The Virtual Personal Cloud

Published on June 2016 | Categories: Types, Research, Internet & Technology | Downloads: 30 | Comments: 0 | Views: 528
of 4
Download PDF   Embed   Report

An idea for returning to individual control of personal data that retains the benefits of Web 2.0 connectivity. Hypothesizes an architecture where all of your personal data and media reside in a personal cloud operated for you by a service provider or that you can host yourself. Applications connect from your browser to your virtual personal cloud, rather than to a large, centralized database. Social connections are maintained peer-to-peer rather than in large walled garden services.

Comments

Content

The Virtual Personal Cloud
Background Computing has gone through a complete cycle: First, we had large, centralized and reliable mainframes/minicomputers with no personal access or control Then we progressed to the PC era with total control and access to one’s personal data but very poor connectivity and reliability Now, we live in the browser + web walled garden era with centralized data, some control and access, great connectivity and reliability (Facebook, Google, etc) This paper explores whether it might be possible to return to a world with more control and access for the individual while still retaining the benefits of Web 2.0 connectivity and reliability. It presumes the continuation of exponential cost reductions in processing and storage, advances in virtualization technology and generous amounts of bandwidth. The Value Proposition Imagine if everyone had a “virtual personal cloud" (VPC) holding all of their media and personal, financial, business, health and lifestreaming data (pictures, microblog/blog entries, sensor data, geotraces, etc, etc). Consumers and small businesses would control access to their data and could grant it to anyone or any app they chose. This VPC service is something people would pay for much as they pay for other service subscriptions today (ISP, cell phone, Xbox Live, etc), with plans based on usage, bandwidth/speed, storage, etc. The VPC is not a backup service or safety deposit box – it’s a live, transactional data store where your online life happens. How would this work? Rather than SaaS apps like Facebook, Twitter and Gmail scaling huge services and databases that accommodate tens or hundreds of millions of users inside them, the VPC service industry uses on-demand virtualization to run hundreds of millions of instances of VPC's in a highly distributed way. The system could leverage the existing DNS system to resolve the location of any VPC – think a new top level domain called .vpc. So a typical HTML5 or AIR client application doesn't connect to www.gmail.com or mail.sbcglobal.net to get at my mail, it connects to the VPC at geraldhuff.vpc, which could be hosted by IBM, Cisco, Verizon, WellsFargo or wherever I choose, or on my own server if it's always on, which could give LAN-like data performance. For mail, as an example, that VPC instance would host both an SMTP server and POP server to accept and deliver email. If you are in my social network, my VPC pushes my status updates to your VPC in a server-to-server version of a P2P network with no "middle man" like Facebook or Twitter, using open standards like Activity Streams. If I want my financial data updated nightly, I can install an app on my VPC which uses credentials securely stored in the VPC to access financial institution data on my behalf. This service and

others might cost money. VPC service providers would have a billing system just like cell phone companies, so I get a consolidated bill that includes all my services. Personal Benefits - You can send me email forever at [email protected], which I can read with many different clients while remaining untethered from an ISP or any other entity - I can move my VPC at any time to another provider with a simple "DNS" change and copy operation - Because of extreme portability, my costs for VPC service will go down with technology cost curves because it will be very competitive - I can also change any client (social network, personal finance, email, etc) at any time, because they all conform to the VPC open standards - In fact, my friend Joe can view my profile data and his social graph with a social network client that he prefers while I view his profile and maintain my own with a different client - When I want to try a new website, I don’t have to register, I just grant it access to store things in (and maybe read certain things from) my VPC. In this way, the open standards and API’s at the front edge of the VPC become “the platform”. - I never have to give my personal and credit card info to a web site I use and find value in because I can get a consolidated bill from my trusted VPC provider - I can make a backup of and easily restore all my data at any time because it's stored in a single "capsule" I can grab a hold of - a major problem in today's world where "my" data is spread out over dozens of web sites. - I can grant and revoke access to my data at any time (needs legal restrictions on 3rd parties making copies of what they read over API’s) - This is just like having data on my own personal hard disk, but better than my hard disk, my VPC is available to me (and other VPC’s) 24/7 from anywhere and any device, auto-backup, etc. Industry Benefits - A huge new recurring revenue stream that simply doesn’t exist today, like the first time cable TV appeared and people started paying for TV content or consumers started paying for internet access - New web startups can scale incredibly without any infrastructure, unleashing tremendously low cost innovation for everyone. That’s because they never store anything. In fact, with AIR and AJAX clients the typical startup could support millions of users storing only, say, 100K of code on a single server. In this model, the client runs in the browser and accesses everyone’s VPC directly. Or, the code could be posted to a central VPC app store, so a startup could support a million users without even a web site. And with VPC-provider billing, all they do is collect checks. So, this is very much like iPhone apps/app store today, but untethered from any device and capable of storing lots of data in VPC’s. - Development is also incredibly easy, because I can run multiple VPC’s on my desktop computer and so my development environment is my deployment

environment. I don’t need to worry about “what happens to this query when I get a million users?” because very VPC is for a single user. It’s the most productive development environment possible, just like single user desktop software used to be! Downsides Such a scheme of highly distributed personal data makes mass public user contribution to a central searchable site a la Flickr or Delicious problematic. People could of course opt in to pushing their status updates or bookmarks to services like that, which could aggregate them. In fact, it would be relatively painless to syndicate ones bookmarks or pictures or other content to multiple aggregation sites. As a counter-argument to the “unleashed innovation” above, some might argue that rapid innovation will be stifled by the reliance on open standards that move slowly. What if an innovator wanted to connect a new kind of content to the nodes in your social graph? To avoid this, it will need to be easy for applications to store proprietary information in the VPC until it can be standardized. Obstacles Of course there are many obstacles to something revolutionary like this, e.g. - All the big centralized SaaS players like google, facebook, twitter, etc. who have a vested interest in the centralized status quo and currently hoard the data - Consumers/small businesses would have to pay for some things they get for free today - A whole host of technical issues around scaling the virtual machines required for hundreds of millions of individual VPC’s, running 3rd party app code safely inside them, etc. - User experience challenges for people managing their own information and access control schemes (they don’t do this explicitly today, they just “use apps”) - The difficulty of creating the many open standards required - Chicken and egg issues – what application provider will write to a VPC back end when they don’t exist, and who wants a VPC with no apps? - The landgrab issue for usernames in the .vpc space and the difficulty of getting a new TLD established - …Just for starters Technical Thought Starters The basic stack of a VPC would have a storage backend like MySQL, a robust data API layer controlled by OpenID/OAuth that implements various open standard data protocols (SMTP/POP/Atom/Activity Stream/Open Social/etc) and a code container that can run sandboxed 3rd party code. The VPC server system would ideally be an open source project. There would be standard VPC host services for billing and account management, including host-to-host protocols like “move account X from you to me”.

An idea not in a vacuum Inspired by slide 32 of this presentation by Jyri Engeström Note that there is a seed of this idea in http://en.wikipedia.org/wiki/Vendor_Relationship_Management (Personal Data Stores) Various bloggers like Om Malik have also talked about breaking down the data silos of Facebook and others The entire Open Stack movement (OpenID, OAuth, OpenSocial) is headed down this path, but still trying to stitch together various highly centralized services Apple’s mobile.me is somewhat like this, except it’s locked into Apple and more about syncing There are "Personal Data Vault" products today, but they are like safe deposit boxes and are not live and transactional Opera’s recent “turn your computer into a web server” effort

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close