IPcentral Weblog
  The DACA Blog

Friday, September 12, 2008

Still Cloudy on Cloud Computing: A Matrix to Guide the Coming Policy Debates
(previous | next)
Today Google's DC office hosted an interesting panel on cloud computing. What was missing was a good definition of what cloud computing actually is.

While Wikipedia has its own broad definition of cloud computing, many think of cloud computing more narrowly as strictly web-based for which clients need nothing but a web browser. But that definition doesn't cover things like Skype and SETI@home. And just because PFF has implemented Outlook Web Access so we can access the Exchange server via the Web, doesn't necessarily mean we've implemented what most people might think of as "cloud computing." Yet these are all variations on a common theme, which leads me to propose my own basic definition: any client/server system that operates over the Internet.

To understand the potential policy and legal issues raised by cloud computing so-defined, one must break down the discussion into a 4-part grid. One axis is divided into private data (e.g., email) and public data (e.g., photo sharing). The other axis is divided into data hosted on a single server or centralized server farm and data hosted on multiple computers in a dynamic peer-to-peer network (e.g., BitTorrent file sharing).

Examples User Data is Public User Data is Private
Centralized Server(s) Blogs Discussion boards Flickr Web-based email servers Windows Terminal Services
Peer-to-Peer BitTorrent FreeNet (article) Skype Wuala

There are also a great number of peer-to-peer cloud computing projects that don't require the sharing of user data. SETI@home may be the most well-known example: When the Search for Extra-Terrestrial Intelligence (SETI) project lost its funding and could no longer afford the massive servers it used to process the data from its radiotelescopes, it realized that it could distribute the work to Internet users in the form of a screensaver (thus the SETI work would only be done when a user's computer was idle).

It is encouraging to see that Congress is no longer considering simply outlawing cloud computing (which used to be called distributed computing), but if there is to be an intelligible debate about policy responses to cloud computing,, we must define our terms and realize that policies beneficial to some forms of cloud computing may complicate-sometimes fatally, in business terms-other forms. For example, regulations imposed on companies storing users' personal data may stymie peer-to-peer backup applications like Wuala, which distributes each user's backup data to other users, but uses encryption to prevent users from accessing the data they're storing for others. Wuala might be forced to close down if regulations requiring companies to keep records for a set period of time or follow separate procedures for minors were interpreted to apply to each Wuala user.

As Georgetown CCT professor Mike Nelson explained at the Google workshop, technology generally follows a clear evolution in the following steps: from hardware to software to people to organizations to policy. It's taken a long time to educate lawmakers about the Internet. Today's panelists all seemed to agree that cloud computing could be "the next big thing." That necessarily means that the education process for lawmakers needs to start all over again, explaining the ways in which cloud computing is similar to prior technologies, the ways it's different, and the salient differences among the four broad categories of cloud computing described above. Until that's done, any talk of legislation in this area is simply premature.

posted by Adam Marcus @ 5:50 PM | Internet

Share |

Link to this Entry | Printer-Friendly


Actually the Wikipedia article (which I've been putting together over the last few months) both covers and specifically mentions Skype and SETI@home.

The vision for cloud computing is similar for that of grid computing (a computing 'grid' like the electricity 'grid' that you can 'plug into' from anywhere), only grid was essentially hijacked and pigeon holed with high performance computing tasks. There are still a handful of vendors trying to rebadge everything as 'cloud computing' (which mostly misses the point of not having to care about the nuts and bolts) but hopefully sanity will prevail despite the presence of significant marketing budgets.


Posted by: Sam J at September 14, 2008 6:45 AM

Sam: I like your definition too--Cloud computing is not having to care about the nuts and bolts. But is it the end-user that doesn't have to care or the administrator? For end users, all they usually need to care about is having a current Web browser. But for the administrator, even when there's no central server to manage, the peer-to-peer network still needs to be managed (via software updates). Is cloud computing just another term for IT outsourcing? Some at the Google talk equated "clouds" to "internets", pointing out that just how many small "internets" were connected to create the "Internet", he wants to be able to create a small "cloud" just for his organization's internal purposes.

So it seems like the term cloud computing is being applied to things that can be mutually exclusive: thin-client computing and server outsourcing.

Whether cloud computing should remain this all-encompassing term depends on whether these different aspects require different approaches--both from implementers and policy-makers. I believe they do.

Posted by: Adam Marcus at September 15, 2008 9:15 AM

Post a Comment:

Blog Main
RSS Feed  
Recent Posts
  EFF-PFF Amicus Brief in Schwarzenegger v. EMA Supreme Court Videogame Violence Case
New OECD Study Finds That Improved IPR Protections Benefit Developing Countries
Hubris, Cowardice, File-sharing, and TechDirt
iPhones, DRM, and Doom-Mongers
"Rogue Archivist" Carl Malamud On How to Fix Gov2.0
Coping with Information Overload: Thoughts on Hamlet's BlackBerry by William Powers
How Many Times Has Michael "Dr. Doom" Copps Forecast an Internet Apocalypse?
Google / Verizon Proposal May Be Important Compromise, But Regulatory Trajectory Concerns Many
Two Schools of Internet Pessimism
GAO: Wireless Prices Plummeting; Public Knowledge: We Must Regulate!
Archives by Month
  September 2010
August 2010
July 2010
June 2010
  - (see all)
Archives by Topic
  - A La Carte
- Add category
- Advertising & Marketing
- Antitrust & Competition Policy
- Appleplectics
- Books & Book Reviews
- Broadband
- Cable
- Campaign Finance Law
- Capitalism
- Capitol Hill
- China
- Commons
- Communications
- Copyright
- Cutting the Video Cord
- Cyber-Security
- Digital Americas
- Digital Europe
- Digital Europe 2006
- Digital TV
- E-commerce
- e-Government & Transparency
- Economics
- Education
- Electricity
- Energy
- Events
- Exaflood
- Free Speech
- Gambling
- General
- Generic Rant
- Global Innovation
- Googlephobia
- Googlephobia
- Human Capital
- Innovation
- Intermediary Deputization & Section 230
- Internet
- Internet Governance
- Internet TV
- Interoperability
- IP
- Local Franchising
- Mass Media
- Media Regulation
- Monetary Policy
- Municipal Ownership
- Net Neutrality
- Neutrality
- Non-PFF Podcasts
- Ongoing Series
- Online Safety & Parental Controls
- Open Source
- PFF Podcasts
- Philosophy / Cyber-Libertarianism
- Privacy
- Privacy Solutions
- Regulation
- Search
- Security
- Software
- Space
- Spectrum
- Sports
- State Policy
- Supreme Court
- Taxes
- The FCC
- The FTC
- The News Frontier
- Think Tanks
- Trade
- Trademark
- Universal Service
- Video Games & Virtual Worlds
- VoIP
- What We're Reading
- Wireless
- Wireline
Archives by Author
PFF Blogosphere Archives
We welcome comments by email - look for a link to the author's email address in the byline of each post. Please let us know if we may publish your remarks.

The Progress & Freedom Foundation