Privacy as a Service sounds like a very plausible business model for the future (or even today).
edit: I forgot about this[1], so apparently it already exists!
[1]: http://news.cnet.com/8301-31921_3-57412225-281/this-internet...
edit: I forgot about this[1], so apparently it already exists!
[1]: http://news.cnet.com/8301-31921_3-57412225-281/this-internet...
At Blekko we've been very aggressive at building privacy in from the start, and we talk about that with folks all the time.
When I was at Sun I was deeply involved in doing security work both in the kernel and on the network. There were lots of places to improve.
In both the Blekko and Sun cases the 'better' version doesn't seem to carry a lot of retail influence. Which is to say if you offer a consumer a choice, the 'secure' one which costs a bit more (either in cash or in complexity) or the 'insecure' one which is cheaper, not enough people pick 'secure' to make the investment pay off. You have to pay for your investment somehow, and the fewer people who are willing to pay for a feature, and the more it costs to implement, means a higher per-consumer price for that feature. Security often pushed that cost threshold over the limit into 'non-business'.
When we were discussing implementing the best privacy policy and technology in any search engine I was pretty clear that having the 'best privacy policy of any search engine' was a nice add-on feature but you couldn't base your business on it. We had to have a search engine that was just as good or better than the competition, and the fact that 'oh by the way its got the best privacy policy' could be icing on the cake, but for the majority of people it would not be their reason for using it.
You can see this sort of effect in lots of different communities.
Perhaps there's an opportunity for a company based in Switzerland to run a private TOR network for obfuscation with guaranteed bandwidth. Private TOR network has its own problems though. You need a large number of users to anonymize each other's data, and the block of assigned IPs can be treated as a single entity and blocked / rerouted as a result.
Admittedly, it is still complicated to set-up your existing web-browser to use it, and latency is still a huge issue - not helped by the limited number of exit nodes.
and of course Phil Zimmerman is giving it a go:
[ though all of that ex-SEAL presence among the founders sets off some of my paranoid alarms about US government backdoors ]
I certainly hope there's enough demand to sustain privacy protection services…but I'm not optimistic. I've never had much luck convincing "regular" people that privacy concerns are worth abandoning the convenience/entertainment of things like Google and Facebook.