<<

Personal Cloud Compung Monica Lam, Stanford University

Seok‐Won Seong, Sudheendra Hangal, Chris Brigham, Debangsu Sengupta, Greg Bayer, Jiwon Seo, Mahew Nasielski, Michael Fischer, Ben Dobson Programmable Open Mobile Internet

NSF Expedition, August 2008

Education Paul Kim Applications Leo Guibas HCI Scott Klemmer Security Dan Boneh John Mitchell Monica Lam Distributed Systems David Mazieres OS Phil Levis Mendel Rosenblum Architecture Christos Kozyrakis Economics Ramesh Johari Guru Parulkar Nick McKeown Networking Fouad Tobagi Andrea Goldschmidt Radio Arogyaswami Paulraj Opportunity for a Clean‐Slate Approach

mainframe mini workstation PC laptop phone

10x cheaper, 10x more users Convergence of CE, PC, phone & Networks

computation internet

media

communication games Tian Di Ren (Heaven‐Earth‐Man) 3‐Tier Architecture

servers data Heaven Internet 天

My key, cache, window into my Man digital ID, digital personality, digital assets, 人 and the internet Personalize the generic PC, borrow the power, display, keyboard, memory, etc

Earth PC,TV 地 at home, on the road, in hotels, on the plane Background: First Approximaon

Internet LivePCs User Data

• NSF‐funded Stanford Collecve project, Moka5 • LivePCs: managed virtual machines in the cloud • PCs: generic compung plaorm • Phone: mobile access, write‐through cache, network accelerator Social Networking 1.0

Applicaon‐centric design • , MySpace, LinkedIn, Yahoo, Google, Apple, … • Flickr, Shuerfly, Picasa, Adobe, , … • Twier, , Crowdvine, Friendfeed, … • , WAYN, , Frengo, , Plaxo, … • 3Jam, Zinn, Loopt, …

Strengths: convenience, available everywhere, free Weaknesses • Data lock‐in, loss of data ownership and privacy

• Disparate data sources • Inconvenient to upload data and relaons • Inefficient, difficult to scale The Omniscient Monopoly

In year 2024 …

What’s happening today … Social Networking 2.0

Distributed social networking infrastructure • Maintenance‐free home servers, part of an ISP service • Technically more scalable, but more complex Design goals • Data privacy – Separate data ownership from storage & app provider • Compeve and praccal – Convenient – Interoperates with exisng services, enables easy migraon • Beer usability PrPl Personal Cloud Infrastructure The PrPl Infrastructure

Person‐Centric Social Networking

My personal cloud • Unified view of all my stuff My Personal Cloud accessible on any device • Federated storage system My Devices, My Relaons Resources Distributed social networking My Friends’ Public • Fine‐grain access control Cloud Cloud • Egalitarian: data privacy, scalability High‐level programming • Locaon‐agnosc semanc query • Generic data navigaon system System Architecture

PrPl Applicaons Collecve Phone Movie Social Personal Memory Book Card Reviews Feeds

SPARQL query

Friend’s Personal‐Cloud Butler Friends of Blobs A Personal PrPl Semanc Index friends Butler

Phone Laptop Home Server imap Google Yelp Facebook Data Stewards Personal‐Cloud Butler

Privacy and confidenality – Lives at home – Knows all confidenal info Intelligent assistance – Saves you me – Shows discreetness, suggests ideas Social networking – Gossips and negoates with other Butlers – Provides plausible deniability Uber Phone Book

Combines informaon from: Gmail, iPhone, Facebook 100 lines of code Collecve Memory

Coherent, real‐me view of photos from many users 550 lines of code Personal Yelp

Filters Yelp informaon with personal friends list 580 lines of code Personal Feeds

Real‐me feeds from friends No central server 480 lines of code Movie Card

Suggests movie dates – Friends’ calendars – Movie schedule 620 lines of code Resistance is Fule?

• To provide a choice • To educate the public • To provoke a response Key: Monezaon of Privacy

Applicaons requiring privacy – Health care, finances – Blind date

Markeng based on whole personal digital history – Client‐based targeted markeng – Market research: sampling of private data Conclusion

Tian Di Ren 3‐Tier Architecture – The Collecve system The Omniscient Monopoly Distributed Social Networking Infrastructure – Person‐centric versus applicaon‐centric – Separate applicaons, data storage, ownership – High‐level programming abstracon Monezaon of Privacy