Media Influencer
helping people break out of pigeonholes since 2003
Media Influencer
- Daily links 07/15/2016 July 15, 2016Picturing an Imaginary Animal Kingdom | PROOF from strother, note the last question and answer of the interview tags: [...] […]Adriana
- Daily links 03/26/2016 March 26, 2016Ethereum: Rise of the World Computer – Silicon Valley Data Science tags: [...] […]Adriana
- Daily links 01/31/2016 January 31, 2016Made in Hong Kong // 香港製造 | Angelina Hue tags: [...] […]Adriana
- Daily links 01/13/2016 January 13, 2016You say advertising, I say block that malware good article tags: [...] […]Adriana
- Daily links 09/22/2015 September 22, 2015Unfit-Bits tags: self-tracking [...] […]Adriana
- Daily links 08/05/2015 August 5, 2015The adblocking revolution is months away (with iOS 9) – with trouble for advertisers, publishers and Google | The Overspill: when there’s more that I want to say “discussion of this post on Hacker [...] […]Adriana
- Daily links 08/04/2015 August 4, 2015Return Detail A Howlett Posted from Diigo. The rest of my favorite links are here. […]Adriana
- Daily links 07/27/2015 July 27, 2015This Works | Perfect Legs Skin Miracle, 120ml | NET-A-PORTER.COM Click to zoom Click to zoom [...] […]Adriana
- Daily links 07/15/2015 July 15, 2015Greece: Here Is Why #ThisIsNotACoup | Andreas Souvaliotis tags: [...] […]Adriana
- Daily links 06/22/2015 June 22, 2015Contact us – Simply Secure Contact us To join the conversation about usable security, follow us on Twitter (@simplysecureorg) or join us on Slack (email slack@simplysecure.org Posted from Diigo. The rest of my favorite links are here. […]Adriana
© 2009 Media Influencer. All Rights Reserved.
This blog is powered by Wordpress and Magatheme by Bryan Helmig.
Dave Walker
on Aug 23rd, 2014
@ 19:27 pm:
Depends on how you look at the problem.
Trust models are far simpler in client-server, but you add risk around discoverability.
In peer-to-peer, there’s more intermediaries which need to be trusted, unless you either build in redundancy (consider the matter of quorum size and data repository count in erasure coding or k of n key splitting, similarly parity in RAID) such that no peer has sufficient data to be able to do anything useful with it, or implement some kind of node reputation model.
A peer-to-peer network, or mesh network, especially where both are formed ad-hoc, is like a crowd of people; a peer-to-peer network which comes together over time, is like a “social network for machines”. Both can have rogue nodes which do things other than what they advertise, and the flipside of privacy for a machine, is whether it needs a hardware-persistent, unspoofable identity so that it can both keep track of the reputations of peers it might consider connecting to, and also be kept track of for its own reputation.
The “noise” element in the article, doesn’t ring true for me – there’s enough researh around identifying systems based on bias and variance in physically-imperfect RNGs…