the private net
This article has been gestating in my mind for quite some time. It was pushed closer to being written by an incident a couple of weeks ago. I run a number of different websites (in this context namely iphone-essentials.ch). Also, since a couple of years, I had a Google AdSense account and ran ads, not to cover my costs (the payouts were too low anyway) but as matter of getting a feel for this method of monetization on the web. To make a long story short: It doesn’t work! A reasonable successful website (500 - 1000 unique users per day, on iphone-essentials.ch) brings in 1 to 3 USD in AdSense commisions per day. In my experience AdSense earnings scale very linearly with number of visitors, so I could have made maybe 20 USD if I had 10'000 uniques per day.
But I can’t. Not because I don’t have the visitors (I don’t), but because Google has decided that I’m a threat to their AdWords advertisers. How they arrived at this conclusion is a mystery to me: Google doesn’t explain what was wrong. They just cancelled my account, kept the balance of 160 USD that wasn’t yet paid to me and basically told me to f*** off.
With that introduction (which also might give you a hint of the bias I might have) onwards to the rest of the article.
The other gestation point is the upcoming Web 2.0 Expo in Berlin. I will be attending as part of the blogging programme. I will try to see, if the things I describe in this post are at all a part of the discussion.
I love Google. I have been using their search engine ever since I saw it the first time. Their web apps are pure joy. Goggle Maps, Google Mail, Google Docs: Wonders of usability and good value. I had seriously contemplated moving our complete mail infrastructure to Google Apps, for a very reasonable 60USD/user/year. I had started on using the collaboration features in Google Docs, writing, computing, sharing and collaboration on internal documents, and documents for clients. Google Chrome looks like a beautiful, speedy, reduce-to-the-max web browser and the few tests I made with it show great potential.
But this action of Google, cancelling my account with them, has changed that. We have all heard the discussions about Google and privacy. Anybody web savvy knows, that there ultimately is no privacy to be had on the internet. Too many traces are left on our trails through the vast web space. Government agencies are routinely listening in on communications, ISPs are required to keep track of all logins and all email communication, Google wants to know who you are and what you read to optimize the ads they are serving to you.
Enough! While we can’t erase all traces of our online lives, we can take care to not leave too much in this Web 2.0.
As an individual, or a small company, I have some limited possibilities. Each one does not amount to much, but together they make me and my employees less transparent, less open to snooping and give us back a bit of the privacy that I think we are entitled to.
And yes: This comes at a price… Both financially and in terms of ease of use. This price I am willing to pay.
We used to host our companies EMail at Joyent, a hosting company that I have used (and will continue to use) for many years. Additionally, I had a number of GMail accounts (for mailing lists, for sites where I didn’t want to divulge my real email, etc.). Our companies EMail was moved to Zimbra, hosted on my own server, in a data center down the street. Zimbra has a lot of features, email, calendaring, collaboration, iPhone integration etc. It’s fairly easy to set up, and has been performing well so far. The big advantage over Joyent? My data is hosted in Switzerland, under swiss jurisdiction. While I trust Joyent and the people behind Joyent, my data could still be seized by a court order. While this is far fetched in our companies case, large companies that move their mail to someone based in the US (like Google) might find all their private EMails subpoenaed by a US court. I’m divesting my Google mail accounts too, moving them, depending on need, to other free mail providers (like Yahoo) or my own server. The cost? The server costs, the co-location costs, Zimbra costs, setting it up costs. We are planning on divesting some of the cost, by offering hosted Zimbra to our customers. If you are interested in an email provider in Switzerland, give me a holler, and you can get a ride on my box. The benefits? My email is completely under my control. We get better and more features than previously (company calendaring, iPhone integration, great web UI)
Search
It’s almost impossible to be better than Google. MS Live Search and Yahoo search account for almost no traffic to a website (as any web master will tell you). New companies (like Cuil) sometimes make a big splash, but have little impact. Still, search is probably the most important application on the web. What to do? A project named YaCy is building a distributed, peer to peer, anonymous search engine. It is fledgling, indexes around 600 million web pages (depends on how many peers are online at any given moment), has a UI that could be improved quite a bit, needs a lot more work on getting search results as good as Googles (the PageRank algorithm really does make a difference!) and the software is in development. But the search is anonymous, the ranking algorithms are public, there is a way to help the search engine make better decisions. I have started to run a 24/7 peer in the YaCy network. So far it has crawled a bit over 2 million pages and thus contributes a tiny part to the web known to YaCy. I have set it up so that it tries to capture areas of the web that interest me: Rails, Ruby, Smalltalk. Also YaCy acts as proxy to my web browser and sees every page I read and uses that to expand the areas of the web it’s crawling. I have no illusion of YaCy being as good as Google any time soon. But I find it important to start to try to break the dominance of the 700 pound gorilla. My peer is open: search.invisible.ch, feel free to come by for your search needs. The cost? The server and the bandwidth (which I already pay for for our email). Not as good search results as the ones provided by Google. The benefits? Helping an important project. Giving less information to Google.
Network
The IP based network is here to stay, and little can be made of changing the infrastructure. However, the infrastructure can be ‘used’ in a slightly different way to thwart eavesdropping etc. An interesting project is Freenet which basically is a content-delivery darknet. It’s goal is to distribute information in a way, that the information can’t be censored, deleted and that participating nodes have no idea and no way of knowing, what content is stored and delivered on their servers. In addition, the servers themselves are stealthy: it’s not easy to detect them. I ran a Freenet node a couple of years ago and have donated to the project. However I have stopped, because I had no immediate use case for it. The available content (if you could get access to it) was of questionable or legally problematic nature. While the goals of Freenet are noble, and the technology is interesting, I don’t think it’s a viable solution for general use.
Of more interest to me is TOR. TOR consists of a series of servers, that route IP traffic in a non-traceable way. It’s not possible to tell for an eavesdropper, which websites I looked at, and the websites have no way to know, who I am based on my IP address. The TOR network is based on a large number of TOR servers, and users proxying through them. There are a few problems: even though the TOR net gives anonymity, it does not give you security, as the recent case of ‘hacked’ embassy email passwords showed. A ‘poisoned’ exit-node (the last server in a chain of servers, that contacts the intended destination server) was logging all mail related traffic and was able to extract hundreds of passwords.
If one runs an exit node, large amount of ‘inappropriate’ trafic (spam, if you allow mail) originates from your server: a sure way to get your IP listed on a blacklist.
I have been thinking of running a TOR node on my co-located server, and force all trafic from my office through it, but I’m not quite sure about the implications yet. One concern is speed: the routing of all traffic through multiple servers takes time and is limited by the bandwidth of the slowest link.
EMail, take2
There is yet another problem with EMail: It get’s transfered throug ISPs on the way and can be intercepted and read. Encryption of EMail is possible (PGP, S/MIME) but non standard and few people use it. A couple of years ago, I worked with 2 friends on Zappata which evolved over time into a secure P2P email system. We had a proof of concept working, could send and receive email, had built a PKI infrastructure that allowed authentication and encryption but in the end, the project fell apart because of lack of funding. I still think that the idea behind it is viable, and I would love to resurrect it: if you happen to have around USD 100'000, a strong interest in wide spread encryption, give me a shout - we could bring the project back to life and deliver a working solution within a few months.
Conclusion
While the net, and the web, are inherently insecure and traceable and there are strong corporate interests in collecting as much information about you as possible, there are ways to take back some control - at a cost. It’s up to the individual or the corporations to decide how much value they place on control over their data, their usage patterns, and how much they are willing to cede to get things free.
Me and my company have made the decision to ‘pay the price’ and start to become independent of the sweet smelling and looking offers from corporations like Google. I would hope, that more people will follow.