> >I don't think people think encryption solves problems (at least not for ISPs). > As encryption becomes more pervasive (customers' reaction to their DNS being >logged; or their HTTP being logged; or whatever ISPs end up compelled to do), >so the cost to ISPs increases. The end-game of the crypto arms race is that >$our customers buy VPNs, and UK government forces $us to block consumer VPNs >that are terminated outside UK jurisdiction (because $terrorism or $copyright >or whatever). From there, one only needs to look at what little we know of >the Great Firewall of China to get a glimpse of the challenge and costs to $us >to disrupt VPNs.
How many customers will buy VPNs though? How many actually have a clue what DNS is? > >e.g. http://blog.zorinaq.com/?e=81 - "It is only after I reached this point in >my tests that I did some deeper reading and learned that the GFW uses machine >learning algorithms to learn, discover, and block VPNs and proxies." > >Machine learning firewalls sound like a significant bite at profits for any >ISP. It sounds like a fantastic engineering challenge to build some thing that does this with low cost. Us engineers can do pretty much anything we want and we’ve proven this before when the same challenge applied a few years ago on content filtering and that was solved for peanuts. I’ve heard the cost argument in these debates over and over again, so when I look at what the EU is doing on privacy and other legislation on data protection, we (the ISPs) seem to be happy to do things like EU-DPR without arguing the costs (which are significant). Ultimately, in the short to medium term the cost gets passed onto the customer and we the smart engineers deal with the costs over time to stay competitive. For me the key issue in this is the sheer volume of data that will be generated and the ability for any organisation to make use of it in a credible way and turn it into valuable information. Regards, Neil. (who isn’t in favour of a lot of this bill but loves engineering challenges).
