OrgMaps

IBM Research


Problem – Attending and Managing Large Events

Drawbacks of current Web sites

One size fits all; users cannot

– customize views

– tailor agenda

� No collaboration; users cannot

– tag, discuss, rate, network

� No event mgmt console/facilities

� No preconference planning

� Limited post conference use

� Abrupt changes between viewing overview and details

Example above: DAC 2010 Web Site

IBM Research


EventMaps – Concept and Value Proposition

� A framework for creating compelling event sites that facilitate interaction, discovery and collaboration utilizing the following principles:

– Visualize an event as a large canvas that can be zoomed and panned

• Support smooth transitions between overview and detail views

• Support hierarchical organization into tracks, sessions, etc.

– Make the canvas a mashup substrate for overlaying information

• View and contribute tags, ratings, discussions etc. on the various parts of the event

• Extensible to mashups from other data sources

– Personal/community/collaboration elements are embedded/anchored in the substrate.

• Provides automatic context for communication

– Interoperate with other visualizations, such as planograms for mapping the event locations, or event details panels

� Manage large events with visualization, collaboration, and gathered information

– E.g., Capacity planning for meetings scheduled, their size, # accepts from users, location of participants, etc

IBM Research


EventMaps: Mega Meetings Made Fun & Easy

• Fluid interface with selective zoom and Pan,

• Personalization/search/filter and calendar export

• Social Sharing (Annotations, Comments, Ratings)

• Location Maps

• Deployed at CHI 2010, Lotusphere 2010, ICIA https://researcher.ibm.com/eventmaps/

EventMaps: Mega Meetings Made Fun & Easy

IBM Research


Track View

ViewReset

Button

Search

BoxVie

w

Selecto

rSession

Categories

Tracks

Favorites

Filter b

y

Favorites

ShowPopularity

FavoritesPreferences

Location-track View

LocationTracks

Locations(rooms)

HelpPage Fe

edba

ckPag

e

IntroPage

Table View

row foreach

session

Location-map View

Thumbnailsof all maps

Map currentlyviewed

IBM Research


Celadon – Zone-based services with clouds

Hospital

Train Terminal

AirportSchool

Theme

park

Sport Ground

Home/OfficeMuseumFinancial Service

Audio

Service

Entertainment

Video Service

U-Learning

Information & Data Service

Health Care

Celadon ServicesMobile Devices

Ubiquitous Collaboration Infrastructure

Diverse Devices Many Spaces Plethora of Services

Celadon Zones with Facility Devices

IBM Research


Building Block Services

� User Maintenance Services

– Enrollment Data, User preference services

� Zone Operating Services

– Member information, location data, service directory, interaction information

� Content Delivery Services

– Focused on delivery to device rather than web channel

– Device could be in-store (personalized ads to kiosk) or mobile (short

messages, mobile web enabled pages).

– SMS data service

� Management and Monitoring Services

– Activity data service, analytics data service, authentication services

� Retail Intrinsic Services

– Loyalty information data, product information data, Active Shopper data

22

IBM TJ Watson Research Center

© 2009 IBM CorporationBeyond Mobile: Pervasive Computing and the Cloud

Mobiloyalty: link loyalty program to on-the-spot coupon delivery

� User enrolls in program.

� User taps-in on entering store.

� Coupons delivered to user’s mobile phone.

� Timed push of coupons.

Tag Reader/Display

Device RecognitionCoupon Retrieval

Product Selection

Coupon Delivery(SMS/WEB)

Enrollment Web

Application

Enrollment, Tag Enablement

Weekly SMS Push Campaign

In Store Experience

23



CeladonMobiloyaltyOperationsCenter

A: Mobile Commerce SiteDisplay coupons to device

B: Active Shopper Service:Display user status/coupon issuance

C: Activity ServiceLogs roundtrip interactions with users

D: Analytics ServiceConverts activities into time of week/dayHistogram, computes expected visit times

E: Analytics ServiceIndividual user view

F: mapping activity to store layout

24



Campaign Analyzer

25



Campaign Analyzer

Occluded for Privacy Reasons

Occluded for Privacy Reasons

26

S&D/Research


POS

subway

vending

POS

Mobile Wallet – value added services

� A cloud-based service� Stores and manages heterogeneous collection of wallet content

types

• Payment Instruments

• Digital coupons and receipts

� Implements user customizable access rules

• Coupling basic access operations with authenticated roles

� Manages user preferences

• e.g. which merchants or merchant classes have read access to certain content types

� Tooled and built using above service frameworks

� Providing secure user access via:� Mobile phone browser

� Mobile application on device

� Desktop computer browser

� And providing secure programmatic access via:� Content providers

• e.g., manufacturer wishes to submit coupon to user wallet

� Content consumers

• e.g., retailer redeems coupon from user wallet, presented at POS

� Transaction engines

• e.g. financial services provider reads payment instruments in wallet, updates balance

Note: Graphics above from other web sites

IBM Research


“submit/updatepayment

instruments”

“establish wallet,set policies”

redeemcoupon

extract coupons, submitreceipt

“display applicable offers, payment instruments

“submitpost-salecoupons”

users content

providers

choosepaymentinstrument

Digital Wallet

Data Service

Persistent

storage

access

providersco

ntent

consumers

wallet host

“submittargetedcoupons”

“view/arrange contents”

“select/enable coupons for POS

checkout”dis-connected mobile userconnected

mobile user

desktopuser

coupon clearing house

subwayNFCgate

merchant POS system

bankweb site

merchant commerce

site

POS touch station

financial services provider

consumer products supplier

marketing engine

“display payment instruments

and balances in wallet

extract e-token,

decrement balance

Value Added Functions and EcosystemNew mobile commerce functions will arise at the intersection of mobile payments and mobile marketing and will initiate customer adoption of mobile commerce in North American markets and expand adoption in Asia and Europe

MobilePayments

MobileMarketing

MobileSocial Networking

Mobile Digital WalletBringing new services

to mobile payments

MobileBanking

“Whilst NFC trial results so far have been encouraging, the industry

as a whole will need to convince both consumers and merchants of the merits of yet another payment

mechanism on top of cash, cheques, credit and debit cards, and to allay understandable (even if

unfounded) fears and scepticism about the security of the mobile wallet.”

Howard Wilcox, Juniper Research [1]

28



Zone-based services in malls – Symbiotic Business Collaboration

Mobile Device User Integrated with Store Finder Business Service

In Shopping Mall Using Interactive Large Format Display

29



Attract ScreenSelect Offers To Redeem

New Offers Uploaded

Versatile POS Scenario

Select Payment Instrument

(Optional) Preparatory Station One Touch Checkout

Transaction Complete

In Store - Mobile Promotions and Offers

14443

14443

30



Conclusion

� Mobile devices and sensors can’t do it all

� Clouds less interesting without data from mobile and pervasive devices

� New cloud-based services will emerge

� Financial and business models will be key

IBM Research


Pervasive Computing and Clouds

The future is exciting

A new era looms over the horizon

Beyond Mobile: Pervasive Computing and Clouds · 2019-03-13 · Simple Taxonomies of Clouds Deployment Model – Public, Private, Hybrid Functional – Test and Development clouds,

Documents